Sync configure with configure_n

Fix use of SuperLU_Dist, configure checks and ifdefs
Add message to configure
2026-10-07 07:04:59 +00:00 · 2022-04-14 20:33:02 +02:00 · 2022-04-13 16:36:32 +02:00 · 2022-04-07 10:36:28 +02:00 · 2022-04-06 11:02:32 +02:00 · 2022-04-06 09:02:20 +02:00
120 changed files with 4379 additions and 3313 deletions
@@ -126,7 +126,7 @@ module amg_c_base_aggregator_mod
           & psb_c_coo_sparse_mat, amg_sml_parms, psb_spk_, psb_ipk_, psb_lpk_
      implicit none
      type(psb_c_csr_sparse_mat), intent(inout) :: a_csr
-      type(psb_desc_type), intent(in)            :: desc_a
+      type(psb_desc_type), intent(inout)         :: desc_a
      integer(psb_lpk_), intent(inout)           :: nlaggr(:)
      type(amg_sml_parms), intent(inout)         :: parms 
      type(psb_c_coo_sparse_mat), intent(inout) :: coo_prol, coo_restr
@@ -144,7 +144,7 @@ module amg_c_base_aggregator_mod
           & psb_c_coo_sparse_mat, amg_sml_parms, psb_spk_, psb_ipk_, psb_lpk_
      implicit none
      type(psb_c_csr_sparse_mat), intent(inout) :: a_csr
-      type(psb_desc_type), intent(in)            :: desc_a
+      type(psb_desc_type), intent(inout)         :: desc_a
      integer(psb_lpk_), intent(inout)           :: nlaggr(:)
      type(amg_sml_parms), intent(inout)         :: parms 
      type(psb_c_coo_sparse_mat), intent(inout) :: coo_prol, coo_restr
@@ -126,7 +126,7 @@ module amg_d_base_aggregator_mod
           & psb_d_coo_sparse_mat, amg_dml_parms, psb_dpk_, psb_ipk_, psb_lpk_
      implicit none
      type(psb_d_csr_sparse_mat), intent(inout) :: a_csr
-      type(psb_desc_type), intent(in)            :: desc_a
+      type(psb_desc_type), intent(inout)         :: desc_a
      integer(psb_lpk_), intent(inout)           :: nlaggr(:)
      type(amg_dml_parms), intent(inout)         :: parms 
      type(psb_d_coo_sparse_mat), intent(inout) :: coo_prol, coo_restr
@@ -144,7 +144,7 @@ module amg_d_base_aggregator_mod
           & psb_d_coo_sparse_mat, amg_dml_parms, psb_dpk_, psb_ipk_, psb_lpk_
      implicit none
      type(psb_d_csr_sparse_mat), intent(inout) :: a_csr
-      type(psb_desc_type), intent(in)            :: desc_a
+      type(psb_desc_type), intent(inout)         :: desc_a
      integer(psb_lpk_), intent(inout)           :: nlaggr(:)
      type(amg_dml_parms), intent(inout)         :: parms 
      type(psb_d_coo_sparse_mat), intent(inout) :: coo_prol, coo_restr
@@ -68,7 +68,7 @@
 !    ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
 !    POSSIBILITY OF SUCH DAMAGE.
 !
-module dmatchboxp_mod
+module amg_d_matchboxp_mod

  use iso_c_binding
  use psb_base_cbind_mod
@@ -94,33 +94,25 @@ module dmatchboxp_mod
    end subroutine dMatchBoxPC
  end interface MatchBoxPC

-  interface i_aggr_assign
-    module procedure i_daggr_assign
-  end interface i_aggr_assign
+  interface amg_i_aggr_assign
+    module procedure amg_i_d_aggr_assign
+  end interface amg_i_aggr_assign

-  interface build_matching
-    module procedure dbuild_matching
-  end interface build_matching
+  interface amg_par_build_matching
+    module procedure amg_d_par_build_matching
+  end interface amg_par_build_matching

-  interface build_ahat
-    module procedure dbuild_ahat
-  end interface build_ahat
+  interface amg_par_build_ahat
+    module procedure amg_d_par_build_ahat
+  end interface amg_par_build_ahat

-  interface psb_gtranspose
-    module procedure psb_dgtranspose
-  end interface psb_gtranspose
-
-  interface psb_htranspose
-    module procedure psb_dhtranspose
-  end interface psb_htranspose
-
-  interface PMatchBox
-    module procedure dPMatchBox
-  end interface PMatchBox
+  interface amg_PMatchBox
+    module procedure amg_d_PMatchBox
+  end interface amg_PMatchBox

 contains

-  subroutine dmatchboxp_build_prol(w,a,desc_a,ilaggr,nlaggr,prol,info,&
+  subroutine amg_d_matchboxp_build_prol(w,a,desc_a,ilaggr,nlaggr,prol,info,&
       & symmetrize,reproducible,display_inp, display_out, print_out)
    use psb_base_mod
    use psb_util_mod
@@ -213,7 +205,7 @@ contains
    end if
    if (do_timings) call psb_toc(idx_phase1)
    if (do_timings) call psb_tic(idx_bldmtc)
-    call build_matching(w,a,desc_a,mate,info,display_inp=display_inp,symmetrize=symmetrize)
+    call amg_par_build_matching(w,a,desc_a,mate,info,display_inp=display_inp,symmetrize=symmetrize)
    if (do_timings) call psb_toc(idx_bldmtc)
    if (debug) write(0,*) iam,' buildprol from buildmatching:',&
         & info
@@ -311,7 +303,7 @@ contains
                  ! Should be a symmetric function.
                  !
                  call desc_a%indxmap%qry_halo_owner(idx,iown,info)
-                  ip    = i_aggr_assign(iam, iown, kg, idxg, wk, widx, nrmagg)
+                  ip    = amg_i_aggr_assign(iam, iown, kg, idxg, wk, widx, nrmagg)
                  if (iam == ip) then
                    nlaggr(iam) = nlaggr(iam) + 1
                    ilaggr(k)   = nlaggr(iam)
@@ -513,9 +505,9 @@ contains
      write(0,*) iam,' :  error from Matching: ',info
    end if

-  end subroutine dmatchboxp_build_prol
+  end subroutine amg_d_matchboxp_build_prol

-  function i_daggr_assign(iam, iown, kg, idxg, wk, widx, nrmagg) &
+  function amg_i_d_aggr_assign(iam, iown, kg, idxg, wk, widx, nrmagg) &
       & result(iproc)
    !
    ! How to break ties? This
@@ -557,10 +549,10 @@ contains
        iproc = iown
      end if
    end if
-  end function i_daggr_assign
+  end function amg_i_d_aggr_assign


-  subroutine dbuild_matching(w,a,desc_a,mate,info,display_inp, symmetrize)
+  subroutine amg_d_par_build_matching(w,a,desc_a,mate,info,display_inp, symmetrize)
    use psb_base_mod
    use psb_util_mod
    use iso_c_binding
@@ -609,7 +601,7 @@ contains
      if (iam == 0) write(0,*)' Into build_ahat:'
    end if
    if (do_timings) call psb_tic(idx_bldahat)
-    call build_ahat(w,a,ahatnd,desc_a,info,symmetrize=symmetrize)
+    call amg_par_build_ahat(w,a,ahatnd,desc_a,info,symmetrize=symmetrize)
    if (do_timings) call psb_toc(idx_bldahat)
    if (info /= 0) then
      write(0,*) 'Error from build_ahat ', info
@@ -700,7 +692,7 @@ contains
      !
      if (debug) write(0,*) iam,' buildmatching into PMatchBox:'
      if (do_timings) call psb_tic(idx_cmboxp)
-      call PMatchBox(nr,nz,vlptr,vlind,ewght,&
+      call amg_PMatchBox(nr,nz,vlptr,vlind,ewght,&
           & vnl, mate, iam, np,ictxt,&
           & msgis,msgas,msgprc,ph0t,ph1t,ph2t,ph1crd,ph2crd,info,display_inp)
      if (do_timings) call psb_toc(idx_cmboxp)
@@ -764,9 +756,9 @@ contains
      val(1:n) = tmp(1:n)
    end subroutine fix_order

-  end subroutine dbuild_matching
+  end subroutine amg_d_par_build_matching

-  subroutine dbuild_ahat(w,a,ahat,desc_a,info,symmetrize)
+  subroutine amg_d_par_build_ahat(w,a,ahat,desc_a,info,symmetrize)
    use psb_base_mod
    implicit none
    real(psb_dpk_), intent(in)           :: w(:)
@@ -1002,301 +994,9 @@ contains
      end block
    end if

-  end subroutine dbuild_ahat
+  end subroutine amg_d_par_build_ahat

-  subroutine psb_dgtranspose(ain,aout,desc_a,info)
-    use psb_base_mod
-    implicit none
-    type(psb_ldspmat_type), intent(in)  :: ain
-    type(psb_ldspmat_type), intent(out) :: aout
-    type(psb_desc_type)           :: desc_a
-    integer(psb_ipk_), intent(out) :: info
-
-    !
-    ! BEWARE: This routine works under the assumption
-    ! that the same DESC_A works for both A and A^T, which
-    ! essentially means that A has a symmetric pattern.
-    !
-    type(psb_ldspmat_type) :: atmp, ahalo, aglb
-    type(psb_ld_coo_sparse_mat) :: tmpcoo
-    type(psb_ld_csr_sparse_mat) :: tmpcsr
-    type(psb_ctxt_type) :: ictxt
-    integer(psb_ipk_) :: me, np
-    integer(psb_lpk_) :: i, j, k, nrow, ncol
-    integer(psb_lpk_), allocatable :: ilv(:)
-    character(len=80) :: aname
-    logical, parameter :: debug=.false., dump=.false., debug_sync=.false.
-
-    ictxt = desc_a%get_context()
-    call psb_info(ictxt,me,np)
-
-    nrow = desc_a%get_local_rows()
-    ncol = desc_a%get_local_cols()
-    if (debug_sync) then
-      call psb_barrier(ictxt)
-      if (me == 0) write(0,*) 'Start gtranspose '
-    end if
-    call ain%cscnv(tmpcsr,info)
-
-    if (debug) then
-      ilv = [(i,i=1,ncol)]
-      call desc_a%l2gip(ilv,info,owned=.false.)
-      write(aname,'(a,i3.3,a)') 'atmp-preh-',me,'.mtx'
-      call ain%print(fname=aname,head='atmp before haloTest ',iv=ilv)
-    end if
-    if (dump) then
-      call ain%cscnv(atmp,info)
-      call psb_gather(aglb,atmp,desc_a,info)
-      if (me==psb_root_) then
-        write(aname,'(a,i3.3,a)') 'aglob-prehalo.mtx'
-        call aglb%print(fname=aname,head='Test ')
-      end if
-    end if
-
-    !call psb_loc_to_glob(tmpcsr%ja,desc_a,info)
-    call atmp%mv_from(tmpcsr)
-
-    if (debug) then
-      write(aname,'(a,i3.3,a)') 'tmpcsr-',me,'.mtx'
-      call atmp%print(fname=aname,head='tmpcsr ',iv=ilv)
-      !call psb_set_debug_level(9999)
-    end if
-
-    ! FIXME THIS NEEDS REWORKING
-    if (debug) write(0,*) me,' Gtranspose into sphalo :',atmp%get_nrows(),atmp%get_ncols()
-    call psb_sphalo(atmp,desc_a,ahalo,info,rowscale=.true.)
-    if (debug) write(0,*) me,' Gtranspose from sphalo :',ahalo%get_nrows(),ahalo%get_ncols()
-    if (info == psb_success_) call psb_rwextd(ncol,atmp,info,b=ahalo)
-
-    if (debug) then
-      write(aname,'(a,i3.3,a)') 'ahalo-',me,'.mtx'
-      call ahalo%print(fname=aname,head='ahalo after haloTest ',iv=ilv)
-      write(aname,'(a,i3.3,a)') 'atmp-h-',me,'.mtx'
-      call atmp%print(fname=aname,head='atmp after haloTest ',iv=ilv)
-    end if
-
-    if (info == psb_success_) call ahalo%free()
-
-    call atmp%cp_to(tmpcoo)
-    call tmpcoo%transp()
-    !call psb_glob_to_loc(tmpcoo%ia,desc_a,info,iact='I')
-    if (debug) write(0,*) 'Before cleanup:',tmpcoo%get_nzeros()
-
-    j = 0
-    do k=1, tmpcoo%get_nzeros()
-      if ((tmpcoo%ia(k) > 0).and.(tmpcoo%ja(k)>0)) then
-        j = j+1
-        tmpcoo%ia(j)  = tmpcoo%ia(k)
-        tmpcoo%ja(j)  = tmpcoo%ja(k)
-        tmpcoo%val(j) = tmpcoo%val(k)
-      end if
-    end do
-    call tmpcoo%set_nzeros(j)
-
-    if (debug) write(0,*) 'After cleanup:',tmpcoo%get_nzeros()
-
-    call ahalo%mv_from(tmpcoo)
-    if (dump) then
-      call psb_gather(aglb,ahalo,desc_a,info)
-      if (me==psb_root_) then
-        write(aname,'(a,i3.3,a)') 'atran-preclip.mtx'
-        call aglb%print(fname=aname,head='Test ')
-      end if
-    end if
-
-
-    call ahalo%csclip(aout,info,imax=nrow)
-
-    if (debug) write(0,*) 'After clip:',aout%get_nzeros()
-
-    if (debug_sync) then
-      call psb_barrier(ictxt)
-      if (me == 0) write(0,*) 'End gtranspose '
-    end if
-    !call aout%cscnv(info,type='csr')
-
-    if (dump) then
-      write(aname,'(a,i3.3,a)') 'atran-',me,'.mtx'
-      call aout%print(fname=aname,head='atrans ',iv=ilv)
-      call psb_gather(aglb,aout,desc_a,info)
-      if (me==psb_root_) then
-        write(aname,'(a,i3.3,a)') 'atran.mtx'
-        call aglb%print(fname=aname,head='Test ')
-      end if
-    end if
-
-  end subroutine psb_dgtranspose
-
-  subroutine psb_dhtranspose(ain,aout,desc_a,info)
-    use psb_base_mod
-    implicit none
-    type(psb_ldspmat_type), intent(in)  :: ain
-    type(psb_ldspmat_type), intent(out) :: aout
-    type(psb_desc_type)           :: desc_a
-    integer(psb_ipk_), intent(out) :: info
-
-    !
-    ! BEWARE: This routine works under the assumption
-    ! that the same DESC_A works for both A and A^T, which
-    ! essentially means that A has a symmetric pattern.
-    !
-    type(psb_ldspmat_type) :: atmp, ahalo, aglb
-    type(psb_ld_coo_sparse_mat) :: tmpcoo, tmpc1, tmpc2, tmpch
-    type(psb_ld_csr_sparse_mat) :: tmpcsr
-    integer(psb_ipk_) :: nz1, nz2, nzh, nz
-    type(psb_ctxt_type) :: ictxt
-    integer(psb_ipk_) :: me, np
-    integer(psb_lpk_) :: i, j, k, nrow, ncol, nlz
-    integer(psb_lpk_), allocatable :: ilv(:)
-    character(len=80) :: aname
-    logical, parameter :: debug=.false., dump=.false., debug_sync=.false.
-
-    ictxt = desc_a%get_context()
-    call psb_info(ictxt,me,np)
-
-    nrow = desc_a%get_local_rows()
-    ncol = desc_a%get_local_cols()
-    if (debug_sync) then
-      call psb_barrier(ictxt)
-      if (me == 0) write(0,*) 'Start htranspose '
-    end if
-    call ain%cscnv(tmpcsr,info)
-
-    if (debug) then
-      ilv = [(i,i=1,ncol)]
-      call desc_a%l2gip(ilv,info,owned=.false.)
-      write(aname,'(a,i3.3,a)') 'atmp-preh-',me,'.mtx'
-      call ain%print(fname=aname,head='atmp before haloTest ',iv=ilv)
-    end if
-    if (dump) then
-      call ain%cscnv(atmp,info)
-      call psb_gather(aglb,atmp,desc_a,info)
-      if (me==psb_root_) then
-        write(aname,'(a,i3.3,a)') 'aglob-prehalo.mtx'
-        call aglb%print(fname=aname,head='Test ')
-      end if
-    end if
-
-    !call psb_loc_to_glob(tmpcsr%ja,desc_a,info)
-    call atmp%mv_from(tmpcsr)
-
-    if (debug) then
-      write(aname,'(a,i3.3,a)') 'tmpcsr-',me,'.mtx'
-      call atmp%print(fname=aname,head='tmpcsr ',iv=ilv)
-      !call psb_set_debug_level(9999)
-    end if
-
-    ! FIXME THIS NEEDS REWORKING
-    if (debug) write(0,*) me,' Htranspose into sphalo :',atmp%get_nrows(),atmp%get_ncols()
-    if (.true.) then
-      call psb_sphalo(atmp,desc_a,ahalo,info, outfmt='coo  ')
-      call atmp%mv_to(tmpc1)
-      call ahalo%mv_to(tmpch)
-      nz1 = tmpc1%get_nzeros()
-      call psb_loc_to_glob(tmpc1%ia(1:nz1),desc_a,info,iact='I')
-      call psb_loc_to_glob(tmpc1%ja(1:nz1),desc_a,info,iact='I')
-      nzh = tmpch%get_nzeros()
-      call psb_loc_to_glob(tmpch%ia(1:nzh),desc_a,info,iact='I')
-      call psb_loc_to_glob(tmpch%ja(1:nzh),desc_a,info,iact='I')
-      nlz = nz1+nzh
-      call tmpcoo%allocate(ncol,ncol,nlz)
-      tmpcoo%ia(1:nz1) = tmpc1%ia(1:nz1)
-      tmpcoo%ja(1:nz1) = tmpc1%ja(1:nz1)
-      tmpcoo%val(1:nz1) = tmpc1%val(1:nz1)
-      tmpcoo%ia(nz1+1:nz1+nzh) = tmpch%ia(1:nzh)
-      tmpcoo%ja(nz1+1:nz1+nzh) = tmpch%ja(1:nzh)
-      tmpcoo%val(nz1+1:nz1+nzh) = tmpch%val(1:nzh)
-      call tmpcoo%set_nzeros(nlz)
-      call tmpcoo%transp()
-      nz = tmpcoo%get_nzeros()
-      call psb_glob_to_loc(tmpcoo%ia(1:nz),desc_a,info,iact='I')
-      call psb_glob_to_loc(tmpcoo%ja(1:nz),desc_a,info,iact='I')
-      if (.true.) then
-        call tmpcoo%clean_negidx(info)
-      else
-        j = 0
-        do k=1, tmpcoo%get_nzeros()
-          if ((tmpcoo%ia(k) > 0).and.(tmpcoo%ja(k)>0)) then
-            j = j+1
-            tmpcoo%ia(j)  = tmpcoo%ia(k)
-            tmpcoo%ja(j)  = tmpcoo%ja(k)
-            tmpcoo%val(j) = tmpcoo%val(k)
-          end if
-        end do
-        call tmpcoo%set_nzeros(j)
-      end if
-      call ahalo%mv_from(tmpcoo)
-      call ahalo%csclip(aout,info,imax=nrow)
-
-    else
-      call psb_sphalo(atmp,desc_a,ahalo,info, rowscale=.true.)
-      if (debug) write(0,*) me,' Htranspose from sphalo :',ahalo%get_nrows(),ahalo%get_ncols()
-      if (info == psb_success_) call psb_rwextd(ncol,atmp,info,b=ahalo)
-
-      if (debug) then
-        write(aname,'(a,i3.3,a)') 'ahalo-',me,'.mtx'
-        call ahalo%print(fname=aname,head='ahalo after haloTest ',iv=ilv)
-        write(aname,'(a,i3.3,a)') 'atmp-h-',me,'.mtx'
-        call atmp%print(fname=aname,head='atmp after haloTest ',iv=ilv)
-      end if
-
-      if (info == psb_success_) call ahalo%free()
-
-      call atmp%cp_to(tmpcoo)
-      call tmpcoo%transp()
-      if (debug) write(0,*) 'Before cleanup:',tmpcoo%get_nzeros()
-      if (.true.) then
-        call tmpcoo%clean_negidx(info)
-      else
-
-        j = 0
-        do k=1, tmpcoo%get_nzeros()
-          if ((tmpcoo%ia(k) > 0).and.(tmpcoo%ja(k)>0)) then
-            j = j+1
-            tmpcoo%ia(j)  = tmpcoo%ia(k)
-            tmpcoo%ja(j)  = tmpcoo%ja(k)
-            tmpcoo%val(j) = tmpcoo%val(k)
-          end if
-        end do
-        call tmpcoo%set_nzeros(j)
-      end if
-
-      if (debug) write(0,*) 'After cleanup:',tmpcoo%get_nzeros()
-
-      call ahalo%mv_from(tmpcoo)
-      if (dump) then
-        call psb_gather(aglb,ahalo,desc_a,info)
-        if (me==psb_root_) then
-          write(aname,'(a,i3.3,a)') 'atran-preclip.mtx'
-          call aglb%print(fname=aname,head='Test ')
-        end if
-      end if
-
-
-      call ahalo%csclip(aout,info,imax=nrow)
-    end if
-
-    if (debug) write(0,*) 'After clip:',aout%get_nzeros()
-
-    if (debug_sync) then
-      call psb_barrier(ictxt)
-      if (me == 0) write(0,*) 'End htranspose '
-    end if
-    !call aout%cscnv(info,type='csr')
-
-    if (dump) then
-      write(aname,'(a,i3.3,a)') 'atran-',me,'.mtx'
-      call aout%print(fname=aname,head='atrans ',iv=ilv)
-      call psb_gather(aglb,aout,desc_a,info)
-      if (me==psb_root_) then
-        write(aname,'(a,i3.3,a)') 'atran.mtx'
-        call aglb%print(fname=aname,head='Test ')
-      end if
-    end if
-
-  end subroutine psb_dhtranspose
-
-  subroutine dPMatchBox(nlver,nledge,verlocptr,verlocind,edgelocweight,&
+  subroutine amg_d_PMatchBox(nlver,nledge,verlocptr,verlocind,edgelocweight,&
       & verdistance, mate, myrank, numprocs, ictxt,&
       & msgindsent,msgactualsent,msgpercent,&
       & ph0_time, ph1_time, ph2_time, ph1_card, ph2_card,info,display_inp)
@@ -1431,6 +1131,6 @@ contains
    end if
    where(mate>=0) mate = mate + 1

-  end subroutine dPMatchBox
+  end subroutine amg_d_PMatchBox

-end module dmatchboxp_mod
+end module amg_d_matchboxp_mod
@@ -118,7 +118,7 @@

 module amg_d_parmatch_aggregator_mod
  use amg_d_base_aggregator_mod
-  use dmatchboxp_mod
+  use amg_d_matchboxp_mod
 #if defined(SERIAL_MPI)
  type, extends(amg_d_base_aggregator_type) :: amg_d_parmatch_aggregator_type
  end type amg_d_parmatch_aggregator_type
@@ -132,8 +132,6 @@ module amg_d_parmatch_aggregator_mod
    type(psb_dspmat_type), allocatable  :: prol, restr
    type(psb_dspmat_type), allocatable  :: ac, base_a, rwa
    type(psb_desc_type), allocatable    :: desc_ac, desc_ax, base_desc, rwdesc
-    integer(psb_ipk_) :: max_csize
-    integer(psb_ipk_) :: max_nlevels
    logical           :: reproducible_matching = .false.
    logical           :: need_symmetrize       = .false.
    logical           :: unsmoothed_hierarchy  = .true.
@@ -143,18 +141,18 @@ module amg_d_parmatch_aggregator_mod
    procedure, pass(ag) :: mat_asb      => amg_d_parmatch_aggregator_mat_asb
    procedure, pass(ag) :: inner_mat_asb      => amg_d_parmatch_aggregator_inner_mat_asb
    procedure, pass(ag) :: bld_map      => amg_d_parmatch_aggregator_bld_map
-    procedure, pass(ag) :: csetc        => d_parmatch_aggr_csetc
-    procedure, pass(ag) :: cseti        => d_parmatch_aggr_cseti
-    procedure, pass(ag) :: default      => d_parmatch_aggr_set_default
-    procedure, pass(ag) :: sizeof       => d_parmatch_aggregator_sizeof
-    procedure, pass(ag) :: update_next  => d_parmatch_aggregator_update_next
-    procedure, pass(ag) :: bld_wnxt     => d_parmatch_bld_wnxt
-    procedure, pass(ag) :: bld_default_w    => d_bld_default_w
-    procedure, pass(ag) :: set_c_default_w  => d_set_prm_c_default_w
-    procedure, pass(ag) :: descr        => d_parmatch_aggregator_descr
-    procedure, pass(ag) :: clone        => d_parmatch_aggregator_clone
-    procedure, pass(ag) :: free         => d_parmatch_aggregator_free
-    procedure, nopass   :: fmt          => d_parmatch_aggregator_fmt
+    procedure, pass(ag) :: csetc        => amg_d_parmatch_aggr_csetc
+    procedure, pass(ag) :: cseti        => amg_d_parmatch_aggr_cseti
+    procedure, pass(ag) :: default      => amg_d_parmatch_aggr_set_default
+    procedure, pass(ag) :: sizeof       => amg_d_parmatch_aggregator_sizeof
+    procedure, pass(ag) :: update_next  => amg_d_parmatch_aggregator_update_next
+    procedure, pass(ag) :: bld_wnxt     => amg_d_parmatch_bld_wnxt
+    procedure, pass(ag) :: bld_default_w    => amg_d_bld_default_w
+    procedure, pass(ag) :: set_c_default_w  => amg_d_set_prm_c_default_w
+    procedure, pass(ag) :: descr        => amg_d_parmatch_aggregator_descr
+    procedure, pass(ag) :: clone        => amg_d_parmatch_aggregator_clone
+    procedure, pass(ag) :: free         => amg_d_parmatch_aggregator_free
+    procedure, nopass   :: fmt          => amg_d_parmatch_aggregator_fmt
    procedure, nopass   :: xt_desc      => amg_d_parmatch_aggregator_xt_desc
  end type amg_d_parmatch_aggregator_type

@@ -168,7 +166,7 @@ module amg_d_parmatch_aggregator_mod
      type(amg_dml_parms), intent(inout)   :: parms
      type(amg_daggr_data), intent(in)     :: ag_data
      type(psb_dspmat_type), intent(inout) :: a
-      type(psb_desc_type), intent(inout)   :: desc_a
+      type(psb_desc_type), intent(inout)     :: desc_a
      integer(psb_lpk_), allocatable, intent(out) :: ilaggr(:),nlaggr(:)
      type(psb_ldspmat_type), intent(out)  :: t_prol
      integer(psb_ipk_), intent(out)      :: info
@@ -235,7 +233,7 @@ module amg_d_parmatch_aggregator_mod
           & psb_ldspmat_type, psb_dpk_, psb_ipk_, psb_lpk_, amg_dml_parms, amg_daggr_data
      implicit none
      type(psb_dspmat_type), intent(in)       :: a
-      type(psb_desc_type), intent(in)         :: desc_a
+      type(psb_desc_type), intent(inout)       :: desc_a
      integer(psb_lpk_), intent(inout)        :: ilaggr(:), nlaggr(:)
      type(amg_dml_parms), intent(inout)      :: parms
      type(psb_ldspmat_type), intent(inout)   :: t_prol
@@ -257,7 +255,7 @@ module amg_d_parmatch_aggregator_mod
      integer(psb_lpk_), intent(inout)        :: ilaggr(:), nlaggr(:)
      type(amg_dml_parms), intent(inout)      :: parms
      type(psb_ldspmat_type), intent(inout)   :: t_prol
-      type(psb_dspmat_type), intent(out)      :: op_prol,ac, op_restr
+      type(psb_dspmat_type), intent(inout)      :: op_prol,ac, op_restr
      type(psb_desc_type), intent(inout)    :: desc_ac
      integer(psb_ipk_), intent(out)          :: info
    end subroutine amg_d_parmatch_unsmth_bld
@@ -275,7 +273,7 @@ module amg_d_parmatch_aggregator_mod
      integer(psb_lpk_), intent(inout)        :: ilaggr(:), nlaggr(:)
      type(amg_dml_parms), intent(inout)      :: parms
      type(psb_ldspmat_type), intent(inout)   :: t_prol
-      type(psb_dspmat_type), intent(out)      :: op_prol,ac, op_restr
+      type(psb_dspmat_type), intent(inout)    :: op_prol,ac, op_restr
      type(psb_desc_type), intent(inout)    :: desc_ac
      integer(psb_ipk_), intent(out)          :: info
    end subroutine amg_d_parmatch_smth_bld
@@ -288,11 +286,11 @@ module amg_d_parmatch_aggregator_mod
           & psb_ldspmat_type, psb_dpk_, psb_ipk_, psb_lpk_, amg_dml_parms, amg_daggr_data
      implicit none
      type(psb_dspmat_type), intent(inout)    :: a
-      type(psb_desc_type), intent(in)         :: desc_a
+      type(psb_desc_type), intent(inout)       :: desc_a
      integer(psb_lpk_), intent(inout)        :: ilaggr(:), nlaggr(:)
      type(amg_dml_parms), intent(inout)      :: parms
      type(psb_ldspmat_type), intent(inout)   :: t_prol
-      type(psb_dspmat_type), intent(out)      :: op_prol,ac, op_restr
+      type(psb_dspmat_type), intent(inout)    :: op_prol,ac, op_restr
      type(psb_desc_type), intent(out)          :: desc_ac
      integer(psb_ipk_), intent(out)          :: info
    end subroutine amg_d_parmatch_spmm_bld_ov
@@ -306,11 +304,11 @@ module amg_d_parmatch_aggregator_mod
           & psb_d_csr_sparse_mat, psb_ld_csr_sparse_mat
      implicit none
      type(psb_d_csr_sparse_mat), intent(inout) :: a
-      type(psb_desc_type), intent(in)           :: desc_a
+      type(psb_desc_type), intent(inout)          :: desc_a
      integer(psb_lpk_), intent(inout)          :: ilaggr(:), nlaggr(:)
      type(amg_dml_parms), intent(inout)        :: parms
      type(psb_ldspmat_type), intent(inout)     :: t_prol
-      type(psb_dspmat_type), intent(out)        :: op_prol,ac, op_restr
+      type(psb_dspmat_type), intent(inout)      :: op_prol,ac, op_restr
      type(psb_desc_type), intent(out)          :: desc_ac
      integer(psb_ipk_), intent(out)            :: info
    end subroutine amg_d_parmatch_spmm_bld_inner
@@ -320,7 +318,7 @@ module amg_d_parmatch_aggregator_mod

 contains

-  subroutine d_bld_default_w(ag,nr)
+  subroutine amg_d_bld_default_w(ag,nr)
    use psb_realloc_mod
    implicit none
    class(amg_d_parmatch_aggregator_type), target, intent(inout) :: ag
@@ -330,9 +328,9 @@ contains
    if (info /= psb_success_) return
    ag%w = done
    !call ag%set_c_default_w()
-  end subroutine d_bld_default_w
+  end subroutine amg_d_bld_default_w

-  subroutine d_set_prm_c_default_w(ag)
+  subroutine amg_d_set_prm_c_default_w(ag)
    use psb_realloc_mod
    use iso_c_binding
    implicit none
@@ -342,9 +340,9 @@ contains
    !write(0,*) 'prm_c_deafult_w '
    call psb_safe_ab_cpy(ag%w,ag%w_nxt,info)

-  end subroutine d_set_prm_c_default_w
+  end subroutine amg_d_set_prm_c_default_w

-  subroutine d_parmatch_bld_wnxt(ag,ilaggr,valaggr,nx)
+  subroutine amg_d_parmatch_bld_wnxt(ag,ilaggr,valaggr,nx)
    use psb_realloc_mod
    implicit none
    class(amg_d_parmatch_aggregator_type), target, intent(inout) :: ag
@@ -358,14 +356,14 @@ contains
    !write(0,*) 'Executing bld_wnxt ',nx
    call psb_realloc(nx,ag%w_nxt,info)

-  end subroutine d_parmatch_bld_wnxt
+  end subroutine amg_d_parmatch_bld_wnxt

-  function d_parmatch_aggregator_fmt() result(val)
+  function amg_d_parmatch_aggregator_fmt() result(val)
    implicit none
    character(len=32)  :: val

    val = "Parallel Matching aggregation"
-  end function d_parmatch_aggregator_fmt
+  end function amg_d_parmatch_aggregator_fmt

  function amg_d_parmatch_aggregator_xt_desc() result(val)
    implicit none
@@ -374,7 +372,7 @@ contains
    val = .true.
  end function amg_d_parmatch_aggregator_xt_desc

-  function d_parmatch_aggregator_sizeof(ag) result(val)
+  function amg_d_parmatch_aggregator_sizeof(ag) result(val)
    use psb_realloc_mod
    implicit none
    class(amg_d_parmatch_aggregator_type), intent(in)  :: ag
@@ -390,9 +388,9 @@ contains
    if (allocated(ag%base_desc)) val = val + ag%base_desc%sizeof()
    if (allocated(ag%desc_ax)) val = val + ag%desc_ax%sizeof()

-  end function d_parmatch_aggregator_sizeof
+  end function amg_d_parmatch_aggregator_sizeof

-  subroutine  d_parmatch_aggregator_descr(ag,parms,iout,info)
+  subroutine  amg_d_parmatch_aggregator_descr(ag,parms,iout,info)
    implicit none
    class(amg_d_parmatch_aggregator_type), intent(in) :: ag
    type(amg_dml_parms), intent(in)   :: parms
@@ -406,7 +404,7 @@ contains
    call parms%mldescr(iout,info)

    return
-  end subroutine d_parmatch_aggregator_descr
+  end subroutine amg_d_parmatch_aggregator_descr

  function is_legal_malg(alg) result(val)
    logical :: val
@@ -437,7 +435,7 @@ contains
  end function is_legal_nlevels


-  subroutine  d_parmatch_aggregator_update_next(ag,agnext,info)
+  subroutine  amg_d_parmatch_aggregator_update_next(ag,agnext,info)
    use psb_realloc_mod
    implicit none
    class(amg_d_parmatch_aggregator_type), target, intent(inout) :: ag
@@ -452,10 +450,10 @@ contains
           & agnext%matching_alg = ag%matching_alg
      if (.not.is_legal_nsweeps(agnext%n_sweeps))&
           & agnext%n_sweeps     = ag%n_sweeps
-      if (.not.is_legal_csize(agnext%max_csize))&
-           & agnext%max_csize    = ag%max_csize
-      if (.not.is_legal_nlevels(agnext%max_nlevels))&
-           & agnext%max_nlevels  = ag%max_nlevels
+!!$      if (.not.is_legal_csize(agnext%max_csize))&
+!!$           & agnext%max_csize    = ag%max_csize
+!!$      if (.not.is_legal_nlevels(agnext%max_nlevels))&
+!!$           & agnext%max_nlevels  = ag%max_nlevels
      ! Is this going to generate shallow copies/memory leaks/double frees?
      ! To be investigated further.
      call psb_safe_ab_cpy(ag%w_nxt,agnext%w,info)
@@ -470,9 +468,9 @@ contains
      ! What should we do here?
    end select
    info = 0
-  end subroutine d_parmatch_aggregator_update_next
+  end subroutine amg_d_parmatch_aggregator_update_next

-  subroutine d_parmatch_aggr_csetc(ag,what,val,info,idx)
+  subroutine amg_d_parmatch_aggr_csetc(ag,what,val,info,idx)

    Implicit None

@@ -514,9 +512,9 @@ contains
      ! Do nothing
    end select
    return
-  end subroutine d_parmatch_aggr_csetc
+  end subroutine amg_d_parmatch_aggr_csetc

-  subroutine d_parmatch_aggr_cseti(ag,what,val,info,idx)
+  subroutine amg_d_parmatch_aggr_cseti(ag,what,val,info,idx)

    Implicit None

@@ -540,10 +538,6 @@ contains
    case('AGGR_SIZE')
      ag%orig_aggr_size = val
      ag%n_sweeps=max(1,ceiling(log(val*1.0)/log(2.0)))
-    case('PRMC_MAX_CSIZE')
-      ag%max_csize=val
-    case('PRMC_MAX_NLEVELS')
-      ag%max_nlevels=val
    case('PRMC_W_SIZE')
      call ag%bld_default_w(val)
    case('PRMC_REPRODUCIBLE_MATCHING')
@@ -556,9 +550,9 @@ contains
      ! Do nothing
    end select
    return
-  end subroutine d_parmatch_aggr_cseti
+  end subroutine amg_d_parmatch_aggr_cseti

-  subroutine d_parmatch_aggr_set_default(ag)
+  subroutine amg_d_parmatch_aggr_set_default(ag)

    Implicit None

@@ -569,8 +563,8 @@ contains
    ag%matching_alg  = 0
    ag%n_sweeps      = 1
    ag%jacobi_sweeps = 0
-    ag%max_nlevels   = 36
-    ag%max_csize     = -1
+!!$    ag%max_nlevels   = 36
+!!$    ag%max_csize     = -1
    !
    ! Apparently BootCMatch works better
    ! by keeping all entries
@@ -579,9 +573,9 @@ contains

    return

-  end subroutine d_parmatch_aggr_set_default
+  end subroutine amg_d_parmatch_aggr_set_default

-  subroutine  d_parmatch_aggregator_free(ag,info)
+  subroutine  amg_d_parmatch_aggregator_free(ag,info)
    use iso_c_binding
    implicit none
    class(amg_d_parmatch_aggregator_type), intent(inout) :: ag
@@ -618,9 +612,9 @@ contains
      call ag%rwdesc%free(info); deallocate(ag%rwdesc,stat=info)
    end if

-  end subroutine d_parmatch_aggregator_free
+  end subroutine amg_d_parmatch_aggregator_free

-  subroutine  d_parmatch_aggregator_clone(ag,agnext,info)
+  subroutine  amg_d_parmatch_aggregator_clone(ag,agnext,info)
    implicit none
    class(amg_d_parmatch_aggregator_type), intent(inout) :: ag
    class(amg_d_base_aggregator_type), allocatable, intent(inout) :: agnext
@@ -640,7 +634,7 @@ contains
      ! Should never ever get here
      info = -1
    end select
-  end subroutine d_parmatch_aggregator_clone
+  end subroutine amg_d_parmatch_aggregator_clone

  subroutine  amg_d_parmatch_aggregator_bld_map(ag,desc_a,desc_ac,ilaggr,nlaggr,&
       & op_restr,op_prol,map,info)
@@ -52,7 +52,7 @@ module amg_d_sludist_solver
  use iso_c_binding
  use amg_d_base_solver_mod

-#if defined(LPK8) 
+#if (!defined(HAVE_SLUDIST_)) || defined(IPK8) 

  type, extends(amg_d_base_solver_type) :: amg_d_sludist_solver_type

@@ -270,11 +270,13 @@ contains
    ! Local variables
    type(psb_dspmat_type) :: atmp
    type(psb_d_csr_sparse_mat) :: acsr
-    integer :: n_row,n_col, nrow_a, nztota, nglob, nzt, npr, npc
-    integer :: ifrst, ibcheck
+    integer(psb_lpk_), allocatable :: gia(:), gja(:)
    type(psb_ctxt_type) :: ctxt
-    integer :: np,me,i, err_act, debug_unit, debug_level
-    character(len=20)  :: name='d_sludist_solver_bld', ch_err
+    integer(psb_lpk_) :: lfrst
+    integer(psb_ipk_) :: n_row,n_col, nrow_a, nztota, nglob, nzt, npr, npc
+    integer(psb_ipk_) :: ifrst, ibcheck
+    integer(psb_ipk_) :: np,me,i, err_act, debug_unit, debug_level
+    character(len=20) :: name='d_sludist_solver_bld', ch_err
    
    info=psb_success_
    call psb_erractionsave(err_act)
@@ -293,19 +295,37 @@ contains
    n_col  = desc_a%get_local_cols()
    nglob  = desc_a%get_global_rows()
    
-    call a%cscnv(atmp,info,type='coo')
+    !
+    ! Strategy here is as follows: because a call to SLUDIST
+    ! as a gobal solver is mostly done  at the coarsest level,
+    ! even if we start from a problem requiring 8 bytes, chances
+    ! are that the global size will be suitable for 4 bytes
+    ! anyway, so we hope for the best, and throw an error
+    ! if something goes wrong.
+    !
+    if (nglob > huge(1_psb_ipk_)) then
+      write(0,*) me,' ',trim(name),': Error: overflow of local indices '
+      info=psb_err_internal_error_
+      call psb_errpush(info,name)
+      goto 9999
+    end if
+    
+    call a%cscnv(atmp,info,type='csr')
+    ! This in case we are dealing with AS
    call psb_rwextd(n_row,atmp,info,b=b) 
-    call atmp%cscnv(info,type='csr',dupl=psb_dupl_add_)
    call atmp%mv_to(acsr)
    nrow_a = acsr%get_nrows()
    nztota = acsr%get_nzeros()
+    call psb_loc_to_glob(ione,lfrst,desc_a,info)
+    
    ! Fix the entries to call C-base SuperLU
-    call psb_loc_to_glob(1,ifrst,desc_a,info) 
-    call psb_loc_to_glob(nrow_a,ibcheck,desc_a,info) 
-    call psb_loc_to_glob(acsr%ja(1:nztota),desc_a,info,iact='I')
+    call psb_realloc(nztota,gja,info)
+    call psb_loc_to_glob(acsr%ja(1:nztota),gja(1:nztota), desc_a, info, iact='I')
+    acsr%ja(1:nztota) = gja(1:nztota)
    acsr%ja(:)  = acsr%ja(:)  - 1
    acsr%irp(:) = acsr%irp(:) - 1
-    ifrst = ifrst - 1
+    ifrst = lfrst - 1
+    
    info = amg_dsludist_fact(nglob,nrow_a,nztota,ifrst,&
         & acsr%val,acsr%irp,acsr%ja,sv%lufactors,&
         & npr,npc)
@@ -318,7 +338,6 @@ contains
    end if
    
    call acsr%free()
-    call atmp%free()

    if (debug_level >= psb_debug_outer_) &
         & write(debug_unit,*) me,' ',trim(name),' end'
@@ -126,7 +126,7 @@ module amg_s_base_aggregator_mod
           & psb_s_coo_sparse_mat, amg_sml_parms, psb_spk_, psb_ipk_, psb_lpk_
      implicit none
      type(psb_s_csr_sparse_mat), intent(inout) :: a_csr
-      type(psb_desc_type), intent(in)            :: desc_a
+      type(psb_desc_type), intent(inout)         :: desc_a
      integer(psb_lpk_), intent(inout)           :: nlaggr(:)
      type(amg_sml_parms), intent(inout)         :: parms 
      type(psb_s_coo_sparse_mat), intent(inout) :: coo_prol, coo_restr
@@ -144,7 +144,7 @@ module amg_s_base_aggregator_mod
           & psb_s_coo_sparse_mat, amg_sml_parms, psb_spk_, psb_ipk_, psb_lpk_
      implicit none
      type(psb_s_csr_sparse_mat), intent(inout) :: a_csr
-      type(psb_desc_type), intent(in)            :: desc_a
+      type(psb_desc_type), intent(inout)         :: desc_a
      integer(psb_lpk_), intent(inout)           :: nlaggr(:)
      type(amg_sml_parms), intent(inout)         :: parms 
      type(psb_s_coo_sparse_mat), intent(inout) :: coo_prol, coo_restr
@@ -68,7 +68,7 @@
 !    ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
 !    POSSIBILITY OF SUCH DAMAGE.
 !
-module smatchboxp_mod
+module amg_s_matchboxp_mod

  use iso_c_binding
  use psb_base_cbind_mod
@@ -94,33 +94,25 @@ module smatchboxp_mod
    end subroutine sMatchBoxPC
  end interface MatchBoxPC

-  interface i_aggr_assign
-    module procedure i_saggr_assign
-  end interface i_aggr_assign
+  interface amg_i_aggr_assign
+    module procedure amg_i_s_aggr_assign
+  end interface amg_i_aggr_assign

-  interface build_matching
-    module procedure sbuild_matching
-  end interface build_matching
+  interface amg_par_build_matching
+    module procedure amg_s_par_build_matching
+  end interface amg_par_build_matching

-  interface build_ahat
-    module procedure sbuild_ahat
-  end interface build_ahat
+  interface amg_par_build_ahat
+    module procedure amg_s_par_build_ahat
+  end interface amg_par_build_ahat

-  interface psb_gtranspose
-    module procedure psb_sgtranspose
-  end interface psb_gtranspose
-
-  interface psb_htranspose
-    module procedure psb_shtranspose
-  end interface psb_htranspose
-
-  interface PMatchBox
-    module procedure sPMatchBox
-  end interface PMatchBox
+  interface amg_PMatchBox
+    module procedure amg_s_PMatchBox
+  end interface amg_PMatchBox

 contains

-  subroutine smatchboxp_build_prol(w,a,desc_a,ilaggr,nlaggr,prol,info,&
+  subroutine amg_s_matchboxp_build_prol(w,a,desc_a,ilaggr,nlaggr,prol,info,&
       & symmetrize,reproducible,display_inp, display_out, print_out)
    use psb_base_mod
    use psb_util_mod
@@ -213,7 +205,7 @@ contains
    end if
    if (do_timings) call psb_toc(idx_phase1)
    if (do_timings) call psb_tic(idx_bldmtc)
-    call build_matching(w,a,desc_a,mate,info,display_inp=display_inp,symmetrize=symmetrize)
+    call amg_par_build_matching(w,a,desc_a,mate,info,display_inp=display_inp,symmetrize=symmetrize)
    if (do_timings) call psb_toc(idx_bldmtc)
    if (debug) write(0,*) iam,' buildprol from buildmatching:',&
         & info
@@ -311,7 +303,7 @@ contains
                  ! Should be a symmetric function.
                  !
                  call desc_a%indxmap%qry_halo_owner(idx,iown,info)
-                  ip    = i_aggr_assign(iam, iown, kg, idxg, wk, widx, nrmagg)
+                  ip    = amg_i_aggr_assign(iam, iown, kg, idxg, wk, widx, nrmagg)
                  if (iam == ip) then
                    nlaggr(iam) = nlaggr(iam) + 1
                    ilaggr(k)   = nlaggr(iam)
@@ -513,9 +505,9 @@ contains
      write(0,*) iam,' :  error from Matching: ',info
    end if

-  end subroutine smatchboxp_build_prol
+  end subroutine amg_s_matchboxp_build_prol

-  function i_saggr_assign(iam, iown, kg, idxg, wk, widx, nrmagg) &
+  function amg_i_s_aggr_assign(iam, iown, kg, idxg, wk, widx, nrmagg) &
       & result(iproc)
    !
    ! How to break ties? This
@@ -557,10 +549,10 @@ contains
        iproc = iown
      end if
    end if
-  end function i_saggr_assign
+  end function amg_i_s_aggr_assign


-  subroutine sbuild_matching(w,a,desc_a,mate,info,display_inp, symmetrize)
+  subroutine amg_s_par_build_matching(w,a,desc_a,mate,info,display_inp, symmetrize)
    use psb_base_mod
    use psb_util_mod
    use iso_c_binding
@@ -609,7 +601,7 @@ contains
      if (iam == 0) write(0,*)' Into build_ahat:'
    end if
    if (do_timings) call psb_tic(idx_bldahat)
-    call build_ahat(w,a,ahatnd,desc_a,info,symmetrize=symmetrize)
+    call amg_par_build_ahat(w,a,ahatnd,desc_a,info,symmetrize=symmetrize)
    if (do_timings) call psb_toc(idx_bldahat)
    if (info /= 0) then
      write(0,*) 'Error from build_ahat ', info
@@ -700,7 +692,7 @@ contains
      !
      if (debug) write(0,*) iam,' buildmatching into PMatchBox:'
      if (do_timings) call psb_tic(idx_cmboxp)
-      call PMatchBox(nr,nz,vlptr,vlind,ewght,&
+      call amg_PMatchBox(nr,nz,vlptr,vlind,ewght,&
           & vnl, mate, iam, np,ictxt,&
           & msgis,msgas,msgprc,ph0t,ph1t,ph2t,ph1crd,ph2crd,info,display_inp)
      if (do_timings) call psb_toc(idx_cmboxp)
@@ -764,9 +756,9 @@ contains
      val(1:n) = tmp(1:n)
    end subroutine fix_order

-  end subroutine sbuild_matching
+  end subroutine amg_s_par_build_matching

-  subroutine sbuild_ahat(w,a,ahat,desc_a,info,symmetrize)
+  subroutine amg_s_par_build_ahat(w,a,ahat,desc_a,info,symmetrize)
    use psb_base_mod
    implicit none
    real(psb_spk_), intent(in)           :: w(:)
@@ -1002,301 +994,9 @@ contains
      end block
    end if

-  end subroutine sbuild_ahat
+  end subroutine amg_s_par_build_ahat

-  subroutine psb_sgtranspose(ain,aout,desc_a,info)
-    use psb_base_mod
-    implicit none
-    type(psb_lsspmat_type), intent(in)  :: ain
-    type(psb_lsspmat_type), intent(out) :: aout
-    type(psb_desc_type)           :: desc_a
-    integer(psb_ipk_), intent(out) :: info
-
-    !
-    ! BEWARE: This routine works under the assumption
-    ! that the same DESC_A works for both A and A^T, which
-    ! essentially means that A has a symmetric pattern.
-    !
-    type(psb_lsspmat_type) :: atmp, ahalo, aglb
-    type(psb_ls_coo_sparse_mat) :: tmpcoo
-    type(psb_ls_csr_sparse_mat) :: tmpcsr
-    type(psb_ctxt_type) :: ictxt
-    integer(psb_ipk_) :: me, np
-    integer(psb_lpk_) :: i, j, k, nrow, ncol
-    integer(psb_lpk_), allocatable :: ilv(:)
-    character(len=80) :: aname
-    logical, parameter :: debug=.false., dump=.false., debug_sync=.false.
-
-    ictxt = desc_a%get_context()
-    call psb_info(ictxt,me,np)
-
-    nrow = desc_a%get_local_rows()
-    ncol = desc_a%get_local_cols()
-    if (debug_sync) then
-      call psb_barrier(ictxt)
-      if (me == 0) write(0,*) 'Start gtranspose '
-    end if
-    call ain%cscnv(tmpcsr,info)
-
-    if (debug) then
-      ilv = [(i,i=1,ncol)]
-      call desc_a%l2gip(ilv,info,owned=.false.)
-      write(aname,'(a,i3.3,a)') 'atmp-preh-',me,'.mtx'
-      call ain%print(fname=aname,head='atmp before haloTest ',iv=ilv)
-    end if
-    if (dump) then
-      call ain%cscnv(atmp,info)
-      call psb_gather(aglb,atmp,desc_a,info)
-      if (me==psb_root_) then
-        write(aname,'(a,i3.3,a)') 'aglob-prehalo.mtx'
-        call aglb%print(fname=aname,head='Test ')
-      end if
-    end if
-
-    !call psb_loc_to_glob(tmpcsr%ja,desc_a,info)
-    call atmp%mv_from(tmpcsr)
-
-    if (debug) then
-      write(aname,'(a,i3.3,a)') 'tmpcsr-',me,'.mtx'
-      call atmp%print(fname=aname,head='tmpcsr ',iv=ilv)
-      !call psb_set_debug_level(9999)
-    end if
-
-    ! FIXME THIS NEEDS REWORKING
-    if (debug) write(0,*) me,' Gtranspose into sphalo :',atmp%get_nrows(),atmp%get_ncols()
-    call psb_sphalo(atmp,desc_a,ahalo,info,rowscale=.true.)
-    if (debug) write(0,*) me,' Gtranspose from sphalo :',ahalo%get_nrows(),ahalo%get_ncols()
-    if (info == psb_success_) call psb_rwextd(ncol,atmp,info,b=ahalo)
-
-    if (debug) then
-      write(aname,'(a,i3.3,a)') 'ahalo-',me,'.mtx'
-      call ahalo%print(fname=aname,head='ahalo after haloTest ',iv=ilv)
-      write(aname,'(a,i3.3,a)') 'atmp-h-',me,'.mtx'
-      call atmp%print(fname=aname,head='atmp after haloTest ',iv=ilv)
-    end if
-
-    if (info == psb_success_) call ahalo%free()
-
-    call atmp%cp_to(tmpcoo)
-    call tmpcoo%transp()
-    !call psb_glob_to_loc(tmpcoo%ia,desc_a,info,iact='I')
-    if (debug) write(0,*) 'Before cleanup:',tmpcoo%get_nzeros()
-
-    j = 0
-    do k=1, tmpcoo%get_nzeros()
-      if ((tmpcoo%ia(k) > 0).and.(tmpcoo%ja(k)>0)) then
-        j = j+1
-        tmpcoo%ia(j)  = tmpcoo%ia(k)
-        tmpcoo%ja(j)  = tmpcoo%ja(k)
-        tmpcoo%val(j) = tmpcoo%val(k)
-      end if
-    end do
-    call tmpcoo%set_nzeros(j)
-
-    if (debug) write(0,*) 'After cleanup:',tmpcoo%get_nzeros()
-
-    call ahalo%mv_from(tmpcoo)
-    if (dump) then
-      call psb_gather(aglb,ahalo,desc_a,info)
-      if (me==psb_root_) then
-        write(aname,'(a,i3.3,a)') 'atran-preclip.mtx'
-        call aglb%print(fname=aname,head='Test ')
-      end if
-    end if
-
-
-    call ahalo%csclip(aout,info,imax=nrow)
-
-    if (debug) write(0,*) 'After clip:',aout%get_nzeros()
-
-    if (debug_sync) then
-      call psb_barrier(ictxt)
-      if (me == 0) write(0,*) 'End gtranspose '
-    end if
-    !call aout%cscnv(info,type='csr')
-
-    if (dump) then
-      write(aname,'(a,i3.3,a)') 'atran-',me,'.mtx'
-      call aout%print(fname=aname,head='atrans ',iv=ilv)
-      call psb_gather(aglb,aout,desc_a,info)
-      if (me==psb_root_) then
-        write(aname,'(a,i3.3,a)') 'atran.mtx'
-        call aglb%print(fname=aname,head='Test ')
-      end if
-    end if
-
-  end subroutine psb_sgtranspose
-
-  subroutine psb_shtranspose(ain,aout,desc_a,info)
-    use psb_base_mod
-    implicit none
-    type(psb_lsspmat_type), intent(in)  :: ain
-    type(psb_lsspmat_type), intent(out) :: aout
-    type(psb_desc_type)           :: desc_a
-    integer(psb_ipk_), intent(out) :: info
-
-    !
-    ! BEWARE: This routine works under the assumption
-    ! that the same DESC_A works for both A and A^T, which
-    ! essentially means that A has a symmetric pattern.
-    !
-    type(psb_lsspmat_type) :: atmp, ahalo, aglb
-    type(psb_ls_coo_sparse_mat) :: tmpcoo, tmpc1, tmpc2, tmpch
-    type(psb_ls_csr_sparse_mat) :: tmpcsr
-    integer(psb_ipk_) :: nz1, nz2, nzh, nz
-    type(psb_ctxt_type) :: ictxt
-    integer(psb_ipk_) :: me, np
-    integer(psb_lpk_) :: i, j, k, nrow, ncol, nlz
-    integer(psb_lpk_), allocatable :: ilv(:)
-    character(len=80) :: aname
-    logical, parameter :: debug=.false., dump=.false., debug_sync=.false.
-
-    ictxt = desc_a%get_context()
-    call psb_info(ictxt,me,np)
-
-    nrow = desc_a%get_local_rows()
-    ncol = desc_a%get_local_cols()
-    if (debug_sync) then
-      call psb_barrier(ictxt)
-      if (me == 0) write(0,*) 'Start htranspose '
-    end if
-    call ain%cscnv(tmpcsr,info)
-
-    if (debug) then
-      ilv = [(i,i=1,ncol)]
-      call desc_a%l2gip(ilv,info,owned=.false.)
-      write(aname,'(a,i3.3,a)') 'atmp-preh-',me,'.mtx'
-      call ain%print(fname=aname,head='atmp before haloTest ',iv=ilv)
-    end if
-    if (dump) then
-      call ain%cscnv(atmp,info)
-      call psb_gather(aglb,atmp,desc_a,info)
-      if (me==psb_root_) then
-        write(aname,'(a,i3.3,a)') 'aglob-prehalo.mtx'
-        call aglb%print(fname=aname,head='Test ')
-      end if
-    end if
-
-    !call psb_loc_to_glob(tmpcsr%ja,desc_a,info)
-    call atmp%mv_from(tmpcsr)
-
-    if (debug) then
-      write(aname,'(a,i3.3,a)') 'tmpcsr-',me,'.mtx'
-      call atmp%print(fname=aname,head='tmpcsr ',iv=ilv)
-      !call psb_set_debug_level(9999)
-    end if
-
-    ! FIXME THIS NEEDS REWORKING
-    if (debug) write(0,*) me,' Htranspose into sphalo :',atmp%get_nrows(),atmp%get_ncols()
-    if (.true.) then
-      call psb_sphalo(atmp,desc_a,ahalo,info, outfmt='coo  ')
-      call atmp%mv_to(tmpc1)
-      call ahalo%mv_to(tmpch)
-      nz1 = tmpc1%get_nzeros()
-      call psb_loc_to_glob(tmpc1%ia(1:nz1),desc_a,info,iact='I')
-      call psb_loc_to_glob(tmpc1%ja(1:nz1),desc_a,info,iact='I')
-      nzh = tmpch%get_nzeros()
-      call psb_loc_to_glob(tmpch%ia(1:nzh),desc_a,info,iact='I')
-      call psb_loc_to_glob(tmpch%ja(1:nzh),desc_a,info,iact='I')
-      nlz = nz1+nzh
-      call tmpcoo%allocate(ncol,ncol,nlz)
-      tmpcoo%ia(1:nz1) = tmpc1%ia(1:nz1)
-      tmpcoo%ja(1:nz1) = tmpc1%ja(1:nz1)
-      tmpcoo%val(1:nz1) = tmpc1%val(1:nz1)
-      tmpcoo%ia(nz1+1:nz1+nzh) = tmpch%ia(1:nzh)
-      tmpcoo%ja(nz1+1:nz1+nzh) = tmpch%ja(1:nzh)
-      tmpcoo%val(nz1+1:nz1+nzh) = tmpch%val(1:nzh)
-      call tmpcoo%set_nzeros(nlz)
-      call tmpcoo%transp()
-      nz = tmpcoo%get_nzeros()
-      call psb_glob_to_loc(tmpcoo%ia(1:nz),desc_a,info,iact='I')
-      call psb_glob_to_loc(tmpcoo%ja(1:nz),desc_a,info,iact='I')
-      if (.true.) then
-        call tmpcoo%clean_negidx(info)
-      else
-        j = 0
-        do k=1, tmpcoo%get_nzeros()
-          if ((tmpcoo%ia(k) > 0).and.(tmpcoo%ja(k)>0)) then
-            j = j+1
-            tmpcoo%ia(j)  = tmpcoo%ia(k)
-            tmpcoo%ja(j)  = tmpcoo%ja(k)
-            tmpcoo%val(j) = tmpcoo%val(k)
-          end if
-        end do
-        call tmpcoo%set_nzeros(j)
-      end if
-      call ahalo%mv_from(tmpcoo)
-      call ahalo%csclip(aout,info,imax=nrow)
-
-    else
-      call psb_sphalo(atmp,desc_a,ahalo,info, rowscale=.true.)
-      if (debug) write(0,*) me,' Htranspose from sphalo :',ahalo%get_nrows(),ahalo%get_ncols()
-      if (info == psb_success_) call psb_rwextd(ncol,atmp,info,b=ahalo)
-
-      if (debug) then
-        write(aname,'(a,i3.3,a)') 'ahalo-',me,'.mtx'
-        call ahalo%print(fname=aname,head='ahalo after haloTest ',iv=ilv)
-        write(aname,'(a,i3.3,a)') 'atmp-h-',me,'.mtx'
-        call atmp%print(fname=aname,head='atmp after haloTest ',iv=ilv)
-      end if
-
-      if (info == psb_success_) call ahalo%free()
-
-      call atmp%cp_to(tmpcoo)
-      call tmpcoo%transp()
-      if (debug) write(0,*) 'Before cleanup:',tmpcoo%get_nzeros()
-      if (.true.) then
-        call tmpcoo%clean_negidx(info)
-      else
-
-        j = 0
-        do k=1, tmpcoo%get_nzeros()
-          if ((tmpcoo%ia(k) > 0).and.(tmpcoo%ja(k)>0)) then
-            j = j+1
-            tmpcoo%ia(j)  = tmpcoo%ia(k)
-            tmpcoo%ja(j)  = tmpcoo%ja(k)
-            tmpcoo%val(j) = tmpcoo%val(k)
-          end if
-        end do
-        call tmpcoo%set_nzeros(j)
-      end if
-
-      if (debug) write(0,*) 'After cleanup:',tmpcoo%get_nzeros()
-
-      call ahalo%mv_from(tmpcoo)
-      if (dump) then
-        call psb_gather(aglb,ahalo,desc_a,info)
-        if (me==psb_root_) then
-          write(aname,'(a,i3.3,a)') 'atran-preclip.mtx'
-          call aglb%print(fname=aname,head='Test ')
-        end if
-      end if
-
-
-      call ahalo%csclip(aout,info,imax=nrow)
-    end if
-
-    if (debug) write(0,*) 'After clip:',aout%get_nzeros()
-
-    if (debug_sync) then
-      call psb_barrier(ictxt)
-      if (me == 0) write(0,*) 'End htranspose '
-    end if
-    !call aout%cscnv(info,type='csr')
-
-    if (dump) then
-      write(aname,'(a,i3.3,a)') 'atran-',me,'.mtx'
-      call aout%print(fname=aname,head='atrans ',iv=ilv)
-      call psb_gather(aglb,aout,desc_a,info)
-      if (me==psb_root_) then
-        write(aname,'(a,i3.3,a)') 'atran.mtx'
-        call aglb%print(fname=aname,head='Test ')
-      end if
-    end if
-
-  end subroutine psb_shtranspose
-
-  subroutine sPMatchBox(nlver,nledge,verlocptr,verlocind,edgelocweight,&
+  subroutine amg_s_PMatchBox(nlver,nledge,verlocptr,verlocind,edgelocweight,&
       & verdistance, mate, myrank, numprocs, ictxt,&
       & msgindsent,msgactualsent,msgpercent,&
       & ph0_time, ph1_time, ph2_time, ph1_card, ph2_card,info,display_inp)
@@ -1431,6 +1131,6 @@ contains
    end if
    where(mate>=0) mate = mate + 1

-  end subroutine sPMatchBox
+  end subroutine amg_s_PMatchBox

-end module smatchboxp_mod
+end module amg_s_matchboxp_mod
@@ -118,7 +118,7 @@

 module amg_s_parmatch_aggregator_mod
  use amg_s_base_aggregator_mod
-  use smatchboxp_mod
+  use amg_s_matchboxp_mod
 #if defined(SERIAL_MPI)
  type, extends(amg_s_base_aggregator_type) :: amg_s_parmatch_aggregator_type
  end type amg_s_parmatch_aggregator_type
@@ -132,8 +132,6 @@ module amg_s_parmatch_aggregator_mod
    type(psb_sspmat_type), allocatable  :: prol, restr
    type(psb_sspmat_type), allocatable  :: ac, base_a, rwa
    type(psb_desc_type), allocatable    :: desc_ac, desc_ax, base_desc, rwdesc
-    integer(psb_ipk_) :: max_csize
-    integer(psb_ipk_) :: max_nlevels
    logical           :: reproducible_matching = .false.
    logical           :: need_symmetrize       = .false.
    logical           :: unsmoothed_hierarchy  = .true.
@@ -143,18 +141,18 @@ module amg_s_parmatch_aggregator_mod
    procedure, pass(ag) :: mat_asb      => amg_s_parmatch_aggregator_mat_asb
    procedure, pass(ag) :: inner_mat_asb      => amg_s_parmatch_aggregator_inner_mat_asb
    procedure, pass(ag) :: bld_map      => amg_s_parmatch_aggregator_bld_map
-    procedure, pass(ag) :: csetc        => s_parmatch_aggr_csetc
-    procedure, pass(ag) :: cseti        => s_parmatch_aggr_cseti
-    procedure, pass(ag) :: default      => s_parmatch_aggr_set_default
-    procedure, pass(ag) :: sizeof       => s_parmatch_aggregator_sizeof
-    procedure, pass(ag) :: update_next  => s_parmatch_aggregator_update_next
-    procedure, pass(ag) :: bld_wnxt     => s_parmatch_bld_wnxt
-    procedure, pass(ag) :: bld_default_w    => s_bld_default_w
-    procedure, pass(ag) :: set_c_default_w  => s_set_prm_c_default_w
-    procedure, pass(ag) :: descr        => s_parmatch_aggregator_descr
-    procedure, pass(ag) :: clone        => s_parmatch_aggregator_clone
-    procedure, pass(ag) :: free         => s_parmatch_aggregator_free
-    procedure, nopass   :: fmt          => s_parmatch_aggregator_fmt
+    procedure, pass(ag) :: csetc        => amg_s_parmatch_aggr_csetc
+    procedure, pass(ag) :: cseti        => amg_s_parmatch_aggr_cseti
+    procedure, pass(ag) :: default      => amg_s_parmatch_aggr_set_default
+    procedure, pass(ag) :: sizeof       => amg_s_parmatch_aggregator_sizeof
+    procedure, pass(ag) :: update_next  => amg_s_parmatch_aggregator_update_next
+    procedure, pass(ag) :: bld_wnxt     => amg_s_parmatch_bld_wnxt
+    procedure, pass(ag) :: bld_default_w    => amg_s_bld_default_w
+    procedure, pass(ag) :: set_c_default_w  => amg_s_set_prm_c_default_w
+    procedure, pass(ag) :: descr        => amg_s_parmatch_aggregator_descr
+    procedure, pass(ag) :: clone        => amg_s_parmatch_aggregator_clone
+    procedure, pass(ag) :: free         => amg_s_parmatch_aggregator_free
+    procedure, nopass   :: fmt          => amg_s_parmatch_aggregator_fmt
    procedure, nopass   :: xt_desc      => amg_s_parmatch_aggregator_xt_desc
  end type amg_s_parmatch_aggregator_type

@@ -168,7 +166,7 @@ module amg_s_parmatch_aggregator_mod
      type(amg_sml_parms), intent(inout)   :: parms
      type(amg_saggr_data), intent(in)     :: ag_data
      type(psb_sspmat_type), intent(inout) :: a
-      type(psb_desc_type), intent(inout)   :: desc_a
+      type(psb_desc_type), intent(inout)     :: desc_a
      integer(psb_lpk_), allocatable, intent(out) :: ilaggr(:),nlaggr(:)
      type(psb_lsspmat_type), intent(out)  :: t_prol
      integer(psb_ipk_), intent(out)      :: info
@@ -235,7 +233,7 @@ module amg_s_parmatch_aggregator_mod
           & psb_lsspmat_type, psb_dpk_, psb_ipk_, psb_lpk_, amg_sml_parms, amg_saggr_data
      implicit none
      type(psb_sspmat_type), intent(in)       :: a
-      type(psb_desc_type), intent(in)         :: desc_a
+      type(psb_desc_type), intent(inout)       :: desc_a
      integer(psb_lpk_), intent(inout)        :: ilaggr(:), nlaggr(:)
      type(amg_sml_parms), intent(inout)      :: parms
      type(psb_lsspmat_type), intent(inout)   :: t_prol
@@ -257,7 +255,7 @@ module amg_s_parmatch_aggregator_mod
      integer(psb_lpk_), intent(inout)        :: ilaggr(:), nlaggr(:)
      type(amg_sml_parms), intent(inout)      :: parms
      type(psb_lsspmat_type), intent(inout)   :: t_prol
-      type(psb_sspmat_type), intent(out)      :: op_prol,ac, op_restr
+      type(psb_sspmat_type), intent(inout)      :: op_prol,ac, op_restr
      type(psb_desc_type), intent(inout)    :: desc_ac
      integer(psb_ipk_), intent(out)          :: info
    end subroutine amg_s_parmatch_unsmth_bld
@@ -275,7 +273,7 @@ module amg_s_parmatch_aggregator_mod
      integer(psb_lpk_), intent(inout)        :: ilaggr(:), nlaggr(:)
      type(amg_sml_parms), intent(inout)      :: parms
      type(psb_lsspmat_type), intent(inout)   :: t_prol
-      type(psb_sspmat_type), intent(out)      :: op_prol,ac, op_restr
+      type(psb_sspmat_type), intent(inout)    :: op_prol,ac, op_restr
      type(psb_desc_type), intent(inout)    :: desc_ac
      integer(psb_ipk_), intent(out)          :: info
    end subroutine amg_s_parmatch_smth_bld
@@ -288,11 +286,11 @@ module amg_s_parmatch_aggregator_mod
           & psb_lsspmat_type, psb_dpk_, psb_ipk_, psb_lpk_, amg_sml_parms, amg_saggr_data
      implicit none
      type(psb_sspmat_type), intent(inout)    :: a
-      type(psb_desc_type), intent(in)         :: desc_a
+      type(psb_desc_type), intent(inout)       :: desc_a
      integer(psb_lpk_), intent(inout)        :: ilaggr(:), nlaggr(:)
      type(amg_sml_parms), intent(inout)      :: parms
      type(psb_lsspmat_type), intent(inout)   :: t_prol
-      type(psb_sspmat_type), intent(out)      :: op_prol,ac, op_restr
+      type(psb_sspmat_type), intent(inout)    :: op_prol,ac, op_restr
      type(psb_desc_type), intent(out)          :: desc_ac
      integer(psb_ipk_), intent(out)          :: info
    end subroutine amg_s_parmatch_spmm_bld_ov
@@ -306,11 +304,11 @@ module amg_s_parmatch_aggregator_mod
           & psb_s_csr_sparse_mat, psb_ls_csr_sparse_mat
      implicit none
      type(psb_s_csr_sparse_mat), intent(inout) :: a
-      type(psb_desc_type), intent(in)           :: desc_a
+      type(psb_desc_type), intent(inout)          :: desc_a
      integer(psb_lpk_), intent(inout)          :: ilaggr(:), nlaggr(:)
      type(amg_sml_parms), intent(inout)        :: parms
      type(psb_lsspmat_type), intent(inout)     :: t_prol
-      type(psb_sspmat_type), intent(out)        :: op_prol,ac, op_restr
+      type(psb_sspmat_type), intent(inout)      :: op_prol,ac, op_restr
      type(psb_desc_type), intent(out)          :: desc_ac
      integer(psb_ipk_), intent(out)            :: info
    end subroutine amg_s_parmatch_spmm_bld_inner
@@ -320,7 +318,7 @@ module amg_s_parmatch_aggregator_mod

 contains

-  subroutine s_bld_default_w(ag,nr)
+  subroutine amg_s_bld_default_w(ag,nr)
    use psb_realloc_mod
    implicit none
    class(amg_s_parmatch_aggregator_type), target, intent(inout) :: ag
@@ -330,9 +328,9 @@ contains
    if (info /= psb_success_) return
    ag%w = done
    !call ag%set_c_default_w()
-  end subroutine s_bld_default_w
+  end subroutine amg_s_bld_default_w

-  subroutine s_set_prm_c_default_w(ag)
+  subroutine amg_s_set_prm_c_default_w(ag)
    use psb_realloc_mod
    use iso_c_binding
    implicit none
@@ -342,9 +340,9 @@ contains
    !write(0,*) 'prm_c_deafult_w '
    call psb_safe_ab_cpy(ag%w,ag%w_nxt,info)

-  end subroutine s_set_prm_c_default_w
+  end subroutine amg_s_set_prm_c_default_w

-  subroutine s_parmatch_bld_wnxt(ag,ilaggr,valaggr,nx)
+  subroutine amg_s_parmatch_bld_wnxt(ag,ilaggr,valaggr,nx)
    use psb_realloc_mod
    implicit none
    class(amg_s_parmatch_aggregator_type), target, intent(inout) :: ag
@@ -358,14 +356,14 @@ contains
    !write(0,*) 'Executing bld_wnxt ',nx
    call psb_realloc(nx,ag%w_nxt,info)

-  end subroutine s_parmatch_bld_wnxt
+  end subroutine amg_s_parmatch_bld_wnxt

-  function s_parmatch_aggregator_fmt() result(val)
+  function amg_s_parmatch_aggregator_fmt() result(val)
    implicit none
    character(len=32)  :: val

    val = "Parallel Matching aggregation"
-  end function s_parmatch_aggregator_fmt
+  end function amg_s_parmatch_aggregator_fmt

  function amg_s_parmatch_aggregator_xt_desc() result(val)
    implicit none
@@ -374,7 +372,7 @@ contains
    val = .true.
  end function amg_s_parmatch_aggregator_xt_desc

-  function s_parmatch_aggregator_sizeof(ag) result(val)
+  function amg_s_parmatch_aggregator_sizeof(ag) result(val)
    use psb_realloc_mod
    implicit none
    class(amg_s_parmatch_aggregator_type), intent(in)  :: ag
@@ -390,9 +388,9 @@ contains
    if (allocated(ag%base_desc)) val = val + ag%base_desc%sizeof()
    if (allocated(ag%desc_ax)) val = val + ag%desc_ax%sizeof()

-  end function s_parmatch_aggregator_sizeof
+  end function amg_s_parmatch_aggregator_sizeof

-  subroutine  s_parmatch_aggregator_descr(ag,parms,iout,info)
+  subroutine  amg_s_parmatch_aggregator_descr(ag,parms,iout,info)
    implicit none
    class(amg_s_parmatch_aggregator_type), intent(in) :: ag
    type(amg_sml_parms), intent(in)   :: parms
@@ -406,7 +404,7 @@ contains
    call parms%mldescr(iout,info)

    return
-  end subroutine s_parmatch_aggregator_descr
+  end subroutine amg_s_parmatch_aggregator_descr

  function is_legal_malg(alg) result(val)
    logical :: val
@@ -437,7 +435,7 @@ contains
  end function is_legal_nlevels


-  subroutine  s_parmatch_aggregator_update_next(ag,agnext,info)
+  subroutine  amg_s_parmatch_aggregator_update_next(ag,agnext,info)
    use psb_realloc_mod
    implicit none
    class(amg_s_parmatch_aggregator_type), target, intent(inout) :: ag
@@ -452,10 +450,10 @@ contains
           & agnext%matching_alg = ag%matching_alg
      if (.not.is_legal_nsweeps(agnext%n_sweeps))&
           & agnext%n_sweeps     = ag%n_sweeps
-      if (.not.is_legal_csize(agnext%max_csize))&
-           & agnext%max_csize    = ag%max_csize
-      if (.not.is_legal_nlevels(agnext%max_nlevels))&
-           & agnext%max_nlevels  = ag%max_nlevels
+!!$      if (.not.is_legal_csize(agnext%max_csize))&
+!!$           & agnext%max_csize    = ag%max_csize
+!!$      if (.not.is_legal_nlevels(agnext%max_nlevels))&
+!!$           & agnext%max_nlevels  = ag%max_nlevels
      ! Is this going to generate shallow copies/memory leaks/double frees?
      ! To be investigated further.
      call psb_safe_ab_cpy(ag%w_nxt,agnext%w,info)
@@ -470,9 +468,9 @@ contains
      ! What should we do here?
    end select
    info = 0
-  end subroutine s_parmatch_aggregator_update_next
+  end subroutine amg_s_parmatch_aggregator_update_next

-  subroutine s_parmatch_aggr_csetc(ag,what,val,info,idx)
+  subroutine amg_s_parmatch_aggr_csetc(ag,what,val,info,idx)

    Implicit None

@@ -514,9 +512,9 @@ contains
      ! Do nothing
    end select
    return
-  end subroutine s_parmatch_aggr_csetc
+  end subroutine amg_s_parmatch_aggr_csetc

-  subroutine s_parmatch_aggr_cseti(ag,what,val,info,idx)
+  subroutine amg_s_parmatch_aggr_cseti(ag,what,val,info,idx)

    Implicit None

@@ -540,10 +538,6 @@ contains
    case('AGGR_SIZE')
      ag%orig_aggr_size = val
      ag%n_sweeps=max(1,ceiling(log(val*1.0)/log(2.0)))
-    case('PRMC_MAX_CSIZE')
-      ag%max_csize=val
-    case('PRMC_MAX_NLEVELS')
-      ag%max_nlevels=val
    case('PRMC_W_SIZE')
      call ag%bld_default_w(val)
    case('PRMC_REPRODUCIBLE_MATCHING')
@@ -556,9 +550,9 @@ contains
      ! Do nothing
    end select
    return
-  end subroutine s_parmatch_aggr_cseti
+  end subroutine amg_s_parmatch_aggr_cseti

-  subroutine s_parmatch_aggr_set_default(ag)
+  subroutine amg_s_parmatch_aggr_set_default(ag)

    Implicit None

@@ -569,8 +563,8 @@ contains
    ag%matching_alg  = 0
    ag%n_sweeps      = 1
    ag%jacobi_sweeps = 0
-    ag%max_nlevels   = 36
-    ag%max_csize     = -1
+!!$    ag%max_nlevels   = 36
+!!$    ag%max_csize     = -1
    !
    ! Apparently BootCMatch works better
    ! by keeping all entries
@@ -579,9 +573,9 @@ contains

    return

-  end subroutine s_parmatch_aggr_set_default
+  end subroutine amg_s_parmatch_aggr_set_default

-  subroutine  s_parmatch_aggregator_free(ag,info)
+  subroutine  amg_s_parmatch_aggregator_free(ag,info)
    use iso_c_binding
    implicit none
    class(amg_s_parmatch_aggregator_type), intent(inout) :: ag
@@ -618,9 +612,9 @@ contains
      call ag%rwdesc%free(info); deallocate(ag%rwdesc,stat=info)
    end if

-  end subroutine s_parmatch_aggregator_free
+  end subroutine amg_s_parmatch_aggregator_free

-  subroutine  s_parmatch_aggregator_clone(ag,agnext,info)
+  subroutine  amg_s_parmatch_aggregator_clone(ag,agnext,info)
    implicit none
    class(amg_s_parmatch_aggregator_type), intent(inout) :: ag
    class(amg_s_base_aggregator_type), allocatable, intent(inout) :: agnext
@@ -640,7 +634,7 @@ contains
      ! Should never ever get here
      info = -1
    end select
-  end subroutine s_parmatch_aggregator_clone
+  end subroutine amg_s_parmatch_aggregator_clone

  subroutine  amg_s_parmatch_aggregator_bld_map(ag,desc_a,desc_ac,ilaggr,nlaggr,&
       & op_restr,op_prol,map,info)
@@ -126,7 +126,7 @@ module amg_z_base_aggregator_mod
           & psb_z_coo_sparse_mat, amg_dml_parms, psb_dpk_, psb_ipk_, psb_lpk_
      implicit none
      type(psb_z_csr_sparse_mat), intent(inout) :: a_csr
-      type(psb_desc_type), intent(in)            :: desc_a
+      type(psb_desc_type), intent(inout)         :: desc_a
      integer(psb_lpk_), intent(inout)           :: nlaggr(:)
      type(amg_dml_parms), intent(inout)         :: parms 
      type(psb_z_coo_sparse_mat), intent(inout) :: coo_prol, coo_restr
@@ -144,7 +144,7 @@ module amg_z_base_aggregator_mod
           & psb_z_coo_sparse_mat, amg_dml_parms, psb_dpk_, psb_ipk_, psb_lpk_
      implicit none
      type(psb_z_csr_sparse_mat), intent(inout) :: a_csr
-      type(psb_desc_type), intent(in)            :: desc_a
+      type(psb_desc_type), intent(inout)         :: desc_a
      integer(psb_lpk_), intent(inout)           :: nlaggr(:)
      type(amg_dml_parms), intent(inout)         :: parms 
      type(psb_z_coo_sparse_mat), intent(inout) :: coo_prol, coo_restr
@@ -52,7 +52,7 @@ module amg_z_sludist_solver
  use iso_c_binding
  use amg_z_base_solver_mod

-#if defined(LPK8) 
+#if (!defined(HAVE_SLUDIST_)) || defined(IPK8) 

  type, extends(amg_z_base_solver_type) :: amg_z_sludist_solver_type

@@ -270,11 +270,13 @@ contains
    ! Local variables
    type(psb_zspmat_type) :: atmp
    type(psb_z_csr_sparse_mat) :: acsr
-    integer :: n_row,n_col, nrow_a, nztota, nglob, nzt, npr, npc
-    integer :: ifrst, ibcheck
+    integer(psb_lpk_), allocatable :: gia(:), gja(:)
    type(psb_ctxt_type) :: ctxt
-    integer :: np,me,i, err_act, debug_unit, debug_level
-    character(len=20)  :: name='z_sludist_solver_bld', ch_err
+    integer(psb_lpk_) :: lfrst
+    integer(psb_ipk_) :: n_row,n_col, nrow_a, nztota, nglob, nzt, npr, npc
+    integer(psb_ipk_) :: ifrst, ibcheck
+    integer(psb_ipk_) :: np,me,i, err_act, debug_unit, debug_level
+    character(len=20) :: name='z_sludist_solver_bld', ch_err
    
    info=psb_success_
    call psb_erractionsave(err_act)
@@ -293,19 +295,36 @@ contains
    n_col  = desc_a%get_local_cols()
    nglob  = desc_a%get_global_rows()
    
-    call a%cscnv(atmp,info,type='coo')
+    !
+    ! Strategy here is as follows: because a call to SLUDIST
+    ! as a gobal solver is mostly done  at the coarsest level,
+    ! even if we start from a problem requiring 8 bytes, chances
+    ! are that the global size will be suitable for 4 bytes
+    ! anyway, so we hope for the best, and throw an error
+    ! if something goes wrong.
+    !
+    if (nglob > huge(1_psb_ipk_)) then
+      write(0,*) me,' ',trim(name),': Error: overflow of local indices '
+      info=psb_err_internal_error_
+      call psb_errpush(info,name)
+      goto 9999
+    end if
+    
+    call a%cscnv(atmp,info,type='csr')
+    ! This in case we are dealing with AS
    call psb_rwextd(n_row,atmp,info,b=b) 
-    call atmp%cscnv(info,type='csr',dupl=psb_dupl_add_)
    call atmp%mv_to(acsr)
    nrow_a = acsr%get_nrows()
    nztota = acsr%get_nzeros()
+    call psb_loc_to_glob(ione,lfrst,desc_a,info)
+    
    ! Fix the entries to call C-base SuperLU
-    call psb_loc_to_glob(1,ifrst,desc_a,info) 
-    call psb_loc_to_glob(nrow_a,ibcheck,desc_a,info) 
-    call psb_loc_to_glob(acsr%ja(1:nztota),desc_a,info,iact='I')
+    call psb_realloc(nztota,gja,info)
+    call psb_loc_to_glob(acsr%ja(1:nztota),gja(1:nztota), desc_a, info, iact='I')
+    acsr%ja(1:nztota) = gja(1:nztota)
    acsr%ja(:)  = acsr%ja(:)  - 1
    acsr%irp(:) = acsr%irp(:) - 1
-    ifrst = ifrst - 1
+    ifrst = lfrst - 1    
    info = amg_zsludist_fact(nglob,nrow_a,nztota,ifrst,&
         & acsr%val,acsr%irp,acsr%ja,sv%lufactors,&
         & npr,npc)
@@ -318,7 +337,6 @@ contains
    end if
    
    call acsr%free()
-    call atmp%free()

    if (debug_level >= psb_debug_outer_) &
         & write(debug_unit,*) me,' ',trim(name),' end'
@@ -83,8 +83,8 @@ subroutine  amg_c_dec_aggregator_build_tprol(ag,parms,ag_data,&
  class(amg_c_dec_aggregator_type), target, intent(inout) :: ag
  type(amg_sml_parms), intent(inout)  :: parms
  type(amg_saggr_data), intent(in)    :: ag_data
-  type(psb_cspmat_type), intent(in)   :: a
-  type(psb_desc_type), intent(in)     :: desc_a
+  type(psb_cspmat_type), intent(inout) :: a
+  type(psb_desc_type), intent(inout)    :: desc_a
  integer(psb_lpk_), allocatable, intent(out) :: ilaggr(:),nlaggr(:)
  type(psb_lcspmat_type), intent(out)  :: t_prol
  integer(psb_ipk_), intent(out)      :: info
@@ -86,8 +86,8 @@ subroutine  amg_c_symdec_aggregator_build_tprol(ag,parms,ag_data,&
  class(amg_c_symdec_aggregator_type), target, intent(inout) :: ag
  type(amg_sml_parms), intent(inout)  :: parms
  type(amg_saggr_data), intent(in)    :: ag_data
-  type(psb_cspmat_type), intent(in)   :: a
-  type(psb_desc_type), intent(in)     :: desc_a
+  type(psb_cspmat_type), intent(inout) :: a
+  type(psb_desc_type), intent(inout)   :: desc_a
  integer(psb_lpk_), allocatable, intent(out) :: ilaggr(:),nlaggr(:)
  type(psb_lcspmat_type), intent(out)  :: op_prol
  integer(psb_ipk_), intent(out)      :: info
@@ -105,7 +105,7 @@
 !
 !
 subroutine amg_caggrmat_minnrg_bld(a,desc_a,ilaggr,nlaggr,parms,&
-     & ac,desc_ac,op_prol,op_restr,info)
+     & ac,desc_ac,op_prol,op_restr,t_prol,info)
  use psb_base_mod
  use amg_base_prec_type
  use amg_c_inner_mod, amg_protect_name => amg_caggrmat_minnrg_bld
@@ -117,8 +117,8 @@ subroutine amg_caggrmat_minnrg_bld(a,desc_a,ilaggr,nlaggr,parms,&
  type(psb_desc_type), intent(inout)            :: desc_a
  integer(psb_lpk_), intent(inout)              :: ilaggr(:), nlaggr(:)
  type(amg_sml_parms), intent(inout)         :: parms 
-  type(psb_lcspmat_type), intent(inout)       :: op_prol
-  type(psb_lcspmat_type), intent(out)         :: ac,op_restr
+  type(psb_lcspmat_type), intent(inout)       :: t_prol
+  type(psb_cspmat_type), intent(inout)       :: op_prol, ac,op_restr
  type(psb_desc_type), intent(inout)            :: desc_ac
  integer(psb_ipk_), intent(out)                :: info

@@ -171,6 +171,8 @@ subroutine amg_caggrmat_minnrg_bld(a,desc_a,ilaggr,nlaggr,parms,&

  filter_mat = (parms%aggr_filter == amg_filter_mat_)

+  !NEEDS TO BE REWORKED !!
+  
  ! naggr: number of local aggregates
  ! nrow: local rows. 
  ! 
@@ -183,361 +185,361 @@ subroutine amg_caggrmat_minnrg_bld(a,desc_a,ilaggr,nlaggr,parms,&
    goto 9999      
  end if

-  ! Get the diagonal D
-  adiag =  a%get_diag(info)
-  if (info == psb_success_) &
-       & call psb_realloc(ncol,adiag,info)    
-  if (info == psb_success_) &
-       & call psb_halo(adiag,desc_a,info)
-  if (info == psb_success_) call a%cp_to_l(la)
-  if (info /= psb_success_) then
-    call psb_errpush(psb_err_from_subroutine_,name,a_err='sp_getdiag')
-    goto 9999
-  end if
-
-  do i=1,size(adiag)
-    if (adiag(i) /= czero) then
-      adinv(i) = cone / adiag(i)
-    else
-      adinv(i) = cone
-    end if
-  end do
-
-
-
-  ! 1. Allocate Ptilde in sparse matrix form 
-  call op_prol%mv_to(tmpcoo)
-  call ptilde%mv_from(tmpcoo)
-  call ptilde%cscnv(info,type='csr')
-
-  if (info == psb_success_) call la%cscnv(am3,info,type='csr',dupl=psb_dupl_add_)
-  if (info == psb_success_) call la%cscnv(da,info,type='csr',dupl=psb_dupl_add_)
-  if (info /= psb_success_) then
-    call psb_errpush(psb_err_from_subroutine_,name,a_err='spcnv')
-    goto 9999
-  end if
-  if (debug_level >= psb_debug_outer_) &
-       & write(debug_unit,*) me,' ',trim(name),&
-       & ' Initial copies done.'
-
-  call da%scal(adinv,info)
-
-  call psb_spspmm(da,ptilde,dap,info)
-
-  if(info /= psb_success_) then
-    call psb_errpush(psb_err_from_subroutine_,name,a_err='spspmm 1')
-    goto 9999
-  end if
-
-  call dap%clone(atmp,info)
-
-  call psb_sphalo(atmp,desc_a,am4,info,&
-       & colcnv=.false.,rowscale=.true.,outfmt='CSR  ')
-  if (info == psb_success_) call psb_rwextd(ncol,atmp,info,b=am4)      
-  if (info == psb_success_) call am4%free()
-
-  call psb_spspmm(da,atmp,dadap,info)
-  call atmp%free()
-
-  !  !$  write(0,*) 'Columns of AP',psb_sp_get_ncols(ap)
-  !  !$  write(0,*) 'Columns of ADAP',psb_sp_get_ncols(adap)
-  call dap%mv_to(csc_dap)
-  call dadap%mv_to(csc_dadap)
-
-  call csc_mat_col_prod(csc_dap,csc_dadap,omp,info)
-  call csc_mat_col_prod(csc_dadap,csc_dadap,oden,info)
-  call psb_sum(ctxt,omp)
-  call psb_sum(ctxt,oden)
-  ! !$  write(0,*) trim(name),' OMP :',omp
-  ! !$  write(0,*) trim(name),' ODEN:',oden
-
-  omp = omp/oden
-
-  ! !$  write(0,*) 'Check on output prolongator ',omp(1:min(size(omp),10))
-  if (debug_level >= psb_debug_outer_) &
-       & write(debug_unit,*) me,' ',trim(name),&
-       & 'Done NUMBMM 1'
-
-  call am3%mv_to(acsr3)
-  ! Compute omega_int
-  ommx = czero
-  do i=1, ncol
-    if (ilaggr(i) >0) then 
-      omi(i) = omp(ilaggr(i))
-    else
-      omi(i) = czero
-    end if
-    if(abs(omi(i)) .gt. abs(ommx)) ommx = omi(i)
-  end do
-  ! Compute omega_fine
-  do i=1, nrow
-    omf(i) = ommx
-    do j=acsr3%irp(i),acsr3%irp(i+1)-1
-      if(abs(omi(acsr3%ja(j))) .lt. abs(omf(i))) omf(i)=omi(acsr3%ja(j))
-    end do
-!!$    if(min(real(omf(i)),aimag(omf(i))) < szero) omf(i) = czero
-    if(psb_minreal(omf(i)) < szero) omf(i) = czero
-  end do
-
-  omf(1:nrow) = omf(1:nrow) * adinv(1:nrow)
-
-  if (filter_mat) then
-    !
-    ! Build the filtered matrix Af from A
-    ! 
-    call la%cscnv(acsrf,info,dupl=psb_dupl_add_)
-
-    do i=1,nrow
-      tmp = czero
-      jd  = -1 
-      do j=acsrf%irp(i),acsrf%irp(i+1)-1
-        if (acsrf%ja(j) == i) jd = j 
-        if (abs(acsrf%val(j)) < theta*sqrt(abs(adiag(i)*adiag(acsrf%ja(j))))) then
-          tmp=tmp+acsrf%val(j)
-          acsrf%val(j)=czero
-        endif
-      enddo
-      if (jd == -1) then 
-        write(0,*) 'Wrong input: we need the diagonal!!!!', i
-      else
-        acsrf%val(jd)=acsrf%val(jd)-tmp
-      end if
-    enddo
-    ! Take out zeroed terms 
-    call acsrf%clean_zeros(info)
-
-    !
-    ! Build the smoothed prolongator using the filtered matrix
-    ! 
-    do i=1,acsrf%get_nrows()
-      do j=acsrf%irp(i),acsrf%irp(i+1)-1
-        if (acsrf%ja(j) == i) then 
-          acsrf%val(j) = cone - omf(i)*acsrf%val(j) 
-        else
-          acsrf%val(j) = - omf(i)*acsrf%val(j) 
-        end if
-      end do
-    end do
-
-    if (debug_level >= psb_debug_outer_) &
-         & write(debug_unit,*) me,' ',trim(name),&
-         & 'Done gather, going for SYMBMM 1'
-
-    call af%mv_from(acsrf)
-    !
-    ! op_prol = (I-w*D*Af)Ptilde
-    ! Doing it this way means to consider diag(Af_i)
-    ! 
-    !
-    call psb_spspmm(af,ptilde,op_prol,info)
-    if (debug_level >= psb_debug_outer_) &
-         & write(debug_unit,*) me,' ',trim(name),&
-         & 'Done SPSPMM 1'
-  else
-    !
-    ! Build the smoothed prolongator using the original matrix
-    !
-    do i=1,acsr3%get_nrows()
-      do j=acsr3%irp(i),acsr3%irp(i+1)-1
-        if (acsr3%ja(j) == i) then 
-          acsr3%val(j) = cone - omf(i)*acsr3%val(j) 
-        else
-          acsr3%val(j) = - omf(i)*acsr3%val(j) 
-        end if
-      end do
-    end do
-
-    call am3%mv_from(acsr3)
-    if (debug_level >= psb_debug_outer_) &
-         & write(debug_unit,*) me,' ',trim(name),&
-         & 'Done gather, going for SYMBMM 1'
-    !
-    ! 
-    ! op_prol = (I-w*D*A)Ptilde
-    ! 
-    !
-    call psb_spspmm(am3,ptilde,op_prol,info)
-    if (debug_level >= psb_debug_outer_) &
-         & write(debug_unit,*) me,' ',trim(name),&
-         & 'Done NUMBMM 1'
-
-  end if
-
-
-  !
-  ! Ok, let's start over with the restrictor
-  ! 
-  call ptilde%transc(rtilde)
-  call la%cscnv(atmp,info,type='csr')
-  call psb_sphalo(atmp,desc_a,am4,info,&
-       & colcnv=.true.,rowscale=.true.)
-  nrt  = am4%get_nrows() 
-  call am4%csclip(atmp2,info,lone,nrt,lone,ncol)
-  call atmp2%cscnv(info,type='CSR')
-  if (info == psb_success_) call psb_rwextd(ncol,atmp,info,b=atmp2)      
-  call am4%free()
-  call atmp2%free()
-
-  ! This is to compute the transpose. It ONLY works if the
-  ! original A has a symmetric pattern.
-  call atmp%transc(atmp2) 
-  call atmp2%csclip(dat,info,lone,nrow,lone,ncol)
-  call dat%cscnv(info,type='csr')
-  call dat%scal(adinv,info)
-
-  ! Now for the product. 
-  call psb_spspmm(dat,ptilde,datp,info)
-
-  call datp%clone(atmp2,info)
-  call psb_sphalo(atmp2,desc_a,am4,info,&
-       & colcnv=.false.,rowscale=.true.,outfmt='CSR  ')
-  if (info == psb_success_) call psb_rwextd(ncol,atmp2,info,b=am4)      
-  if (info == psb_success_) call am4%free()
-
-
-  call psb_symbmm(dat,atmp2,datdatp,info)
-  call psb_numbmm(dat,atmp2,datdatp)
-  call atmp2%free()
-
-  call datp%mv_to(csc_datp)    
-  call datdatp%mv_to(csc_datdatp)    
-
-  call csc_mat_col_prod(csc_datp,csc_datdatp,omp,info)
-  call csc_mat_col_prod(csc_datdatp,csc_datdatp,oden,info)
-  call psb_sum(ctxt,omp)
-  call psb_sum(ctxt,oden)
-
-
-  ! !$  write(debug_unit,*) trim(name),' OMP_R :',omp
-  ! ! $  write(debug_unit,*) trim(name),' ODEN_R:',oden
-  omp = omp/oden
-  ! !$  write(0,*) 'Check on output restrictor',omp(1:min(size(omp),10))
-  ! Compute omega_int
-  ommx = czero
-  do i=1, ncol
-    if (ilaggr(i) >0) then 
-      omi(i) = omp(ilaggr(i))
-    else
-      omi(i) = czero
-    end if
-    if(abs(omi(i)) .gt. abs(ommx)) ommx = omi(i)
-  end do
-  ! Compute omega_fine
-  ! Going over the columns of atmp means going over the rows
-  ! of A^T. Hopefully ;-) 
-  call atmp%cp_to(acsc)
-
-  do i=1, nrow
-    omf(i) = ommx
-    do j= acsc%icp(i),acsc%icp(i+1)-1
-      if(abs(omi(acsc%ia(j))) .lt. abs(omf(i))) omf(i)=omi(acsc%ia(j))
-    end do
-!!$    if(min(real(omf(i)),aimag(omf(i))) < szero) omf(i) = czero
-    if(psb_minreal(omf(i)) < szero) omf(i) = czero
-  end do
-  omf(1:nrow) = omf(1:nrow)*adinv(1:nrow)
-  call psb_halo(omf,desc_a,info)
-  call acsc%free() 
-
-
-  call atmp%mv_to(acsr1)
-
-  do i=1,acsr1%get_nrows()
-    do j=acsr1%irp(i),acsr1%irp(i+1)-1
-      if (acsr1%ja(j) == i) then 
-        acsr1%val(j) = cone - acsr1%val(j)*omf(acsr1%ja(j))
-      else
-        acsr1%val(j) =      - acsr1%val(j)*omf(acsr1%ja(j))
-      end if
-    end do
-  end do
-  call atmp%mv_from(acsr1)
-
-  call rtilde%mv_to(tmpcoo)
-  nzl = tmpcoo%get_nzeros()
-  i=0
-  do k=1, nzl
-    if ((naggrm1 < tmpcoo%ia(k)) .and. (tmpcoo%ia(k) <= naggrp1)) then
-      i = i+1
-      tmpcoo%val(i) = tmpcoo%val(k)
-      tmpcoo%ia(i)  = tmpcoo%ia(k)
-      tmpcoo%ja(i)  = tmpcoo%ja(k)
-    end if
-  end do
-  call tmpcoo%set_nzeros(i)
-  call rtilde%mv_from(tmpcoo)
-  call rtilde%cscnv(info,type='csr')
-
-  call psb_spspmm(rtilde,atmp,op_restr,info)
-
-  !
-  ! Now we have to gather the halo of op_prol, and add it to itself
-  ! to multiply it by A,
-  !
-  call op_prol%clone(tmp_prol,info) 
-  if (info == psb_success_) call psb_sphalo(tmp_prol,desc_a,am4,info,&
-       & colcnv=.false.,rowscale=.true.)
-  if (info == psb_success_) call psb_rwextd(ncol,tmp_prol,info,b=am4)      
-  if (info == psb_success_) call am4%free()
-
-  if(info /= psb_success_) then
-    call psb_errpush(psb_err_internal_error_,name,a_err='Halo of op_prol')
-    goto 9999
-  end if
-
-  !
-  ! Now we have to fix this.  The only rows of B that are correct 
-  ! are those corresponding to "local" aggregates, i.e. indices in ilaggr(:)
-  !
-  call op_restr%mv_to(tmpcoo)
-
-  nzl = tmpcoo%get_nzeros()
-  i=0
-  do k=1, nzl
-    if ((naggrm1 < tmpcoo%ia(k)) .and. (tmpcoo%ia(k) <= naggrp1)) then
-      i = i+1
-      tmpcoo%val(i) = tmpcoo%val(k)
-      tmpcoo%ia(i)  = tmpcoo%ia(k)
-      tmpcoo%ja(i)  = tmpcoo%ja(k)
-    end if
-  end do
-  call tmpcoo%set_nzeros(i)
-  call op_restr%mv_from(tmpcoo)
-  call op_restr%cscnv(info,type='csr')
-
-
-  if (debug_level >= psb_debug_outer_) &
-       & write(debug_unit,*) me,' ',trim(name),&
-       & 'starting sphalo/ rwxtd'
-
-  call psb_spspmm(la,tmp_prol,am3,info)
-  if (debug_level >= psb_debug_outer_) &
-       & write(debug_unit,*) me,' ',trim(name),&
-       & 'Done SPSPMM 2'
-
-  call psb_sphalo(am3,desc_a,am4,info,&
-       & colcnv=.false.,rowscale=.true.)
-  if (info == psb_success_) call psb_rwextd(ncol,am3,info,b=am4)      
-  if (info == psb_success_) call am4%free()
-
-  if(info /= psb_success_) then
-    call psb_errpush(psb_err_internal_error_,name,&
-         & a_err='Extend am3')
-    goto 9999
-  end if
-  if (debug_level >= psb_debug_outer_) &
-       & write(debug_unit,*) me,' ',trim(name),&
-       & 'Done sphalo/ rwxtd'
-
-  call psb_spspmm(op_restr,am3,ac,info)
-  if (info == psb_success_) call am3%free()
-  if (info == psb_success_) call ac%cscnv(info,type='coo',dupl=psb_dupl_add_)
-
-  if (info /= psb_success_) then
-    call psb_errpush(psb_err_internal_error_,name,&
-         &a_err='Build ac = op_restr x am3')
-    goto 9999
-  end if
+!!$  ! Get the diagonal D
+!!$  adiag =  a%get_diag(info)
+!!$  if (info == psb_success_) &
+!!$       & call psb_realloc(ncol,adiag,info)    
+!!$  if (info == psb_success_) &
+!!$       & call psb_halo(adiag,desc_a,info)
+!!$  if (info == psb_success_) call a%cp_to_l(la)
+!!$  if (info /= psb_success_) then
+!!$    call psb_errpush(psb_err_from_subroutine_,name,a_err='sp_getdiag')
+!!$    goto 9999
+!!$  end if
+!!$
+!!$  do i=1,size(adiag)
+!!$    if (adiag(i) /= czero) then
+!!$      adinv(i) = cone / adiag(i)
+!!$    else
+!!$      adinv(i) = cone
+!!$    end if
+!!$  end do
+!!$
+!!$
+!!$
+!!$  ! 1. Allocate Ptilde in sparse matrix form 
+!!$  call op_prol%mv_to(tmpcoo)
+!!$  call ptilde%mv_from(tmpcoo)
+!!$  call ptilde%cscnv(info,type='csr')
+!!$
+!!$  if (info == psb_success_) call la%cscnv(am3,info,type='csr',dupl=psb_dupl_add_)
+!!$  if (info == psb_success_) call la%cscnv(da,info,type='csr',dupl=psb_dupl_add_)
+!!$  if (info /= psb_success_) then
+!!$    call psb_errpush(psb_err_from_subroutine_,name,a_err='spcnv')
+!!$    goto 9999
+!!$  end if
+!!$  if (debug_level >= psb_debug_outer_) &
+!!$       & write(debug_unit,*) me,' ',trim(name),&
+!!$       & ' Initial copies done.'
+!!$
+!!$  call da%scal(adinv,info)
+!!$
+!!$  call psb_spspmm(da,ptilde,dap,info)
+!!$
+!!$  if(info /= psb_success_) then
+!!$    call psb_errpush(psb_err_from_subroutine_,name,a_err='spspmm 1')
+!!$    goto 9999
+!!$  end if
+!!$
+!!$  call dap%clone(atmp,info)
+!!$
+!!$  call psb_sphalo(atmp,desc_a,am4,info,&
+!!$       & colcnv=.false.,rowscale=.true.,outfmt='CSR  ')
+!!$  if (info == psb_success_) call psb_rwextd(ncol,atmp,info,b=am4)      
+!!$  if (info == psb_success_) call am4%free()
+!!$
+!!$  call psb_spspmm(da,atmp,dadap,info)
+!!$  call atmp%free()
+!!$
+!!$  !  !$  write(0,*) 'Columns of AP',psb_sp_get_ncols(ap)
+!!$  !  !$  write(0,*) 'Columns of ADAP',psb_sp_get_ncols(adap)
+!!$  call dap%mv_to(csc_dap)
+!!$  call dadap%mv_to(csc_dadap)
+!!$
+!!$  call csc_mat_col_prod(csc_dap,csc_dadap,omp,info)
+!!$  call csc_mat_col_prod(csc_dadap,csc_dadap,oden,info)
+!!$  call psb_sum(ctxt,omp)
+!!$  call psb_sum(ctxt,oden)
+!!$  ! !$  write(0,*) trim(name),' OMP :',omp
+!!$  ! !$  write(0,*) trim(name),' ODEN:',oden
+!!$
+!!$  omp = omp/oden
+!!$
+!!$  ! !$  write(0,*) 'Check on output prolongator ',omp(1:min(size(omp),10))
+!!$  if (debug_level >= psb_debug_outer_) &
+!!$       & write(debug_unit,*) me,' ',trim(name),&
+!!$       & 'Done NUMBMM 1'
+!!$
+!!$  call am3%mv_to(acsr3)
+!!$  ! Compute omega_int
+!!$  ommx = czero
+!!$  do i=1, ncol
+!!$    if (ilaggr(i) >0) then 
+!!$      omi(i) = omp(ilaggr(i))
+!!$    else
+!!$      omi(i) = czero
+!!$    end if
+!!$    if(abs(omi(i)) .gt. abs(ommx)) ommx = omi(i)
+!!$  end do
+!!$  ! Compute omega_fine
+!!$  do i=1, nrow
+!!$    omf(i) = ommx
+!!$    do j=acsr3%irp(i),acsr3%irp(i+1)-1
+!!$      if(abs(omi(acsr3%ja(j))) .lt. abs(omf(i))) omf(i)=omi(acsr3%ja(j))
+!!$    end do
+!!$ ! !     if(min(real(omf(i)),aimag(omf(i))) < szero) omf(i) = czero
+!!$    if(psb_minreal(omf(i)) < szero) omf(i) = czero
+!!$  end do
+!!$
+!!$  omf(1:nrow) = omf(1:nrow) * adinv(1:nrow)
+!!$
+!!$  if (filter_mat) then
+!!$    !
+!!$    ! Build the filtered matrix Af from A
+!!$    ! 
+!!$    call la%cscnv(acsrf,info,dupl=psb_dupl_add_)
+!!$
+!!$    do i=1,nrow
+!!$      tmp = czero
+!!$      jd  = -1 
+!!$      do j=acsrf%irp(i),acsrf%irp(i+1)-1
+!!$        if (acsrf%ja(j) == i) jd = j 
+!!$        if (abs(acsrf%val(j)) < theta*sqrt(abs(adiag(i)*adiag(acsrf%ja(j))))) then
+!!$          tmp=tmp+acsrf%val(j)
+!!$          acsrf%val(j)=czero
+!!$        endif
+!!$      enddo
+!!$      if (jd == -1) then 
+!!$        write(0,*) 'Wrong input: we need the diagonal!!!!', i
+!!$      else
+!!$        acsrf%val(jd)=acsrf%val(jd)-tmp
+!!$      end if
+!!$    enddo
+!!$    ! Take out zeroed terms 
+!!$    call acsrf%clean_zeros(info)
+!!$
+!!$    !
+!!$    ! Build the smoothed prolongator using the filtered matrix
+!!$    ! 
+!!$    do i=1,acsrf%get_nrows()
+!!$      do j=acsrf%irp(i),acsrf%irp(i+1)-1
+!!$        if (acsrf%ja(j) == i) then 
+!!$          acsrf%val(j) = cone - omf(i)*acsrf%val(j) 
+!!$        else
+!!$          acsrf%val(j) = - omf(i)*acsrf%val(j) 
+!!$        end if
+!!$      end do
+!!$    end do
+!!$
+!!$    if (debug_level >= psb_debug_outer_) &
+!!$         & write(debug_unit,*) me,' ',trim(name),&
+!!$         & 'Done gather, going for SYMBMM 1'
+!!$
+!!$    call af%mv_from(acsrf)
+!!$    !
+!!$    ! op_prol = (I-w*D*Af)Ptilde
+!!$    ! Doing it this way means to consider diag(Af_i)
+!!$    ! 
+!!$    !
+!!$    call psb_spspmm(af,ptilde,op_prol,info)
+!!$    if (debug_level >= psb_debug_outer_) &
+!!$         & write(debug_unit,*) me,' ',trim(name),&
+!!$         & 'Done SPSPMM 1'
+!!$  else
+!!$    !
+!!$    ! Build the smoothed prolongator using the original matrix
+!!$    !
+!!$    do i=1,acsr3%get_nrows()
+!!$      do j=acsr3%irp(i),acsr3%irp(i+1)-1
+!!$        if (acsr3%ja(j) == i) then 
+!!$          acsr3%val(j) = cone - omf(i)*acsr3%val(j) 
+!!$        else
+!!$          acsr3%val(j) = - omf(i)*acsr3%val(j) 
+!!$        end if
+!!$      end do
+!!$    end do
+!!$
+!!$    call am3%mv_from(acsr3)
+!!$    if (debug_level >= psb_debug_outer_) &
+!!$         & write(debug_unit,*) me,' ',trim(name),&
+!!$         & 'Done gather, going for SYMBMM 1'
+!!$    !
+!!$    ! 
+!!$    ! op_prol = (I-w*D*A)Ptilde
+!!$    ! 
+!!$    !
+!!$    call psb_spspmm(am3,ptilde,op_prol,info)
+!!$    if (debug_level >= psb_debug_outer_) &
+!!$         & write(debug_unit,*) me,' ',trim(name),&
+!!$         & 'Done NUMBMM 1'
+!!$
+!!$  end if
+!!$
+!!$
+!!$  !
+!!$  ! Ok, let's start over with the restrictor
+!!$  ! 
+!!$  call ptilde%transc(rtilde)
+!!$  call la%cscnv(atmp,info,type='csr')
+!!$  call psb_sphalo(atmp,desc_a,am4,info,&
+!!$       & colcnv=.true.,rowscale=.true.)
+!!$  nrt  = am4%get_nrows() 
+!!$  call am4%csclip(atmp2,info,lone,nrt,lone,ncol)
+!!$  call atmp2%cscnv(info,type='CSR')
+!!$  if (info == psb_success_) call psb_rwextd(ncol,atmp,info,b=atmp2)      
+!!$  call am4%free()
+!!$  call atmp2%free()
+!!$
+!!$  ! This is to compute the transpose. It ONLY works if the
+!!$  ! original A has a symmetric pattern.
+!!$  call atmp%transc(atmp2) 
+!!$  call atmp2%csclip(dat,info,lone,nrow,lone,ncol)
+!!$  call dat%cscnv(info,type='csr')
+!!$  call dat%scal(adinv,info)
+!!$
+!!$  ! Now for the product. 
+!!$  call psb_spspmm(dat,ptilde,datp,info)
+!!$
+!!$  call datp%clone(atmp2,info)
+!!$  call psb_sphalo(atmp2,desc_a,am4,info,&
+!!$       & colcnv=.false.,rowscale=.true.,outfmt='CSR  ')
+!!$  if (info == psb_success_) call psb_rwextd(ncol,atmp2,info,b=am4)      
+!!$  if (info == psb_success_) call am4%free()
+!!$
+!!$
+!!$  call psb_symbmm(dat,atmp2,datdatp,info)
+!!$  call psb_numbmm(dat,atmp2,datdatp)
+!!$  call atmp2%free()
+!!$
+!!$  call datp%mv_to(csc_datp)    
+!!$  call datdatp%mv_to(csc_datdatp)    
+!!$
+!!$  call csc_mat_col_prod(csc_datp,csc_datdatp,omp,info)
+!!$  call csc_mat_col_prod(csc_datdatp,csc_datdatp,oden,info)
+!!$  call psb_sum(ctxt,omp)
+!!$  call psb_sum(ctxt,oden)
+!!$
+!!$
+!!$  ! !$  write(debug_unit,*) trim(name),' OMP_R :',omp
+!!$  ! ! $  write(debug_unit,*) trim(name),' ODEN_R:',oden
+!!$  omp = omp/oden
+!!$  ! !$  write(0,*) 'Check on output restrictor',omp(1:min(size(omp),10))
+!!$  ! Compute omega_int
+!!$  ommx = czero
+!!$  do i=1, ncol
+!!$    if (ilaggr(i) >0) then 
+!!$      omi(i) = omp(ilaggr(i))
+!!$    else
+!!$      omi(i) = czero
+!!$    end if
+!!$    if(abs(omi(i)) .gt. abs(ommx)) ommx = omi(i)
+!!$  end do
+!!$  ! Compute omega_fine
+!!$  ! Going over the columns of atmp means going over the rows
+!!$  ! of A^T. Hopefully ;-) 
+!!$  call atmp%cp_to(acsc)
+!!$
+!!$  do i=1, nrow
+!!$    omf(i) = ommx
+!!$    do j= acsc%icp(i),acsc%icp(i+1)-1
+!!$      if(abs(omi(acsc%ia(j))) .lt. abs(omf(i))) omf(i)=omi(acsc%ia(j))
+!!$    end do
+!!$ ! !    if(min(real(omf(i)),aimag(omf(i))) < szero) omf(i) = czero
+!!$    if(psb_minreal(omf(i)) < szero) omf(i) = czero
+!!$  end do
+!!$  omf(1:nrow) = omf(1:nrow)*adinv(1:nrow)
+!!$  call psb_halo(omf,desc_a,info)
+!!$  call acsc%free() 
+!!$
+!!$
+!!$  call atmp%mv_to(acsr1)
+!!$
+!!$  do i=1,acsr1%get_nrows()
+!!$    do j=acsr1%irp(i),acsr1%irp(i+1)-1
+!!$      if (acsr1%ja(j) == i) then 
+!!$        acsr1%val(j) = cone - acsr1%val(j)*omf(acsr1%ja(j))
+!!$      else
+!!$        acsr1%val(j) =      - acsr1%val(j)*omf(acsr1%ja(j))
+!!$      end if
+!!$    end do
+!!$  end do
+!!$  call atmp%mv_from(acsr1)
+!!$
+!!$  call rtilde%mv_to(tmpcoo)
+!!$  nzl = tmpcoo%get_nzeros()
+!!$  i=0
+!!$  do k=1, nzl
+!!$    if ((naggrm1 < tmpcoo%ia(k)) .and. (tmpcoo%ia(k) <= naggrp1)) then
+!!$      i = i+1
+!!$      tmpcoo%val(i) = tmpcoo%val(k)
+!!$      tmpcoo%ia(i)  = tmpcoo%ia(k)
+!!$      tmpcoo%ja(i)  = tmpcoo%ja(k)
+!!$    end if
+!!$  end do
+!!$  call tmpcoo%set_nzeros(i)
+!!$  call rtilde%mv_from(tmpcoo)
+!!$  call rtilde%cscnv(info,type='csr')
+!!$
+!!$  call psb_spspmm(rtilde,atmp,op_restr,info)
+!!$
+!!$  !
+!!$  ! Now we have to gather the halo of op_prol, and add it to itself
+!!$  ! to multiply it by A,
+!!$  !
+!!$  call op_prol%clone(tmp_prol,info) 
+!!$  if (info == psb_success_) call psb_sphalo(tmp_prol,desc_a,am4,info,&
+!!$       & colcnv=.false.,rowscale=.true.)
+!!$  if (info == psb_success_) call psb_rwextd(ncol,tmp_prol,info,b=am4)      
+!!$  if (info == psb_success_) call am4%free()
+!!$
+!!$  if(info /= psb_success_) then
+!!$    call psb_errpush(psb_err_internal_error_,name,a_err='Halo of op_prol')
+!!$    goto 9999
+!!$  end if
+!!$
+!!$  !
+!!$  ! Now we have to fix this.  The only rows of B that are correct 
+!!$  ! are those corresponding to "local" aggregates, i.e. indices in ilaggr(:)
+!!$  !
+!!$  call op_restr%mv_to(tmpcoo)
+!!$
+!!$  nzl = tmpcoo%get_nzeros()
+!!$  i=0
+!!$  do k=1, nzl
+!!$    if ((naggrm1 < tmpcoo%ia(k)) .and. (tmpcoo%ia(k) <= naggrp1)) then
+!!$      i = i+1
+!!$      tmpcoo%val(i) = tmpcoo%val(k)
+!!$      tmpcoo%ia(i)  = tmpcoo%ia(k)
+!!$      tmpcoo%ja(i)  = tmpcoo%ja(k)
+!!$    end if
+!!$  end do
+!!$  call tmpcoo%set_nzeros(i)
+!!$  call op_restr%mv_from(tmpcoo)
+!!$  call op_restr%cscnv(info,type='csr')
+!!$
+!!$
+!!$  if (debug_level >= psb_debug_outer_) &
+!!$       & write(debug_unit,*) me,' ',trim(name),&
+!!$       & 'starting sphalo/ rwxtd'
+!!$
+!!$  call psb_spspmm(la,tmp_prol,am3,info)
+!!$  if (debug_level >= psb_debug_outer_) &
+!!$       & write(debug_unit,*) me,' ',trim(name),&
+!!$       & 'Done SPSPMM 2'
+!!$
+!!$  call psb_sphalo(am3,desc_a,am4,info,&
+!!$       & colcnv=.false.,rowscale=.true.)
+!!$  if (info == psb_success_) call psb_rwextd(ncol,am3,info,b=am4)      
+!!$  if (info == psb_success_) call am4%free()
+!!$
+!!$  if(info /= psb_success_) then
+!!$    call psb_errpush(psb_err_internal_error_,name,&
+!!$         & a_err='Extend am3')
+!!$    goto 9999
+!!$  end if
+!!$  if (debug_level >= psb_debug_outer_) &
+!!$       & write(debug_unit,*) me,' ',trim(name),&
+!!$       & 'Done sphalo/ rwxtd'
+!!$
+!!$  call psb_spspmm(op_restr,am3,ac,info)
+!!$  if (info == psb_success_) call am3%free()
+!!$  if (info == psb_success_) call ac%cscnv(info,type='coo',dupl=psb_dupl_add_)
+!!$
+!!$  if (info /= psb_success_) then
+!!$    call psb_errpush(psb_err_internal_error_,name,&
+!!$         &a_err='Build ac = op_restr x am3')
+!!$    goto 9999
+!!$  end if



@@ -116,7 +116,7 @@ subroutine amg_caggrmat_smth_bld(a,desc_a,ilaggr,nlaggr,parms,&
  type(psb_desc_type), intent(inout)       :: desc_a
  integer(psb_lpk_), intent(inout)         :: ilaggr(:), nlaggr(:)
  type(amg_sml_parms), intent(inout)      :: parms 
-  type(psb_cspmat_type), intent(out)      :: op_prol,ac,op_restr
+  type(psb_cspmat_type), intent(inout)      :: op_prol,ac,op_restr
  type(psb_lcspmat_type), intent(inout)    :: t_prol
  type(psb_desc_type), intent(inout)       :: desc_ac
  integer(psb_ipk_), intent(out)           :: info
@@ -83,8 +83,8 @@ subroutine  amg_d_dec_aggregator_build_tprol(ag,parms,ag_data,&
  class(amg_d_dec_aggregator_type), target, intent(inout) :: ag
  type(amg_dml_parms), intent(inout)  :: parms
  type(amg_daggr_data), intent(in)    :: ag_data
-  type(psb_dspmat_type), intent(in)   :: a
-  type(psb_desc_type), intent(in)     :: desc_a
+  type(psb_dspmat_type), intent(inout) :: a
+  type(psb_desc_type), intent(inout)    :: desc_a
  integer(psb_lpk_), allocatable, intent(out) :: ilaggr(:),nlaggr(:)
  type(psb_ldspmat_type), intent(out)  :: t_prol
  integer(psb_ipk_), intent(out)      :: info
@@ -48,7 +48,7 @@ subroutine  amg_d_parmatch_aggregator_build_tprol(ag,parms,ag_data,&
  use amg_base_prec_type
  use amg_d_inner_mod
 #if defined(SERIAL_MPI)
-    use amg_d_parmatch_aggregator_mod
+  use amg_d_parmatch_aggregator_mod
 #else
  use amg_d_parmatch_aggregator_mod, amg_protect_name => amg_d_parmatch_aggregator_build_tprol
 #endif
@@ -58,7 +58,7 @@ subroutine  amg_d_parmatch_aggregator_build_tprol(ag,parms,ag_data,&
  type(amg_dml_parms), intent(inout)   :: parms
  type(amg_daggr_data), intent(in)     :: ag_data
  type(psb_dspmat_type), intent(inout) :: a
-  type(psb_desc_type), intent(inout)   :: desc_a
+  type(psb_desc_type), intent(inout)      :: desc_a
  integer(psb_lpk_), allocatable, intent(out) :: ilaggr(:),nlaggr(:)
  type(psb_ldspmat_type), intent(out)  :: t_prol
  integer(psb_ipk_), intent(out)      :: info
@@ -68,7 +68,8 @@ subroutine  amg_d_parmatch_aggregator_build_tprol(ag,parms,ag_data,&
  real(psb_dpk_), allocatable    :: tmpw(:), tmpwnxt(:)
  integer(psb_lpk_), allocatable :: ixaggr(:), nxaggr(:), tlaggr(:), ivr(:)
  type(psb_dspmat_type)          :: a_tmp
-  integer(c_int) :: match_algorithm, n_sweeps, max_csize, max_nlevels
+  integer(psb_ipk_)    :: match_algorithm, n_sweeps
+  integer(psb_lpk_)    :: target_csize
  character(len=40)    :: name, ch_err
  character(len=80)    :: fname, prefix_
  type(psb_ctxt_type)  :: ictxt
@@ -128,27 +129,22 @@ subroutine  amg_d_parmatch_aggregator_build_tprol(ag,parms,ag_data,&
      write(debug_unit, *) 'Warning: AGGR_SIZE reset to value ',2**n_sweeps
    end if
  end if
-  if (ag%max_csize > 0) then
-    max_csize       = ag%max_csize
+  if (ag_data%target_coarse_size > 0) then
+    target_csize       = ag_data%target_coarse_size
  else
-    max_csize       = ag_data%min_coarse_size
-  end if
-  if (ag%max_nlevels > 0) then
-    max_nlevels     = ag%max_nlevels
-  else
-    max_nlevels = ag_data%max_levs
+    target_csize       = ag_data%min_coarse_size
  end if
  if (.true.) then
    block
      integer(psb_ipk_) :: ipv(2)
-      ipv(1) = max_csize
+      ipv(1) = target_csize
      ipv(2) = n_sweeps
      call psb_bcast(ictxt,ipv)
-      max_csize = ipv(1)
+      target_csize = ipv(1)
      n_sweeps  = ipv(2)
    end block
  else
-    call psb_bcast(ictxt,max_csize)
+    call psb_bcast(ictxt,target_csize)
    call psb_bcast(ictxt,n_sweeps)
  end if
  if (n_sweeps /= ag%n_sweeps) then
@@ -156,7 +152,7 @@ subroutine  amg_d_parmatch_aggregator_build_tprol(ag,parms,ag_data,&
  end if
 !!$  if (me==0) write(0,*) 'Matching sweeps: ',n_sweeps
  n_sweeps = max(1,n_sweeps)
-  if (debug) write(0,*) me,' Copies, with n_sweeps: ',n_sweeps,max_csize
+  if (debug) write(0,*) me,' Copies, with n_sweeps: ',n_sweeps,target_csize
  if (ag%unsmoothed_hierarchy.and.allocated(ag%base_a)) then
    call ag%base_a%cp_to(acsr)
    if (ag%do_clean_zeros) call acsr%clean_zeros(info)
@@ -242,7 +238,7 @@ subroutine  amg_d_parmatch_aggregator_build_tprol(ag,parms,ag_data,&

  if (debug) then
    call psb_barrier(ictxt)
-    if (me == 0) write(0,*) 'N_sweeps ',n_sweeps,nr,desc_acv(0)%is_ok(),max_csize
+    if (me == 0) write(0,*) 'N_sweeps ',n_sweeps,nr,desc_acv(0)%is_ok(),target_csize
  end if

  !
@@ -264,7 +260,7 @@ subroutine  amg_d_parmatch_aggregator_build_tprol(ag,parms,ag_data,&
    !
    if (debug) write(0,*) me,' Into matchbox_build_prol ',info
    if (do_timings) call psb_tic(idx_mboxp)
-    call dmatchboxp_build_prol(tmpw,acv(i-1),desc_acv(i-1),ixaggr,nxaggr,tmp_prol,info,&
+    call amg_d_matchboxp_build_prol(tmpw,acv(i-1),desc_acv(i-1),ixaggr,nxaggr,tmp_prol,info,&
         & symmetrize=ag%need_symmetrize,reproducible=ag%reproducible_matching)
    if (do_timings) call psb_toc(idx_mboxp)
    if (debug) write(0,*) me,' Out from matchbox_build_prol ',info
@@ -300,11 +296,11 @@ subroutine  amg_d_parmatch_aggregator_build_tprol(ag,parms,ag_data,&

    if (debug) then
      call psb_barrier(ictxt)
-      if (me==0) write(0,*) me,trim(name),' Done mat_asb:',i,sum(nxaggr),max_csize,info
+      if (me==0) write(0,*) me,trim(name),' Done mat_asb:',i,sum(nxaggr),target_csize,info
      csz = sum(nxaggr)
      call psb_bcast(ictxt,csz)
      if (csz /= sum(nxaggr)) write(0,*) me,trim(name),' Mismatch matasb',&
-           & csz,sum(nxaggr),max_csize
+           & csz,sum(nxaggr),target_csize
    end if
    if (psb_errstatus_fatal())  write(0,*)me,trim(name),'Error fatal on entry to tmpwnxt 2'

@@ -342,10 +338,10 @@ subroutine  amg_d_parmatch_aggregator_build_tprol(ag,parms,ag_data,&
    call move_alloc(tmpwnxt,tmpw)
    if (debug) then
      if (csz /= sum(nlaggr)) write(0,*) me,trim(name),' Mismatch 2 matasb',&
-           & csz,sum(nlaggr),max_csize, info
+           & csz,sum(nlaggr),target_csize, info
    end if
    call acv(i-1)%free()
-    if ((sum(nlaggr) <= max_csize).or.(any(nlaggr==0))) then
+    if ((sum(nlaggr) <= target_csize).or.(any(nlaggr==0))) then
      x_sweeps = i
      exit sweeps_loop
    end if
@@ -122,7 +122,7 @@ subroutine amg_d_parmatch_smth_bld(ag,a,desc_a,ilaggr,nlaggr,parms,&
  integer(psb_lpk_), intent(inout)         :: ilaggr(:), nlaggr(:)
  type(amg_dml_parms), intent(inout)    :: parms
  type(psb_ldspmat_type), intent(inout)  :: t_prol
-  type(psb_dspmat_type), intent(out)    :: op_prol,ac,op_restr
+  type(psb_dspmat_type), intent(inout)   :: op_prol,ac,op_restr
  type(psb_desc_type), intent(inout)    :: desc_ac
  integer(psb_ipk_), intent(out)           :: info

@@ -108,7 +108,7 @@ subroutine amg_d_parmatch_spmm_bld(a,desc_a,ilaggr,nlaggr,parms,&

  ! Arguments
  type(psb_dspmat_type), intent(in)       :: a
-  type(psb_desc_type), intent(in)         :: desc_a
+  type(psb_desc_type), intent(inout)       :: desc_a
  integer(psb_lpk_), intent(inout)        :: ilaggr(:), nlaggr(:)
  type(amg_dml_parms), intent(inout)      :: parms
  type(psb_ldspmat_type), intent(inout)   :: t_prol
@@ -108,11 +108,11 @@ subroutine amg_d_parmatch_spmm_bld_inner(a_csr,desc_a,ilaggr,nlaggr,parms,&

  ! Arguments
  type(psb_d_csr_sparse_mat), intent(inout) :: a_csr
-  type(psb_desc_type), intent(in)           :: desc_a
+  type(psb_desc_type), intent(inout)          :: desc_a
  integer(psb_lpk_), intent(inout)          :: ilaggr(:), nlaggr(:)
  type(amg_dml_parms), intent(inout)        :: parms
  type(psb_ldspmat_type), intent(inout)     :: t_prol
-  type(psb_dspmat_type), intent(out)        :: ac, op_prol, op_restr
+  type(psb_dspmat_type), intent(inout)      :: ac, op_prol, op_restr
  type(psb_desc_type), intent(out)          :: desc_ac
  integer(psb_ipk_), intent(out)            :: info

@@ -108,7 +108,7 @@ subroutine amg_d_parmatch_spmm_bld_ov(a,desc_a,ilaggr,nlaggr,parms,&

  ! Arguments
  type(psb_dspmat_type), intent(inout)    :: a
-  type(psb_desc_type), intent(in)         :: desc_a
+  type(psb_desc_type), intent(inout)       :: desc_a
  integer(psb_lpk_), intent(inout)        :: ilaggr(:), nlaggr(:)
  type(amg_dml_parms), intent(inout)      :: parms
  type(psb_ldspmat_type), intent(inout)   :: t_prol
@@ -86,8 +86,8 @@ subroutine  amg_d_symdec_aggregator_build_tprol(ag,parms,ag_data,&
  class(amg_d_symdec_aggregator_type), target, intent(inout) :: ag
  type(amg_dml_parms), intent(inout)  :: parms
  type(amg_daggr_data), intent(in)    :: ag_data
-  type(psb_dspmat_type), intent(in)   :: a
-  type(psb_desc_type), intent(in)     :: desc_a
+  type(psb_dspmat_type), intent(inout) :: a
+  type(psb_desc_type), intent(inout)   :: desc_a
  integer(psb_lpk_), allocatable, intent(out) :: ilaggr(:),nlaggr(:)
  type(psb_ldspmat_type), intent(out)  :: op_prol
  integer(psb_ipk_), intent(out)      :: info
@@ -105,7 +105,7 @@
 !
 !
 subroutine amg_daggrmat_minnrg_bld(a,desc_a,ilaggr,nlaggr,parms,&
-     & ac,desc_ac,op_prol,op_restr,info)
+     & ac,desc_ac,op_prol,op_restr,t_prol,info)
  use psb_base_mod
  use amg_base_prec_type
  use amg_d_inner_mod, amg_protect_name => amg_daggrmat_minnrg_bld
@@ -117,8 +117,8 @@ subroutine amg_daggrmat_minnrg_bld(a,desc_a,ilaggr,nlaggr,parms,&
  type(psb_desc_type), intent(inout)            :: desc_a
  integer(psb_lpk_), intent(inout)              :: ilaggr(:), nlaggr(:)
  type(amg_dml_parms), intent(inout)         :: parms 
-  type(psb_ldspmat_type), intent(inout)       :: op_prol
-  type(psb_ldspmat_type), intent(out)         :: ac,op_restr
+  type(psb_ldspmat_type), intent(inout)       :: t_prol
+  type(psb_dspmat_type), intent(inout)       :: op_prol, ac,op_restr
  type(psb_desc_type), intent(inout)            :: desc_ac
  integer(psb_ipk_), intent(out)                :: info

@@ -171,6 +171,8 @@ subroutine amg_daggrmat_minnrg_bld(a,desc_a,ilaggr,nlaggr,parms,&

  filter_mat = (parms%aggr_filter == amg_filter_mat_)

+  !NEEDS TO BE REWORKED !!
+  
  ! naggr: number of local aggregates
  ! nrow: local rows. 
  ! 
@@ -183,361 +185,361 @@ subroutine amg_daggrmat_minnrg_bld(a,desc_a,ilaggr,nlaggr,parms,&
    goto 9999      
  end if

-  ! Get the diagonal D
-  adiag =  a%get_diag(info)
-  if (info == psb_success_) &
-       & call psb_realloc(ncol,adiag,info)    
-  if (info == psb_success_) &
-       & call psb_halo(adiag,desc_a,info)
-  if (info == psb_success_) call a%cp_to_l(la)
-  if (info /= psb_success_) then
-    call psb_errpush(psb_err_from_subroutine_,name,a_err='sp_getdiag')
-    goto 9999
-  end if
-
-  do i=1,size(adiag)
-    if (adiag(i) /= dzero) then
-      adinv(i) = done / adiag(i)
-    else
-      adinv(i) = done
-    end if
-  end do
-
-
-
-  ! 1. Allocate Ptilde in sparse matrix form 
-  call op_prol%mv_to(tmpcoo)
-  call ptilde%mv_from(tmpcoo)
-  call ptilde%cscnv(info,type='csr')
-
-  if (info == psb_success_) call la%cscnv(am3,info,type='csr',dupl=psb_dupl_add_)
-  if (info == psb_success_) call la%cscnv(da,info,type='csr',dupl=psb_dupl_add_)
-  if (info /= psb_success_) then
-    call psb_errpush(psb_err_from_subroutine_,name,a_err='spcnv')
-    goto 9999
-  end if
-  if (debug_level >= psb_debug_outer_) &
-       & write(debug_unit,*) me,' ',trim(name),&
-       & ' Initial copies done.'
-
-  call da%scal(adinv,info)
-
-  call psb_spspmm(da,ptilde,dap,info)
-
-  if(info /= psb_success_) then
-    call psb_errpush(psb_err_from_subroutine_,name,a_err='spspmm 1')
-    goto 9999
-  end if
-
-  call dap%clone(atmp,info)
-
-  call psb_sphalo(atmp,desc_a,am4,info,&
-       & colcnv=.false.,rowscale=.true.,outfmt='CSR  ')
-  if (info == psb_success_) call psb_rwextd(ncol,atmp,info,b=am4)      
-  if (info == psb_success_) call am4%free()
-
-  call psb_spspmm(da,atmp,dadap,info)
-  call atmp%free()
-
-  !  !$  write(0,*) 'Columns of AP',psb_sp_get_ncols(ap)
-  !  !$  write(0,*) 'Columns of ADAP',psb_sp_get_ncols(adap)
-  call dap%mv_to(csc_dap)
-  call dadap%mv_to(csc_dadap)
-
-  call csc_mat_col_prod(csc_dap,csc_dadap,omp,info)
-  call csc_mat_col_prod(csc_dadap,csc_dadap,oden,info)
-  call psb_sum(ctxt,omp)
-  call psb_sum(ctxt,oden)
-  ! !$  write(0,*) trim(name),' OMP :',omp
-  ! !$  write(0,*) trim(name),' ODEN:',oden
-
-  omp = omp/oden
-
-  ! !$  write(0,*) 'Check on output prolongator ',omp(1:min(size(omp),10))
-  if (debug_level >= psb_debug_outer_) &
-       & write(debug_unit,*) me,' ',trim(name),&
-       & 'Done NUMBMM 1'
-
-  call am3%mv_to(acsr3)
-  ! Compute omega_int
-  ommx = dzero
-  do i=1, ncol
-    if (ilaggr(i) >0) then 
-      omi(i) = omp(ilaggr(i))
-    else
-      omi(i) = dzero
-    end if
-    if(abs(omi(i)) .gt. abs(ommx)) ommx = omi(i)
-  end do
-  ! Compute omega_fine
-  do i=1, nrow
-    omf(i) = ommx
-    do j=acsr3%irp(i),acsr3%irp(i+1)-1
-      if(abs(omi(acsr3%ja(j))) .lt. abs(omf(i))) omf(i)=omi(acsr3%ja(j))
-    end do
-!!$    if(min(real(omf(i)),aimag(omf(i))) < dzero) omf(i) = dzero
-    if(psb_minreal(omf(i)) < dzero) omf(i) = dzero
-  end do
-
-  omf(1:nrow) = omf(1:nrow) * adinv(1:nrow)
-
-  if (filter_mat) then
-    !
-    ! Build the filtered matrix Af from A
-    ! 
-    call la%cscnv(acsrf,info,dupl=psb_dupl_add_)
-
-    do i=1,nrow
-      tmp = dzero
-      jd  = -1 
-      do j=acsrf%irp(i),acsrf%irp(i+1)-1
-        if (acsrf%ja(j) == i) jd = j 
-        if (abs(acsrf%val(j)) < theta*sqrt(abs(adiag(i)*adiag(acsrf%ja(j))))) then
-          tmp=tmp+acsrf%val(j)
-          acsrf%val(j)=dzero
-        endif
-      enddo
-      if (jd == -1) then 
-        write(0,*) 'Wrong input: we need the diagonal!!!!', i
-      else
-        acsrf%val(jd)=acsrf%val(jd)-tmp
-      end if
-    enddo
-    ! Take out zeroed terms 
-    call acsrf%clean_zeros(info)
-
-    !
-    ! Build the smoothed prolongator using the filtered matrix
-    ! 
-    do i=1,acsrf%get_nrows()
-      do j=acsrf%irp(i),acsrf%irp(i+1)-1
-        if (acsrf%ja(j) == i) then 
-          acsrf%val(j) = done - omf(i)*acsrf%val(j) 
-        else
-          acsrf%val(j) = - omf(i)*acsrf%val(j) 
-        end if
-      end do
-    end do
-
-    if (debug_level >= psb_debug_outer_) &
-         & write(debug_unit,*) me,' ',trim(name),&
-         & 'Done gather, going for SYMBMM 1'
-
-    call af%mv_from(acsrf)
-    !
-    ! op_prol = (I-w*D*Af)Ptilde
-    ! Doing it this way means to consider diag(Af_i)
-    ! 
-    !
-    call psb_spspmm(af,ptilde,op_prol,info)
-    if (debug_level >= psb_debug_outer_) &
-         & write(debug_unit,*) me,' ',trim(name),&
-         & 'Done SPSPMM 1'
-  else
-    !
-    ! Build the smoothed prolongator using the original matrix
-    !
-    do i=1,acsr3%get_nrows()
-      do j=acsr3%irp(i),acsr3%irp(i+1)-1
-        if (acsr3%ja(j) == i) then 
-          acsr3%val(j) = done - omf(i)*acsr3%val(j) 
-        else
-          acsr3%val(j) = - omf(i)*acsr3%val(j) 
-        end if
-      end do
-    end do
-
-    call am3%mv_from(acsr3)
-    if (debug_level >= psb_debug_outer_) &
-         & write(debug_unit,*) me,' ',trim(name),&
-         & 'Done gather, going for SYMBMM 1'
-    !
-    ! 
-    ! op_prol = (I-w*D*A)Ptilde
-    ! 
-    !
-    call psb_spspmm(am3,ptilde,op_prol,info)
-    if (debug_level >= psb_debug_outer_) &
-         & write(debug_unit,*) me,' ',trim(name),&
-         & 'Done NUMBMM 1'
-
-  end if
-
-
-  !
-  ! Ok, let's start over with the restrictor
-  ! 
-  call ptilde%transc(rtilde)
-  call la%cscnv(atmp,info,type='csr')
-  call psb_sphalo(atmp,desc_a,am4,info,&
-       & colcnv=.true.,rowscale=.true.)
-  nrt  = am4%get_nrows() 
-  call am4%csclip(atmp2,info,lone,nrt,lone,ncol)
-  call atmp2%cscnv(info,type='CSR')
-  if (info == psb_success_) call psb_rwextd(ncol,atmp,info,b=atmp2)      
-  call am4%free()
-  call atmp2%free()
-
-  ! This is to compute the transpose. It ONLY works if the
-  ! original A has a symmetric pattern.
-  call atmp%transc(atmp2) 
-  call atmp2%csclip(dat,info,lone,nrow,lone,ncol)
-  call dat%cscnv(info,type='csr')
-  call dat%scal(adinv,info)
-
-  ! Now for the product. 
-  call psb_spspmm(dat,ptilde,datp,info)
-
-  call datp%clone(atmp2,info)
-  call psb_sphalo(atmp2,desc_a,am4,info,&
-       & colcnv=.false.,rowscale=.true.,outfmt='CSR  ')
-  if (info == psb_success_) call psb_rwextd(ncol,atmp2,info,b=am4)      
-  if (info == psb_success_) call am4%free()
-
-
-  call psb_symbmm(dat,atmp2,datdatp,info)
-  call psb_numbmm(dat,atmp2,datdatp)
-  call atmp2%free()
-
-  call datp%mv_to(csc_datp)    
-  call datdatp%mv_to(csc_datdatp)    
-
-  call csc_mat_col_prod(csc_datp,csc_datdatp,omp,info)
-  call csc_mat_col_prod(csc_datdatp,csc_datdatp,oden,info)
-  call psb_sum(ctxt,omp)
-  call psb_sum(ctxt,oden)
-
-
-  ! !$  write(debug_unit,*) trim(name),' OMP_R :',omp
-  ! ! $  write(debug_unit,*) trim(name),' ODEN_R:',oden
-  omp = omp/oden
-  ! !$  write(0,*) 'Check on output restrictor',omp(1:min(size(omp),10))
-  ! Compute omega_int
-  ommx = dzero
-  do i=1, ncol
-    if (ilaggr(i) >0) then 
-      omi(i) = omp(ilaggr(i))
-    else
-      omi(i) = dzero
-    end if
-    if(abs(omi(i)) .gt. abs(ommx)) ommx = omi(i)
-  end do
-  ! Compute omega_fine
-  ! Going over the columns of atmp means going over the rows
-  ! of A^T. Hopefully ;-) 
-  call atmp%cp_to(acsc)
-
-  do i=1, nrow
-    omf(i) = ommx
-    do j= acsc%icp(i),acsc%icp(i+1)-1
-      if(abs(omi(acsc%ia(j))) .lt. abs(omf(i))) omf(i)=omi(acsc%ia(j))
-    end do
-!!$    if(min(real(omf(i)),aimag(omf(i))) < dzero) omf(i) = dzero
-    if(psb_minreal(omf(i)) < dzero) omf(i) = dzero
-  end do
-  omf(1:nrow) = omf(1:nrow)*adinv(1:nrow)
-  call psb_halo(omf,desc_a,info)
-  call acsc%free() 
-
-
-  call atmp%mv_to(acsr1)
-
-  do i=1,acsr1%get_nrows()
-    do j=acsr1%irp(i),acsr1%irp(i+1)-1
-      if (acsr1%ja(j) == i) then 
-        acsr1%val(j) = done - acsr1%val(j)*omf(acsr1%ja(j))
-      else
-        acsr1%val(j) =      - acsr1%val(j)*omf(acsr1%ja(j))
-      end if
-    end do
-  end do
-  call atmp%mv_from(acsr1)
-
-  call rtilde%mv_to(tmpcoo)
-  nzl = tmpcoo%get_nzeros()
-  i=0
-  do k=1, nzl
-    if ((naggrm1 < tmpcoo%ia(k)) .and. (tmpcoo%ia(k) <= naggrp1)) then
-      i = i+1
-      tmpcoo%val(i) = tmpcoo%val(k)
-      tmpcoo%ia(i)  = tmpcoo%ia(k)
-      tmpcoo%ja(i)  = tmpcoo%ja(k)
-    end if
-  end do
-  call tmpcoo%set_nzeros(i)
-  call rtilde%mv_from(tmpcoo)
-  call rtilde%cscnv(info,type='csr')
-
-  call psb_spspmm(rtilde,atmp,op_restr,info)
-
-  !
-  ! Now we have to gather the halo of op_prol, and add it to itself
-  ! to multiply it by A,
-  !
-  call op_prol%clone(tmp_prol,info) 
-  if (info == psb_success_) call psb_sphalo(tmp_prol,desc_a,am4,info,&
-       & colcnv=.false.,rowscale=.true.)
-  if (info == psb_success_) call psb_rwextd(ncol,tmp_prol,info,b=am4)      
-  if (info == psb_success_) call am4%free()
-
-  if(info /= psb_success_) then
-    call psb_errpush(psb_err_internal_error_,name,a_err='Halo of op_prol')
-    goto 9999
-  end if
-
-  !
-  ! Now we have to fix this.  The only rows of B that are correct 
-  ! are those corresponding to "local" aggregates, i.e. indices in ilaggr(:)
-  !
-  call op_restr%mv_to(tmpcoo)
-
-  nzl = tmpcoo%get_nzeros()
-  i=0
-  do k=1, nzl
-    if ((naggrm1 < tmpcoo%ia(k)) .and. (tmpcoo%ia(k) <= naggrp1)) then
-      i = i+1
-      tmpcoo%val(i) = tmpcoo%val(k)
-      tmpcoo%ia(i)  = tmpcoo%ia(k)
-      tmpcoo%ja(i)  = tmpcoo%ja(k)
-    end if
-  end do
-  call tmpcoo%set_nzeros(i)
-  call op_restr%mv_from(tmpcoo)
-  call op_restr%cscnv(info,type='csr')
-
-
-  if (debug_level >= psb_debug_outer_) &
-       & write(debug_unit,*) me,' ',trim(name),&
-       & 'starting sphalo/ rwxtd'
-
-  call psb_spspmm(la,tmp_prol,am3,info)
-  if (debug_level >= psb_debug_outer_) &
-       & write(debug_unit,*) me,' ',trim(name),&
-       & 'Done SPSPMM 2'
-
-  call psb_sphalo(am3,desc_a,am4,info,&
-       & colcnv=.false.,rowscale=.true.)
-  if (info == psb_success_) call psb_rwextd(ncol,am3,info,b=am4)      
-  if (info == psb_success_) call am4%free()
-
-  if(info /= psb_success_) then
-    call psb_errpush(psb_err_internal_error_,name,&
-         & a_err='Extend am3')
-    goto 9999
-  end if
-  if (debug_level >= psb_debug_outer_) &
-       & write(debug_unit,*) me,' ',trim(name),&
-       & 'Done sphalo/ rwxtd'
-
-  call psb_spspmm(op_restr,am3,ac,info)
-  if (info == psb_success_) call am3%free()
-  if (info == psb_success_) call ac%cscnv(info,type='coo',dupl=psb_dupl_add_)
-
-  if (info /= psb_success_) then
-    call psb_errpush(psb_err_internal_error_,name,&
-         &a_err='Build ac = op_restr x am3')
-    goto 9999
-  end if
+!!$  ! Get the diagonal D
+!!$  adiag =  a%get_diag(info)
+!!$  if (info == psb_success_) &
+!!$       & call psb_realloc(ncol,adiag,info)    
+!!$  if (info == psb_success_) &
+!!$       & call psb_halo(adiag,desc_a,info)
+!!$  if (info == psb_success_) call a%cp_to_l(la)
+!!$  if (info /= psb_success_) then
+!!$    call psb_errpush(psb_err_from_subroutine_,name,a_err='sp_getdiag')
+!!$    goto 9999
+!!$  end if
+!!$
+!!$  do i=1,size(adiag)
+!!$    if (adiag(i) /= dzero) then
+!!$      adinv(i) = done / adiag(i)
+!!$    else
+!!$      adinv(i) = done
+!!$    end if
+!!$  end do
+!!$
+!!$
+!!$
+!!$  ! 1. Allocate Ptilde in sparse matrix form 
+!!$  call op_prol%mv_to(tmpcoo)
+!!$  call ptilde%mv_from(tmpcoo)
+!!$  call ptilde%cscnv(info,type='csr')
+!!$
+!!$  if (info == psb_success_) call la%cscnv(am3,info,type='csr',dupl=psb_dupl_add_)
+!!$  if (info == psb_success_) call la%cscnv(da,info,type='csr',dupl=psb_dupl_add_)
+!!$  if (info /= psb_success_) then
+!!$    call psb_errpush(psb_err_from_subroutine_,name,a_err='spcnv')
+!!$    goto 9999
+!!$  end if
+!!$  if (debug_level >= psb_debug_outer_) &
+!!$       & write(debug_unit,*) me,' ',trim(name),&
+!!$       & ' Initial copies done.'
+!!$
+!!$  call da%scal(adinv,info)
+!!$
+!!$  call psb_spspmm(da,ptilde,dap,info)
+!!$
+!!$  if(info /= psb_success_) then
+!!$    call psb_errpush(psb_err_from_subroutine_,name,a_err='spspmm 1')
+!!$    goto 9999
+!!$  end if
+!!$
+!!$  call dap%clone(atmp,info)
+!!$
+!!$  call psb_sphalo(atmp,desc_a,am4,info,&
+!!$       & colcnv=.false.,rowscale=.true.,outfmt='CSR  ')
+!!$  if (info == psb_success_) call psb_rwextd(ncol,atmp,info,b=am4)      
+!!$  if (info == psb_success_) call am4%free()
+!!$
+!!$  call psb_spspmm(da,atmp,dadap,info)
+!!$  call atmp%free()
+!!$
+!!$  !  !$  write(0,*) 'Columns of AP',psb_sp_get_ncols(ap)
+!!$  !  !$  write(0,*) 'Columns of ADAP',psb_sp_get_ncols(adap)
+!!$  call dap%mv_to(csc_dap)
+!!$  call dadap%mv_to(csc_dadap)
+!!$
+!!$  call csc_mat_col_prod(csc_dap,csc_dadap,omp,info)
+!!$  call csc_mat_col_prod(csc_dadap,csc_dadap,oden,info)
+!!$  call psb_sum(ctxt,omp)
+!!$  call psb_sum(ctxt,oden)
+!!$  ! !$  write(0,*) trim(name),' OMP :',omp
+!!$  ! !$  write(0,*) trim(name),' ODEN:',oden
+!!$
+!!$  omp = omp/oden
+!!$
+!!$  ! !$  write(0,*) 'Check on output prolongator ',omp(1:min(size(omp),10))
+!!$  if (debug_level >= psb_debug_outer_) &
+!!$       & write(debug_unit,*) me,' ',trim(name),&
+!!$       & 'Done NUMBMM 1'
+!!$
+!!$  call am3%mv_to(acsr3)
+!!$  ! Compute omega_int
+!!$  ommx = dzero
+!!$  do i=1, ncol
+!!$    if (ilaggr(i) >0) then 
+!!$      omi(i) = omp(ilaggr(i))
+!!$    else
+!!$      omi(i) = dzero
+!!$    end if
+!!$    if(abs(omi(i)) .gt. abs(ommx)) ommx = omi(i)
+!!$  end do
+!!$  ! Compute omega_fine
+!!$  do i=1, nrow
+!!$    omf(i) = ommx
+!!$    do j=acsr3%irp(i),acsr3%irp(i+1)-1
+!!$      if(abs(omi(acsr3%ja(j))) .lt. abs(omf(i))) omf(i)=omi(acsr3%ja(j))
+!!$    end do
+!!$ ! !     if(min(real(omf(i)),aimag(omf(i))) < dzero) omf(i) = dzero
+!!$    if(psb_minreal(omf(i)) < dzero) omf(i) = dzero
+!!$  end do
+!!$
+!!$  omf(1:nrow) = omf(1:nrow) * adinv(1:nrow)
+!!$
+!!$  if (filter_mat) then
+!!$    !
+!!$    ! Build the filtered matrix Af from A
+!!$    ! 
+!!$    call la%cscnv(acsrf,info,dupl=psb_dupl_add_)
+!!$
+!!$    do i=1,nrow
+!!$      tmp = dzero
+!!$      jd  = -1 
+!!$      do j=acsrf%irp(i),acsrf%irp(i+1)-1
+!!$        if (acsrf%ja(j) == i) jd = j 
+!!$        if (abs(acsrf%val(j)) < theta*sqrt(abs(adiag(i)*adiag(acsrf%ja(j))))) then
+!!$          tmp=tmp+acsrf%val(j)
+!!$          acsrf%val(j)=dzero
+!!$        endif
+!!$      enddo
+!!$      if (jd == -1) then 
+!!$        write(0,*) 'Wrong input: we need the diagonal!!!!', i
+!!$      else
+!!$        acsrf%val(jd)=acsrf%val(jd)-tmp
+!!$      end if
+!!$    enddo
+!!$    ! Take out zeroed terms 
+!!$    call acsrf%clean_zeros(info)
+!!$
+!!$    !
+!!$    ! Build the smoothed prolongator using the filtered matrix
+!!$    ! 
+!!$    do i=1,acsrf%get_nrows()
+!!$      do j=acsrf%irp(i),acsrf%irp(i+1)-1
+!!$        if (acsrf%ja(j) == i) then 
+!!$          acsrf%val(j) = done - omf(i)*acsrf%val(j) 
+!!$        else
+!!$          acsrf%val(j) = - omf(i)*acsrf%val(j) 
+!!$        end if
+!!$      end do
+!!$    end do
+!!$
+!!$    if (debug_level >= psb_debug_outer_) &
+!!$         & write(debug_unit,*) me,' ',trim(name),&
+!!$         & 'Done gather, going for SYMBMM 1'
+!!$
+!!$    call af%mv_from(acsrf)
+!!$    !
+!!$    ! op_prol = (I-w*D*Af)Ptilde
+!!$    ! Doing it this way means to consider diag(Af_i)
+!!$    ! 
+!!$    !
+!!$    call psb_spspmm(af,ptilde,op_prol,info)
+!!$    if (debug_level >= psb_debug_outer_) &
+!!$         & write(debug_unit,*) me,' ',trim(name),&
+!!$         & 'Done SPSPMM 1'
+!!$  else
+!!$    !
+!!$    ! Build the smoothed prolongator using the original matrix
+!!$    !
+!!$    do i=1,acsr3%get_nrows()
+!!$      do j=acsr3%irp(i),acsr3%irp(i+1)-1
+!!$        if (acsr3%ja(j) == i) then 
+!!$          acsr3%val(j) = done - omf(i)*acsr3%val(j) 
+!!$        else
+!!$          acsr3%val(j) = - omf(i)*acsr3%val(j) 
+!!$        end if
+!!$      end do
+!!$    end do
+!!$
+!!$    call am3%mv_from(acsr3)
+!!$    if (debug_level >= psb_debug_outer_) &
+!!$         & write(debug_unit,*) me,' ',trim(name),&
+!!$         & 'Done gather, going for SYMBMM 1'
+!!$    !
+!!$    ! 
+!!$    ! op_prol = (I-w*D*A)Ptilde
+!!$    ! 
+!!$    !
+!!$    call psb_spspmm(am3,ptilde,op_prol,info)
+!!$    if (debug_level >= psb_debug_outer_) &
+!!$         & write(debug_unit,*) me,' ',trim(name),&
+!!$         & 'Done NUMBMM 1'
+!!$
+!!$  end if
+!!$
+!!$
+!!$  !
+!!$  ! Ok, let's start over with the restrictor
+!!$  ! 
+!!$  call ptilde%transc(rtilde)
+!!$  call la%cscnv(atmp,info,type='csr')
+!!$  call psb_sphalo(atmp,desc_a,am4,info,&
+!!$       & colcnv=.true.,rowscale=.true.)
+!!$  nrt  = am4%get_nrows() 
+!!$  call am4%csclip(atmp2,info,lone,nrt,lone,ncol)
+!!$  call atmp2%cscnv(info,type='CSR')
+!!$  if (info == psb_success_) call psb_rwextd(ncol,atmp,info,b=atmp2)      
+!!$  call am4%free()
+!!$  call atmp2%free()
+!!$
+!!$  ! This is to compute the transpose. It ONLY works if the
+!!$  ! original A has a symmetric pattern.
+!!$  call atmp%transc(atmp2) 
+!!$  call atmp2%csclip(dat,info,lone,nrow,lone,ncol)
+!!$  call dat%cscnv(info,type='csr')
+!!$  call dat%scal(adinv,info)
+!!$
+!!$  ! Now for the product. 
+!!$  call psb_spspmm(dat,ptilde,datp,info)
+!!$
+!!$  call datp%clone(atmp2,info)
+!!$  call psb_sphalo(atmp2,desc_a,am4,info,&
+!!$       & colcnv=.false.,rowscale=.true.,outfmt='CSR  ')
+!!$  if (info == psb_success_) call psb_rwextd(ncol,atmp2,info,b=am4)      
+!!$  if (info == psb_success_) call am4%free()
+!!$
+!!$
+!!$  call psb_symbmm(dat,atmp2,datdatp,info)
+!!$  call psb_numbmm(dat,atmp2,datdatp)
+!!$  call atmp2%free()
+!!$
+!!$  call datp%mv_to(csc_datp)    
+!!$  call datdatp%mv_to(csc_datdatp)    
+!!$
+!!$  call csc_mat_col_prod(csc_datp,csc_datdatp,omp,info)
+!!$  call csc_mat_col_prod(csc_datdatp,csc_datdatp,oden,info)
+!!$  call psb_sum(ctxt,omp)
+!!$  call psb_sum(ctxt,oden)
+!!$
+!!$
+!!$  ! !$  write(debug_unit,*) trim(name),' OMP_R :',omp
+!!$  ! ! $  write(debug_unit,*) trim(name),' ODEN_R:',oden
+!!$  omp = omp/oden
+!!$  ! !$  write(0,*) 'Check on output restrictor',omp(1:min(size(omp),10))
+!!$  ! Compute omega_int
+!!$  ommx = dzero
+!!$  do i=1, ncol
+!!$    if (ilaggr(i) >0) then 
+!!$      omi(i) = omp(ilaggr(i))
+!!$    else
+!!$      omi(i) = dzero
+!!$    end if
+!!$    if(abs(omi(i)) .gt. abs(ommx)) ommx = omi(i)
+!!$  end do
+!!$  ! Compute omega_fine
+!!$  ! Going over the columns of atmp means going over the rows
+!!$  ! of A^T. Hopefully ;-) 
+!!$  call atmp%cp_to(acsc)
+!!$
+!!$  do i=1, nrow
+!!$    omf(i) = ommx
+!!$    do j= acsc%icp(i),acsc%icp(i+1)-1
+!!$      if(abs(omi(acsc%ia(j))) .lt. abs(omf(i))) omf(i)=omi(acsc%ia(j))
+!!$    end do
+!!$ ! !    if(min(real(omf(i)),aimag(omf(i))) < dzero) omf(i) = dzero
+!!$    if(psb_minreal(omf(i)) < dzero) omf(i) = dzero
+!!$  end do
+!!$  omf(1:nrow) = omf(1:nrow)*adinv(1:nrow)
+!!$  call psb_halo(omf,desc_a,info)
+!!$  call acsc%free() 
+!!$
+!!$
+!!$  call atmp%mv_to(acsr1)
+!!$
+!!$  do i=1,acsr1%get_nrows()
+!!$    do j=acsr1%irp(i),acsr1%irp(i+1)-1
+!!$      if (acsr1%ja(j) == i) then 
+!!$        acsr1%val(j) = done - acsr1%val(j)*omf(acsr1%ja(j))
+!!$      else
+!!$        acsr1%val(j) =      - acsr1%val(j)*omf(acsr1%ja(j))
+!!$      end if
+!!$    end do
+!!$  end do
+!!$  call atmp%mv_from(acsr1)
+!!$
+!!$  call rtilde%mv_to(tmpcoo)
+!!$  nzl = tmpcoo%get_nzeros()
+!!$  i=0
+!!$  do k=1, nzl
+!!$    if ((naggrm1 < tmpcoo%ia(k)) .and. (tmpcoo%ia(k) <= naggrp1)) then
+!!$      i = i+1
+!!$      tmpcoo%val(i) = tmpcoo%val(k)
+!!$      tmpcoo%ia(i)  = tmpcoo%ia(k)
+!!$      tmpcoo%ja(i)  = tmpcoo%ja(k)
+!!$    end if
+!!$  end do
+!!$  call tmpcoo%set_nzeros(i)
+!!$  call rtilde%mv_from(tmpcoo)
+!!$  call rtilde%cscnv(info,type='csr')
+!!$
+!!$  call psb_spspmm(rtilde,atmp,op_restr,info)
+!!$
+!!$  !
+!!$  ! Now we have to gather the halo of op_prol, and add it to itself
+!!$  ! to multiply it by A,
+!!$  !
+!!$  call op_prol%clone(tmp_prol,info) 
+!!$  if (info == psb_success_) call psb_sphalo(tmp_prol,desc_a,am4,info,&
+!!$       & colcnv=.false.,rowscale=.true.)
+!!$  if (info == psb_success_) call psb_rwextd(ncol,tmp_prol,info,b=am4)      
+!!$  if (info == psb_success_) call am4%free()
+!!$
+!!$  if(info /= psb_success_) then
+!!$    call psb_errpush(psb_err_internal_error_,name,a_err='Halo of op_prol')
+!!$    goto 9999
+!!$  end if
+!!$
+!!$  !
+!!$  ! Now we have to fix this.  The only rows of B that are correct 
+!!$  ! are those corresponding to "local" aggregates, i.e. indices in ilaggr(:)
+!!$  !
+!!$  call op_restr%mv_to(tmpcoo)
+!!$
+!!$  nzl = tmpcoo%get_nzeros()
+!!$  i=0
+!!$  do k=1, nzl
+!!$    if ((naggrm1 < tmpcoo%ia(k)) .and. (tmpcoo%ia(k) <= naggrp1)) then
+!!$      i = i+1
+!!$      tmpcoo%val(i) = tmpcoo%val(k)
+!!$      tmpcoo%ia(i)  = tmpcoo%ia(k)
+!!$      tmpcoo%ja(i)  = tmpcoo%ja(k)
+!!$    end if
+!!$  end do
+!!$  call tmpcoo%set_nzeros(i)
+!!$  call op_restr%mv_from(tmpcoo)
+!!$  call op_restr%cscnv(info,type='csr')
+!!$
+!!$
+!!$  if (debug_level >= psb_debug_outer_) &
+!!$       & write(debug_unit,*) me,' ',trim(name),&
+!!$       & 'starting sphalo/ rwxtd'
+!!$
+!!$  call psb_spspmm(la,tmp_prol,am3,info)
+!!$  if (debug_level >= psb_debug_outer_) &
+!!$       & write(debug_unit,*) me,' ',trim(name),&
+!!$       & 'Done SPSPMM 2'
+!!$
+!!$  call psb_sphalo(am3,desc_a,am4,info,&
+!!$       & colcnv=.false.,rowscale=.true.)
+!!$  if (info == psb_success_) call psb_rwextd(ncol,am3,info,b=am4)      
+!!$  if (info == psb_success_) call am4%free()
+!!$
+!!$  if(info /= psb_success_) then
+!!$    call psb_errpush(psb_err_internal_error_,name,&
+!!$         & a_err='Extend am3')
+!!$    goto 9999
+!!$  end if
+!!$  if (debug_level >= psb_debug_outer_) &
+!!$       & write(debug_unit,*) me,' ',trim(name),&
+!!$       & 'Done sphalo/ rwxtd'
+!!$
+!!$  call psb_spspmm(op_restr,am3,ac,info)
+!!$  if (info == psb_success_) call am3%free()
+!!$  if (info == psb_success_) call ac%cscnv(info,type='coo',dupl=psb_dupl_add_)
+!!$
+!!$  if (info /= psb_success_) then
+!!$    call psb_errpush(psb_err_internal_error_,name,&
+!!$         &a_err='Build ac = op_restr x am3')
+!!$    goto 9999
+!!$  end if



@@ -116,7 +116,7 @@ subroutine amg_daggrmat_smth_bld(a,desc_a,ilaggr,nlaggr,parms,&
  type(psb_desc_type), intent(inout)       :: desc_a
  integer(psb_lpk_), intent(inout)         :: ilaggr(:), nlaggr(:)
  type(amg_dml_parms), intent(inout)      :: parms 
-  type(psb_dspmat_type), intent(out)      :: op_prol,ac,op_restr
+  type(psb_dspmat_type), intent(inout)      :: op_prol,ac,op_restr
  type(psb_ldspmat_type), intent(inout)    :: t_prol
  type(psb_desc_type), intent(inout)       :: desc_ac
  integer(psb_ipk_), intent(out)           :: info
@@ -83,8 +83,8 @@ subroutine  amg_s_dec_aggregator_build_tprol(ag,parms,ag_data,&
  class(amg_s_dec_aggregator_type), target, intent(inout) :: ag
  type(amg_sml_parms), intent(inout)  :: parms
  type(amg_saggr_data), intent(in)    :: ag_data
-  type(psb_sspmat_type), intent(in)   :: a
-  type(psb_desc_type), intent(in)     :: desc_a
+  type(psb_sspmat_type), intent(inout) :: a
+  type(psb_desc_type), intent(inout)    :: desc_a
  integer(psb_lpk_), allocatable, intent(out) :: ilaggr(:),nlaggr(:)
  type(psb_lsspmat_type), intent(out)  :: t_prol
  integer(psb_ipk_), intent(out)      :: info
@@ -48,7 +48,7 @@ subroutine  amg_s_parmatch_aggregator_build_tprol(ag,parms,ag_data,&
  use amg_base_prec_type
  use amg_s_inner_mod
 #if defined(SERIAL_MPI)
-    use amg_s_parmatch_aggregator_mod
+  use amg_s_parmatch_aggregator_mod
 #else
  use amg_s_parmatch_aggregator_mod, amg_protect_name => amg_s_parmatch_aggregator_build_tprol
 #endif
@@ -58,7 +58,7 @@ subroutine  amg_s_parmatch_aggregator_build_tprol(ag,parms,ag_data,&
  type(amg_sml_parms), intent(inout)   :: parms
  type(amg_saggr_data), intent(in)     :: ag_data
  type(psb_sspmat_type), intent(inout) :: a
-  type(psb_desc_type), intent(inout)   :: desc_a
+  type(psb_desc_type), intent(inout)      :: desc_a
  integer(psb_lpk_), allocatable, intent(out) :: ilaggr(:),nlaggr(:)
  type(psb_lsspmat_type), intent(out)  :: t_prol
  integer(psb_ipk_), intent(out)      :: info
@@ -68,7 +68,8 @@ subroutine  amg_s_parmatch_aggregator_build_tprol(ag,parms,ag_data,&
  real(psb_spk_), allocatable    :: tmpw(:), tmpwnxt(:)
  integer(psb_lpk_), allocatable :: ixaggr(:), nxaggr(:), tlaggr(:), ivr(:)
  type(psb_sspmat_type)          :: a_tmp
-  integer(c_int) :: match_algorithm, n_sweeps, max_csize, max_nlevels
+  integer(psb_ipk_)    :: match_algorithm, n_sweeps
+  integer(psb_lpk_)    :: target_csize
  character(len=40)    :: name, ch_err
  character(len=80)    :: fname, prefix_
  type(psb_ctxt_type)  :: ictxt
@@ -128,27 +129,22 @@ subroutine  amg_s_parmatch_aggregator_build_tprol(ag,parms,ag_data,&
      write(debug_unit, *) 'Warning: AGGR_SIZE reset to value ',2**n_sweeps
    end if
  end if
-  if (ag%max_csize > 0) then
-    max_csize       = ag%max_csize
+  if (ag_data%target_coarse_size > 0) then
+    target_csize       = ag_data%target_coarse_size
  else
-    max_csize       = ag_data%min_coarse_size
-  end if
-  if (ag%max_nlevels > 0) then
-    max_nlevels     = ag%max_nlevels
-  else
-    max_nlevels = ag_data%max_levs
+    target_csize       = ag_data%min_coarse_size
  end if
  if (.true.) then
    block
      integer(psb_ipk_) :: ipv(2)
-      ipv(1) = max_csize
+      ipv(1) = target_csize
      ipv(2) = n_sweeps
      call psb_bcast(ictxt,ipv)
-      max_csize = ipv(1)
+      target_csize = ipv(1)
      n_sweeps  = ipv(2)
    end block
  else
-    call psb_bcast(ictxt,max_csize)
+    call psb_bcast(ictxt,target_csize)
    call psb_bcast(ictxt,n_sweeps)
  end if
  if (n_sweeps /= ag%n_sweeps) then
@@ -156,7 +152,7 @@ subroutine  amg_s_parmatch_aggregator_build_tprol(ag,parms,ag_data,&
  end if
 !!$  if (me==0) write(0,*) 'Matching sweeps: ',n_sweeps
  n_sweeps = max(1,n_sweeps)
-  if (debug) write(0,*) me,' Copies, with n_sweeps: ',n_sweeps,max_csize
+  if (debug) write(0,*) me,' Copies, with n_sweeps: ',n_sweeps,target_csize
  if (ag%unsmoothed_hierarchy.and.allocated(ag%base_a)) then
    call ag%base_a%cp_to(acsr)
    if (ag%do_clean_zeros) call acsr%clean_zeros(info)
@@ -242,7 +238,7 @@ subroutine  amg_s_parmatch_aggregator_build_tprol(ag,parms,ag_data,&

  if (debug) then
    call psb_barrier(ictxt)
-    if (me == 0) write(0,*) 'N_sweeps ',n_sweeps,nr,desc_acv(0)%is_ok(),max_csize
+    if (me == 0) write(0,*) 'N_sweeps ',n_sweeps,nr,desc_acv(0)%is_ok(),target_csize
  end if

  !
@@ -264,7 +260,7 @@ subroutine  amg_s_parmatch_aggregator_build_tprol(ag,parms,ag_data,&
    !
    if (debug) write(0,*) me,' Into matchbox_build_prol ',info
    if (do_timings) call psb_tic(idx_mboxp)
-    call smatchboxp_build_prol(tmpw,acv(i-1),desc_acv(i-1),ixaggr,nxaggr,tmp_prol,info,&
+    call amg_s_matchboxp_build_prol(tmpw,acv(i-1),desc_acv(i-1),ixaggr,nxaggr,tmp_prol,info,&
         & symmetrize=ag%need_symmetrize,reproducible=ag%reproducible_matching)
    if (do_timings) call psb_toc(idx_mboxp)
    if (debug) write(0,*) me,' Out from matchbox_build_prol ',info
@@ -300,11 +296,11 @@ subroutine  amg_s_parmatch_aggregator_build_tprol(ag,parms,ag_data,&

    if (debug) then
      call psb_barrier(ictxt)
-      if (me==0) write(0,*) me,trim(name),' Done mat_asb:',i,sum(nxaggr),max_csize,info
+      if (me==0) write(0,*) me,trim(name),' Done mat_asb:',i,sum(nxaggr),target_csize,info
      csz = sum(nxaggr)
      call psb_bcast(ictxt,csz)
      if (csz /= sum(nxaggr)) write(0,*) me,trim(name),' Mismatch matasb',&
-           & csz,sum(nxaggr),max_csize
+           & csz,sum(nxaggr),target_csize
    end if
    if (psb_errstatus_fatal())  write(0,*)me,trim(name),'Error fatal on entry to tmpwnxt 2'

@@ -342,10 +338,10 @@ subroutine  amg_s_parmatch_aggregator_build_tprol(ag,parms,ag_data,&
    call move_alloc(tmpwnxt,tmpw)
    if (debug) then
      if (csz /= sum(nlaggr)) write(0,*) me,trim(name),' Mismatch 2 matasb',&
-           & csz,sum(nlaggr),max_csize, info
+           & csz,sum(nlaggr),target_csize, info
    end if
    call acv(i-1)%free()
-    if ((sum(nlaggr) <= max_csize).or.(any(nlaggr==0))) then
+    if ((sum(nlaggr) <= target_csize).or.(any(nlaggr==0))) then
      x_sweeps = i
      exit sweeps_loop
    end if
@@ -122,7 +122,7 @@ subroutine amg_s_parmatch_smth_bld(ag,a,desc_a,ilaggr,nlaggr,parms,&
  integer(psb_lpk_), intent(inout)         :: ilaggr(:), nlaggr(:)
  type(amg_sml_parms), intent(inout)    :: parms
  type(psb_lsspmat_type), intent(inout)  :: t_prol
-  type(psb_sspmat_type), intent(out)    :: op_prol,ac,op_restr
+  type(psb_sspmat_type), intent(inout)   :: op_prol,ac,op_restr
  type(psb_desc_type), intent(inout)    :: desc_ac
  integer(psb_ipk_), intent(out)           :: info

@@ -108,7 +108,7 @@ subroutine amg_s_parmatch_spmm_bld(a,desc_a,ilaggr,nlaggr,parms,&

  ! Arguments
  type(psb_sspmat_type), intent(in)       :: a
-  type(psb_desc_type), intent(in)         :: desc_a
+  type(psb_desc_type), intent(inout)       :: desc_a
  integer(psb_lpk_), intent(inout)        :: ilaggr(:), nlaggr(:)
  type(amg_sml_parms), intent(inout)      :: parms
  type(psb_lsspmat_type), intent(inout)   :: t_prol
@@ -108,11 +108,11 @@ subroutine amg_s_parmatch_spmm_bld_inner(a_csr,desc_a,ilaggr,nlaggr,parms,&

  ! Arguments
  type(psb_s_csr_sparse_mat), intent(inout) :: a_csr
-  type(psb_desc_type), intent(in)           :: desc_a
+  type(psb_desc_type), intent(inout)          :: desc_a
  integer(psb_lpk_), intent(inout)          :: ilaggr(:), nlaggr(:)
  type(amg_sml_parms), intent(inout)        :: parms
  type(psb_lsspmat_type), intent(inout)     :: t_prol
-  type(psb_sspmat_type), intent(out)        :: ac, op_prol, op_restr
+  type(psb_sspmat_type), intent(inout)      :: ac, op_prol, op_restr
  type(psb_desc_type), intent(out)          :: desc_ac
  integer(psb_ipk_), intent(out)            :: info

@@ -108,7 +108,7 @@ subroutine amg_s_parmatch_spmm_bld_ov(a,desc_a,ilaggr,nlaggr,parms,&

  ! Arguments
  type(psb_sspmat_type), intent(inout)    :: a
-  type(psb_desc_type), intent(in)         :: desc_a
+  type(psb_desc_type), intent(inout)       :: desc_a
  integer(psb_lpk_), intent(inout)        :: ilaggr(:), nlaggr(:)
  type(amg_sml_parms), intent(inout)      :: parms
  type(psb_lsspmat_type), intent(inout)   :: t_prol
@@ -86,8 +86,8 @@ subroutine  amg_s_symdec_aggregator_build_tprol(ag,parms,ag_data,&
  class(amg_s_symdec_aggregator_type), target, intent(inout) :: ag
  type(amg_sml_parms), intent(inout)  :: parms
  type(amg_saggr_data), intent(in)    :: ag_data
-  type(psb_sspmat_type), intent(in)   :: a
-  type(psb_desc_type), intent(in)     :: desc_a
+  type(psb_sspmat_type), intent(inout) :: a
+  type(psb_desc_type), intent(inout)   :: desc_a
  integer(psb_lpk_), allocatable, intent(out) :: ilaggr(:),nlaggr(:)
  type(psb_lsspmat_type), intent(out)  :: op_prol
  integer(psb_ipk_), intent(out)      :: info
@@ -105,7 +105,7 @@
 !
 !
 subroutine amg_saggrmat_minnrg_bld(a,desc_a,ilaggr,nlaggr,parms,&
-     & ac,desc_ac,op_prol,op_restr,info)
+     & ac,desc_ac,op_prol,op_restr,t_prol,info)
  use psb_base_mod
  use amg_base_prec_type
  use amg_s_inner_mod, amg_protect_name => amg_saggrmat_minnrg_bld
@@ -117,8 +117,8 @@ subroutine amg_saggrmat_minnrg_bld(a,desc_a,ilaggr,nlaggr,parms,&
  type(psb_desc_type), intent(inout)            :: desc_a
  integer(psb_lpk_), intent(inout)              :: ilaggr(:), nlaggr(:)
  type(amg_sml_parms), intent(inout)         :: parms 
-  type(psb_lsspmat_type), intent(inout)       :: op_prol
-  type(psb_lsspmat_type), intent(out)         :: ac,op_restr
+  type(psb_lsspmat_type), intent(inout)       :: t_prol
+  type(psb_sspmat_type), intent(inout)       :: op_prol, ac,op_restr
  type(psb_desc_type), intent(inout)            :: desc_ac
  integer(psb_ipk_), intent(out)                :: info

@@ -171,6 +171,8 @@ subroutine amg_saggrmat_minnrg_bld(a,desc_a,ilaggr,nlaggr,parms,&

  filter_mat = (parms%aggr_filter == amg_filter_mat_)

+  !NEEDS TO BE REWORKED !!
+  
  ! naggr: number of local aggregates
  ! nrow: local rows. 
  ! 
@@ -183,361 +185,361 @@ subroutine amg_saggrmat_minnrg_bld(a,desc_a,ilaggr,nlaggr,parms,&
    goto 9999      
  end if

-  ! Get the diagonal D
-  adiag =  a%get_diag(info)
-  if (info == psb_success_) &
-       & call psb_realloc(ncol,adiag,info)    
-  if (info == psb_success_) &
-       & call psb_halo(adiag,desc_a,info)
-  if (info == psb_success_) call a%cp_to_l(la)
-  if (info /= psb_success_) then
-    call psb_errpush(psb_err_from_subroutine_,name,a_err='sp_getdiag')
-    goto 9999
-  end if
-
-  do i=1,size(adiag)
-    if (adiag(i) /= szero) then
-      adinv(i) = sone / adiag(i)
-    else
-      adinv(i) = sone
-    end if
-  end do
-
-
-
-  ! 1. Allocate Ptilde in sparse matrix form 
-  call op_prol%mv_to(tmpcoo)
-  call ptilde%mv_from(tmpcoo)
-  call ptilde%cscnv(info,type='csr')
-
-  if (info == psb_success_) call la%cscnv(am3,info,type='csr',dupl=psb_dupl_add_)
-  if (info == psb_success_) call la%cscnv(da,info,type='csr',dupl=psb_dupl_add_)
-  if (info /= psb_success_) then
-    call psb_errpush(psb_err_from_subroutine_,name,a_err='spcnv')
-    goto 9999
-  end if
-  if (debug_level >= psb_debug_outer_) &
-       & write(debug_unit,*) me,' ',trim(name),&
-       & ' Initial copies done.'
-
-  call da%scal(adinv,info)
-
-  call psb_spspmm(da,ptilde,dap,info)
-
-  if(info /= psb_success_) then
-    call psb_errpush(psb_err_from_subroutine_,name,a_err='spspmm 1')
-    goto 9999
-  end if
-
-  call dap%clone(atmp,info)
-
-  call psb_sphalo(atmp,desc_a,am4,info,&
-       & colcnv=.false.,rowscale=.true.,outfmt='CSR  ')
-  if (info == psb_success_) call psb_rwextd(ncol,atmp,info,b=am4)      
-  if (info == psb_success_) call am4%free()
-
-  call psb_spspmm(da,atmp,dadap,info)
-  call atmp%free()
-
-  !  !$  write(0,*) 'Columns of AP',psb_sp_get_ncols(ap)
-  !  !$  write(0,*) 'Columns of ADAP',psb_sp_get_ncols(adap)
-  call dap%mv_to(csc_dap)
-  call dadap%mv_to(csc_dadap)
-
-  call csc_mat_col_prod(csc_dap,csc_dadap,omp,info)
-  call csc_mat_col_prod(csc_dadap,csc_dadap,oden,info)
-  call psb_sum(ctxt,omp)
-  call psb_sum(ctxt,oden)
-  ! !$  write(0,*) trim(name),' OMP :',omp
-  ! !$  write(0,*) trim(name),' ODEN:',oden
-
-  omp = omp/oden
-
-  ! !$  write(0,*) 'Check on output prolongator ',omp(1:min(size(omp),10))
-  if (debug_level >= psb_debug_outer_) &
-       & write(debug_unit,*) me,' ',trim(name),&
-       & 'Done NUMBMM 1'
-
-  call am3%mv_to(acsr3)
-  ! Compute omega_int
-  ommx = szero
-  do i=1, ncol
-    if (ilaggr(i) >0) then 
-      omi(i) = omp(ilaggr(i))
-    else
-      omi(i) = szero
-    end if
-    if(abs(omi(i)) .gt. abs(ommx)) ommx = omi(i)
-  end do
-  ! Compute omega_fine
-  do i=1, nrow
-    omf(i) = ommx
-    do j=acsr3%irp(i),acsr3%irp(i+1)-1
-      if(abs(omi(acsr3%ja(j))) .lt. abs(omf(i))) omf(i)=omi(acsr3%ja(j))
-    end do
-!!$    if(min(real(omf(i)),aimag(omf(i))) < szero) omf(i) = szero
-    if(psb_minreal(omf(i)) < szero) omf(i) = szero
-  end do
-
-  omf(1:nrow) = omf(1:nrow) * adinv(1:nrow)
-
-  if (filter_mat) then
-    !
-    ! Build the filtered matrix Af from A
-    ! 
-    call la%cscnv(acsrf,info,dupl=psb_dupl_add_)
-
-    do i=1,nrow
-      tmp = szero
-      jd  = -1 
-      do j=acsrf%irp(i),acsrf%irp(i+1)-1
-        if (acsrf%ja(j) == i) jd = j 
-        if (abs(acsrf%val(j)) < theta*sqrt(abs(adiag(i)*adiag(acsrf%ja(j))))) then
-          tmp=tmp+acsrf%val(j)
-          acsrf%val(j)=szero
-        endif
-      enddo
-      if (jd == -1) then 
-        write(0,*) 'Wrong input: we need the diagonal!!!!', i
-      else
-        acsrf%val(jd)=acsrf%val(jd)-tmp
-      end if
-    enddo
-    ! Take out zeroed terms 
-    call acsrf%clean_zeros(info)
-
-    !
-    ! Build the smoothed prolongator using the filtered matrix
-    ! 
-    do i=1,acsrf%get_nrows()
-      do j=acsrf%irp(i),acsrf%irp(i+1)-1
-        if (acsrf%ja(j) == i) then 
-          acsrf%val(j) = sone - omf(i)*acsrf%val(j) 
-        else
-          acsrf%val(j) = - omf(i)*acsrf%val(j) 
-        end if
-      end do
-    end do
-
-    if (debug_level >= psb_debug_outer_) &
-         & write(debug_unit,*) me,' ',trim(name),&
-         & 'Done gather, going for SYMBMM 1'
-
-    call af%mv_from(acsrf)
-    !
-    ! op_prol = (I-w*D*Af)Ptilde
-    ! Doing it this way means to consider diag(Af_i)
-    ! 
-    !
-    call psb_spspmm(af,ptilde,op_prol,info)
-    if (debug_level >= psb_debug_outer_) &
-         & write(debug_unit,*) me,' ',trim(name),&
-         & 'Done SPSPMM 1'
-  else
-    !
-    ! Build the smoothed prolongator using the original matrix
-    !
-    do i=1,acsr3%get_nrows()
-      do j=acsr3%irp(i),acsr3%irp(i+1)-1
-        if (acsr3%ja(j) == i) then 
-          acsr3%val(j) = sone - omf(i)*acsr3%val(j) 
-        else
-          acsr3%val(j) = - omf(i)*acsr3%val(j) 
-        end if
-      end do
-    end do
-
-    call am3%mv_from(acsr3)
-    if (debug_level >= psb_debug_outer_) &
-         & write(debug_unit,*) me,' ',trim(name),&
-         & 'Done gather, going for SYMBMM 1'
-    !
-    ! 
-    ! op_prol = (I-w*D*A)Ptilde
-    ! 
-    !
-    call psb_spspmm(am3,ptilde,op_prol,info)
-    if (debug_level >= psb_debug_outer_) &
-         & write(debug_unit,*) me,' ',trim(name),&
-         & 'Done NUMBMM 1'
-
-  end if
-
-
-  !
-  ! Ok, let's start over with the restrictor
-  ! 
-  call ptilde%transc(rtilde)
-  call la%cscnv(atmp,info,type='csr')
-  call psb_sphalo(atmp,desc_a,am4,info,&
-       & colcnv=.true.,rowscale=.true.)
-  nrt  = am4%get_nrows() 
-  call am4%csclip(atmp2,info,lone,nrt,lone,ncol)
-  call atmp2%cscnv(info,type='CSR')
-  if (info == psb_success_) call psb_rwextd(ncol,atmp,info,b=atmp2)      
-  call am4%free()
-  call atmp2%free()
-
-  ! This is to compute the transpose. It ONLY works if the
-  ! original A has a symmetric pattern.
-  call atmp%transc(atmp2) 
-  call atmp2%csclip(dat,info,lone,nrow,lone,ncol)
-  call dat%cscnv(info,type='csr')
-  call dat%scal(adinv,info)
-
-  ! Now for the product. 
-  call psb_spspmm(dat,ptilde,datp,info)
-
-  call datp%clone(atmp2,info)
-  call psb_sphalo(atmp2,desc_a,am4,info,&
-       & colcnv=.false.,rowscale=.true.,outfmt='CSR  ')
-  if (info == psb_success_) call psb_rwextd(ncol,atmp2,info,b=am4)      
-  if (info == psb_success_) call am4%free()
-
-
-  call psb_symbmm(dat,atmp2,datdatp,info)
-  call psb_numbmm(dat,atmp2,datdatp)
-  call atmp2%free()
-
-  call datp%mv_to(csc_datp)    
-  call datdatp%mv_to(csc_datdatp)    
-
-  call csc_mat_col_prod(csc_datp,csc_datdatp,omp,info)
-  call csc_mat_col_prod(csc_datdatp,csc_datdatp,oden,info)
-  call psb_sum(ctxt,omp)
-  call psb_sum(ctxt,oden)
-
-
-  ! !$  write(debug_unit,*) trim(name),' OMP_R :',omp
-  ! ! $  write(debug_unit,*) trim(name),' ODEN_R:',oden
-  omp = omp/oden
-  ! !$  write(0,*) 'Check on output restrictor',omp(1:min(size(omp),10))
-  ! Compute omega_int
-  ommx = szero
-  do i=1, ncol
-    if (ilaggr(i) >0) then 
-      omi(i) = omp(ilaggr(i))
-    else
-      omi(i) = szero
-    end if
-    if(abs(omi(i)) .gt. abs(ommx)) ommx = omi(i)
-  end do
-  ! Compute omega_fine
-  ! Going over the columns of atmp means going over the rows
-  ! of A^T. Hopefully ;-) 
-  call atmp%cp_to(acsc)
-
-  do i=1, nrow
-    omf(i) = ommx
-    do j= acsc%icp(i),acsc%icp(i+1)-1
-      if(abs(omi(acsc%ia(j))) .lt. abs(omf(i))) omf(i)=omi(acsc%ia(j))
-    end do
-!!$    if(min(real(omf(i)),aimag(omf(i))) < szero) omf(i) = szero
-    if(psb_minreal(omf(i)) < szero) omf(i) = szero
-  end do
-  omf(1:nrow) = omf(1:nrow)*adinv(1:nrow)
-  call psb_halo(omf,desc_a,info)
-  call acsc%free() 
-
-
-  call atmp%mv_to(acsr1)
-
-  do i=1,acsr1%get_nrows()
-    do j=acsr1%irp(i),acsr1%irp(i+1)-1
-      if (acsr1%ja(j) == i) then 
-        acsr1%val(j) = sone - acsr1%val(j)*omf(acsr1%ja(j))
-      else
-        acsr1%val(j) =      - acsr1%val(j)*omf(acsr1%ja(j))
-      end if
-    end do
-  end do
-  call atmp%mv_from(acsr1)
-
-  call rtilde%mv_to(tmpcoo)
-  nzl = tmpcoo%get_nzeros()
-  i=0
-  do k=1, nzl
-    if ((naggrm1 < tmpcoo%ia(k)) .and. (tmpcoo%ia(k) <= naggrp1)) then
-      i = i+1
-      tmpcoo%val(i) = tmpcoo%val(k)
-      tmpcoo%ia(i)  = tmpcoo%ia(k)
-      tmpcoo%ja(i)  = tmpcoo%ja(k)
-    end if
-  end do
-  call tmpcoo%set_nzeros(i)
-  call rtilde%mv_from(tmpcoo)
-  call rtilde%cscnv(info,type='csr')
-
-  call psb_spspmm(rtilde,atmp,op_restr,info)
-
-  !
-  ! Now we have to gather the halo of op_prol, and add it to itself
-  ! to multiply it by A,
-  !
-  call op_prol%clone(tmp_prol,info) 
-  if (info == psb_success_) call psb_sphalo(tmp_prol,desc_a,am4,info,&
-       & colcnv=.false.,rowscale=.true.)
-  if (info == psb_success_) call psb_rwextd(ncol,tmp_prol,info,b=am4)      
-  if (info == psb_success_) call am4%free()
-
-  if(info /= psb_success_) then
-    call psb_errpush(psb_err_internal_error_,name,a_err='Halo of op_prol')
-    goto 9999
-  end if
-
-  !
-  ! Now we have to fix this.  The only rows of B that are correct 
-  ! are those corresponding to "local" aggregates, i.e. indices in ilaggr(:)
-  !
-  call op_restr%mv_to(tmpcoo)
-
-  nzl = tmpcoo%get_nzeros()
-  i=0
-  do k=1, nzl
-    if ((naggrm1 < tmpcoo%ia(k)) .and. (tmpcoo%ia(k) <= naggrp1)) then
-      i = i+1
-      tmpcoo%val(i) = tmpcoo%val(k)
-      tmpcoo%ia(i)  = tmpcoo%ia(k)
-      tmpcoo%ja(i)  = tmpcoo%ja(k)
-    end if
-  end do
-  call tmpcoo%set_nzeros(i)
-  call op_restr%mv_from(tmpcoo)
-  call op_restr%cscnv(info,type='csr')
-
-
-  if (debug_level >= psb_debug_outer_) &
-       & write(debug_unit,*) me,' ',trim(name),&
-       & 'starting sphalo/ rwxtd'
-
-  call psb_spspmm(la,tmp_prol,am3,info)
-  if (debug_level >= psb_debug_outer_) &
-       & write(debug_unit,*) me,' ',trim(name),&
-       & 'Done SPSPMM 2'
-
-  call psb_sphalo(am3,desc_a,am4,info,&
-       & colcnv=.false.,rowscale=.true.)
-  if (info == psb_success_) call psb_rwextd(ncol,am3,info,b=am4)      
-  if (info == psb_success_) call am4%free()
-
-  if(info /= psb_success_) then
-    call psb_errpush(psb_err_internal_error_,name,&
-         & a_err='Extend am3')
-    goto 9999
-  end if
-  if (debug_level >= psb_debug_outer_) &
-       & write(debug_unit,*) me,' ',trim(name),&
-       & 'Done sphalo/ rwxtd'
-
-  call psb_spspmm(op_restr,am3,ac,info)
-  if (info == psb_success_) call am3%free()
-  if (info == psb_success_) call ac%cscnv(info,type='coo',dupl=psb_dupl_add_)
-
-  if (info /= psb_success_) then
-    call psb_errpush(psb_err_internal_error_,name,&
-         &a_err='Build ac = op_restr x am3')
-    goto 9999
-  end if
+!!$  ! Get the diagonal D
+!!$  adiag =  a%get_diag(info)
+!!$  if (info == psb_success_) &
+!!$       & call psb_realloc(ncol,adiag,info)    
+!!$  if (info == psb_success_) &
+!!$       & call psb_halo(adiag,desc_a,info)
+!!$  if (info == psb_success_) call a%cp_to_l(la)
+!!$  if (info /= psb_success_) then
+!!$    call psb_errpush(psb_err_from_subroutine_,name,a_err='sp_getdiag')
+!!$    goto 9999
+!!$  end if
+!!$
+!!$  do i=1,size(adiag)
+!!$    if (adiag(i) /= szero) then
+!!$      adinv(i) = sone / adiag(i)
+!!$    else
+!!$      adinv(i) = sone
+!!$    end if
+!!$  end do
+!!$
+!!$
+!!$
+!!$  ! 1. Allocate Ptilde in sparse matrix form 
+!!$  call op_prol%mv_to(tmpcoo)
+!!$  call ptilde%mv_from(tmpcoo)
+!!$  call ptilde%cscnv(info,type='csr')
+!!$
+!!$  if (info == psb_success_) call la%cscnv(am3,info,type='csr',dupl=psb_dupl_add_)
+!!$  if (info == psb_success_) call la%cscnv(da,info,type='csr',dupl=psb_dupl_add_)
+!!$  if (info /= psb_success_) then
+!!$    call psb_errpush(psb_err_from_subroutine_,name,a_err='spcnv')
+!!$    goto 9999
+!!$  end if
+!!$  if (debug_level >= psb_debug_outer_) &
+!!$       & write(debug_unit,*) me,' ',trim(name),&
+!!$       & ' Initial copies done.'
+!!$
+!!$  call da%scal(adinv,info)
+!!$
+!!$  call psb_spspmm(da,ptilde,dap,info)
+!!$
+!!$  if(info /= psb_success_) then
+!!$    call psb_errpush(psb_err_from_subroutine_,name,a_err='spspmm 1')
+!!$    goto 9999
+!!$  end if
+!!$
+!!$  call dap%clone(atmp,info)
+!!$
+!!$  call psb_sphalo(atmp,desc_a,am4,info,&
+!!$       & colcnv=.false.,rowscale=.true.,outfmt='CSR  ')
+!!$  if (info == psb_success_) call psb_rwextd(ncol,atmp,info,b=am4)      
+!!$  if (info == psb_success_) call am4%free()
+!!$
+!!$  call psb_spspmm(da,atmp,dadap,info)
+!!$  call atmp%free()
+!!$
+!!$  !  !$  write(0,*) 'Columns of AP',psb_sp_get_ncols(ap)
+!!$  !  !$  write(0,*) 'Columns of ADAP',psb_sp_get_ncols(adap)
+!!$  call dap%mv_to(csc_dap)
+!!$  call dadap%mv_to(csc_dadap)
+!!$
+!!$  call csc_mat_col_prod(csc_dap,csc_dadap,omp,info)
+!!$  call csc_mat_col_prod(csc_dadap,csc_dadap,oden,info)
+!!$  call psb_sum(ctxt,omp)
+!!$  call psb_sum(ctxt,oden)
+!!$  ! !$  write(0,*) trim(name),' OMP :',omp
+!!$  ! !$  write(0,*) trim(name),' ODEN:',oden
+!!$
+!!$  omp = omp/oden
+!!$
+!!$  ! !$  write(0,*) 'Check on output prolongator ',omp(1:min(size(omp),10))
+!!$  if (debug_level >= psb_debug_outer_) &
+!!$       & write(debug_unit,*) me,' ',trim(name),&
+!!$       & 'Done NUMBMM 1'
+!!$
+!!$  call am3%mv_to(acsr3)
+!!$  ! Compute omega_int
+!!$  ommx = szero
+!!$  do i=1, ncol
+!!$    if (ilaggr(i) >0) then 
+!!$      omi(i) = omp(ilaggr(i))
+!!$    else
+!!$      omi(i) = szero
+!!$    end if
+!!$    if(abs(omi(i)) .gt. abs(ommx)) ommx = omi(i)
+!!$  end do
+!!$  ! Compute omega_fine
+!!$  do i=1, nrow
+!!$    omf(i) = ommx
+!!$    do j=acsr3%irp(i),acsr3%irp(i+1)-1
+!!$      if(abs(omi(acsr3%ja(j))) .lt. abs(omf(i))) omf(i)=omi(acsr3%ja(j))
+!!$    end do
+!!$ ! !     if(min(real(omf(i)),aimag(omf(i))) < szero) omf(i) = szero
+!!$    if(psb_minreal(omf(i)) < szero) omf(i) = szero
+!!$  end do
+!!$
+!!$  omf(1:nrow) = omf(1:nrow) * adinv(1:nrow)
+!!$
+!!$  if (filter_mat) then
+!!$    !
+!!$    ! Build the filtered matrix Af from A
+!!$    ! 
+!!$    call la%cscnv(acsrf,info,dupl=psb_dupl_add_)
+!!$
+!!$    do i=1,nrow
+!!$      tmp = szero
+!!$      jd  = -1 
+!!$      do j=acsrf%irp(i),acsrf%irp(i+1)-1
+!!$        if (acsrf%ja(j) == i) jd = j 
+!!$        if (abs(acsrf%val(j)) < theta*sqrt(abs(adiag(i)*adiag(acsrf%ja(j))))) then
+!!$          tmp=tmp+acsrf%val(j)
+!!$          acsrf%val(j)=szero
+!!$        endif
+!!$      enddo
+!!$      if (jd == -1) then 
+!!$        write(0,*) 'Wrong input: we need the diagonal!!!!', i
+!!$      else
+!!$        acsrf%val(jd)=acsrf%val(jd)-tmp
+!!$      end if
+!!$    enddo
+!!$    ! Take out zeroed terms 
+!!$    call acsrf%clean_zeros(info)
+!!$
+!!$    !
+!!$    ! Build the smoothed prolongator using the filtered matrix
+!!$    ! 
+!!$    do i=1,acsrf%get_nrows()
+!!$      do j=acsrf%irp(i),acsrf%irp(i+1)-1
+!!$        if (acsrf%ja(j) == i) then 
+!!$          acsrf%val(j) = sone - omf(i)*acsrf%val(j) 
+!!$        else
+!!$          acsrf%val(j) = - omf(i)*acsrf%val(j) 
+!!$        end if
+!!$      end do
+!!$    end do
+!!$
+!!$    if (debug_level >= psb_debug_outer_) &
+!!$         & write(debug_unit,*) me,' ',trim(name),&
+!!$         & 'Done gather, going for SYMBMM 1'
+!!$
+!!$    call af%mv_from(acsrf)
+!!$    !
+!!$    ! op_prol = (I-w*D*Af)Ptilde
+!!$    ! Doing it this way means to consider diag(Af_i)
+!!$    ! 
+!!$    !
+!!$    call psb_spspmm(af,ptilde,op_prol,info)
+!!$    if (debug_level >= psb_debug_outer_) &
+!!$         & write(debug_unit,*) me,' ',trim(name),&
+!!$         & 'Done SPSPMM 1'
+!!$  else
+!!$    !
+!!$    ! Build the smoothed prolongator using the original matrix
+!!$    !
+!!$    do i=1,acsr3%get_nrows()
+!!$      do j=acsr3%irp(i),acsr3%irp(i+1)-1
+!!$        if (acsr3%ja(j) == i) then 
+!!$          acsr3%val(j) = sone - omf(i)*acsr3%val(j) 
+!!$        else
+!!$          acsr3%val(j) = - omf(i)*acsr3%val(j) 
+!!$        end if
+!!$      end do
+!!$    end do
+!!$
+!!$    call am3%mv_from(acsr3)
+!!$    if (debug_level >= psb_debug_outer_) &
+!!$         & write(debug_unit,*) me,' ',trim(name),&
+!!$         & 'Done gather, going for SYMBMM 1'
+!!$    !
+!!$    ! 
+!!$    ! op_prol = (I-w*D*A)Ptilde
+!!$    ! 
+!!$    !
+!!$    call psb_spspmm(am3,ptilde,op_prol,info)
+!!$    if (debug_level >= psb_debug_outer_) &
+!!$         & write(debug_unit,*) me,' ',trim(name),&
+!!$         & 'Done NUMBMM 1'
+!!$
+!!$  end if
+!!$
+!!$
+!!$  !
+!!$  ! Ok, let's start over with the restrictor
+!!$  ! 
+!!$  call ptilde%transc(rtilde)
+!!$  call la%cscnv(atmp,info,type='csr')
+!!$  call psb_sphalo(atmp,desc_a,am4,info,&
+!!$       & colcnv=.true.,rowscale=.true.)
+!!$  nrt  = am4%get_nrows() 
+!!$  call am4%csclip(atmp2,info,lone,nrt,lone,ncol)
+!!$  call atmp2%cscnv(info,type='CSR')
+!!$  if (info == psb_success_) call psb_rwextd(ncol,atmp,info,b=atmp2)      
+!!$  call am4%free()
+!!$  call atmp2%free()
+!!$
+!!$  ! This is to compute the transpose. It ONLY works if the
+!!$  ! original A has a symmetric pattern.
+!!$  call atmp%transc(atmp2) 
+!!$  call atmp2%csclip(dat,info,lone,nrow,lone,ncol)
+!!$  call dat%cscnv(info,type='csr')
+!!$  call dat%scal(adinv,info)
+!!$
+!!$  ! Now for the product. 
+!!$  call psb_spspmm(dat,ptilde,datp,info)
+!!$
+!!$  call datp%clone(atmp2,info)
+!!$  call psb_sphalo(atmp2,desc_a,am4,info,&
+!!$       & colcnv=.false.,rowscale=.true.,outfmt='CSR  ')
+!!$  if (info == psb_success_) call psb_rwextd(ncol,atmp2,info,b=am4)      
+!!$  if (info == psb_success_) call am4%free()
+!!$
+!!$
+!!$  call psb_symbmm(dat,atmp2,datdatp,info)
+!!$  call psb_numbmm(dat,atmp2,datdatp)
+!!$  call atmp2%free()
+!!$
+!!$  call datp%mv_to(csc_datp)    
+!!$  call datdatp%mv_to(csc_datdatp)    
+!!$
+!!$  call csc_mat_col_prod(csc_datp,csc_datdatp,omp,info)
+!!$  call csc_mat_col_prod(csc_datdatp,csc_datdatp,oden,info)
+!!$  call psb_sum(ctxt,omp)
+!!$  call psb_sum(ctxt,oden)
+!!$
+!!$
+!!$  ! !$  write(debug_unit,*) trim(name),' OMP_R :',omp
+!!$  ! ! $  write(debug_unit,*) trim(name),' ODEN_R:',oden
+!!$  omp = omp/oden
+!!$  ! !$  write(0,*) 'Check on output restrictor',omp(1:min(size(omp),10))
+!!$  ! Compute omega_int
+!!$  ommx = szero
+!!$  do i=1, ncol
+!!$    if (ilaggr(i) >0) then 
+!!$      omi(i) = omp(ilaggr(i))
+!!$    else
+!!$      omi(i) = szero
+!!$    end if
+!!$    if(abs(omi(i)) .gt. abs(ommx)) ommx = omi(i)
+!!$  end do
+!!$  ! Compute omega_fine
+!!$  ! Going over the columns of atmp means going over the rows
+!!$  ! of A^T. Hopefully ;-) 
+!!$  call atmp%cp_to(acsc)
+!!$
+!!$  do i=1, nrow
+!!$    omf(i) = ommx
+!!$    do j= acsc%icp(i),acsc%icp(i+1)-1
+!!$      if(abs(omi(acsc%ia(j))) .lt. abs(omf(i))) omf(i)=omi(acsc%ia(j))
+!!$    end do
+!!$ ! !    if(min(real(omf(i)),aimag(omf(i))) < szero) omf(i) = szero
+!!$    if(psb_minreal(omf(i)) < szero) omf(i) = szero
+!!$  end do
+!!$  omf(1:nrow) = omf(1:nrow)*adinv(1:nrow)
+!!$  call psb_halo(omf,desc_a,info)
+!!$  call acsc%free() 
+!!$
+!!$
+!!$  call atmp%mv_to(acsr1)
+!!$
+!!$  do i=1,acsr1%get_nrows()
+!!$    do j=acsr1%irp(i),acsr1%irp(i+1)-1
+!!$      if (acsr1%ja(j) == i) then 
+!!$        acsr1%val(j) = sone - acsr1%val(j)*omf(acsr1%ja(j))
+!!$      else
+!!$        acsr1%val(j) =      - acsr1%val(j)*omf(acsr1%ja(j))
+!!$      end if
+!!$    end do
+!!$  end do
+!!$  call atmp%mv_from(acsr1)
+!!$
+!!$  call rtilde%mv_to(tmpcoo)
+!!$  nzl = tmpcoo%get_nzeros()
+!!$  i=0
+!!$  do k=1, nzl
+!!$    if ((naggrm1 < tmpcoo%ia(k)) .and. (tmpcoo%ia(k) <= naggrp1)) then
+!!$      i = i+1
+!!$      tmpcoo%val(i) = tmpcoo%val(k)
+!!$      tmpcoo%ia(i)  = tmpcoo%ia(k)
+!!$      tmpcoo%ja(i)  = tmpcoo%ja(k)
+!!$    end if
+!!$  end do
+!!$  call tmpcoo%set_nzeros(i)
+!!$  call rtilde%mv_from(tmpcoo)
+!!$  call rtilde%cscnv(info,type='csr')
+!!$
+!!$  call psb_spspmm(rtilde,atmp,op_restr,info)
+!!$
+!!$  !
+!!$  ! Now we have to gather the halo of op_prol, and add it to itself
+!!$  ! to multiply it by A,
+!!$  !
+!!$  call op_prol%clone(tmp_prol,info) 
+!!$  if (info == psb_success_) call psb_sphalo(tmp_prol,desc_a,am4,info,&
+!!$       & colcnv=.false.,rowscale=.true.)
+!!$  if (info == psb_success_) call psb_rwextd(ncol,tmp_prol,info,b=am4)      
+!!$  if (info == psb_success_) call am4%free()
+!!$
+!!$  if(info /= psb_success_) then
+!!$    call psb_errpush(psb_err_internal_error_,name,a_err='Halo of op_prol')
+!!$    goto 9999
+!!$  end if
+!!$
+!!$  !
+!!$  ! Now we have to fix this.  The only rows of B that are correct 
+!!$  ! are those corresponding to "local" aggregates, i.e. indices in ilaggr(:)
+!!$  !
+!!$  call op_restr%mv_to(tmpcoo)
+!!$
+!!$  nzl = tmpcoo%get_nzeros()
+!!$  i=0
+!!$  do k=1, nzl
+!!$    if ((naggrm1 < tmpcoo%ia(k)) .and. (tmpcoo%ia(k) <= naggrp1)) then
+!!$      i = i+1
+!!$      tmpcoo%val(i) = tmpcoo%val(k)
+!!$      tmpcoo%ia(i)  = tmpcoo%ia(k)
+!!$      tmpcoo%ja(i)  = tmpcoo%ja(k)
+!!$    end if
+!!$  end do
+!!$  call tmpcoo%set_nzeros(i)
+!!$  call op_restr%mv_from(tmpcoo)
+!!$  call op_restr%cscnv(info,type='csr')
+!!$
+!!$
+!!$  if (debug_level >= psb_debug_outer_) &
+!!$       & write(debug_unit,*) me,' ',trim(name),&
+!!$       & 'starting sphalo/ rwxtd'
+!!$
+!!$  call psb_spspmm(la,tmp_prol,am3,info)
+!!$  if (debug_level >= psb_debug_outer_) &
+!!$       & write(debug_unit,*) me,' ',trim(name),&
+!!$       & 'Done SPSPMM 2'
+!!$
+!!$  call psb_sphalo(am3,desc_a,am4,info,&
+!!$       & colcnv=.false.,rowscale=.true.)
+!!$  if (info == psb_success_) call psb_rwextd(ncol,am3,info,b=am4)      
+!!$  if (info == psb_success_) call am4%free()
+!!$
+!!$  if(info /= psb_success_) then
+!!$    call psb_errpush(psb_err_internal_error_,name,&
+!!$         & a_err='Extend am3')
+!!$    goto 9999
+!!$  end if
+!!$  if (debug_level >= psb_debug_outer_) &
+!!$       & write(debug_unit,*) me,' ',trim(name),&
+!!$       & 'Done sphalo/ rwxtd'
+!!$
+!!$  call psb_spspmm(op_restr,am3,ac,info)
+!!$  if (info == psb_success_) call am3%free()
+!!$  if (info == psb_success_) call ac%cscnv(info,type='coo',dupl=psb_dupl_add_)
+!!$
+!!$  if (info /= psb_success_) then
+!!$    call psb_errpush(psb_err_internal_error_,name,&
+!!$         &a_err='Build ac = op_restr x am3')
+!!$    goto 9999
+!!$  end if



@@ -116,7 +116,7 @@ subroutine amg_saggrmat_smth_bld(a,desc_a,ilaggr,nlaggr,parms,&
  type(psb_desc_type), intent(inout)       :: desc_a
  integer(psb_lpk_), intent(inout)         :: ilaggr(:), nlaggr(:)
  type(amg_sml_parms), intent(inout)      :: parms 
-  type(psb_sspmat_type), intent(out)      :: op_prol,ac,op_restr
+  type(psb_sspmat_type), intent(inout)      :: op_prol,ac,op_restr
  type(psb_lsspmat_type), intent(inout)    :: t_prol
  type(psb_desc_type), intent(inout)       :: desc_ac
  integer(psb_ipk_), intent(out)           :: info
@@ -83,8 +83,8 @@ subroutine  amg_z_dec_aggregator_build_tprol(ag,parms,ag_data,&
  class(amg_z_dec_aggregator_type), target, intent(inout) :: ag
  type(amg_dml_parms), intent(inout)  :: parms
  type(amg_daggr_data), intent(in)    :: ag_data
-  type(psb_zspmat_type), intent(in)   :: a
-  type(psb_desc_type), intent(in)     :: desc_a
+  type(psb_zspmat_type), intent(inout) :: a
+  type(psb_desc_type), intent(inout)    :: desc_a
  integer(psb_lpk_), allocatable, intent(out) :: ilaggr(:),nlaggr(:)
  type(psb_lzspmat_type), intent(out)  :: t_prol
  integer(psb_ipk_), intent(out)      :: info
@@ -86,8 +86,8 @@ subroutine  amg_z_symdec_aggregator_build_tprol(ag,parms,ag_data,&
  class(amg_z_symdec_aggregator_type), target, intent(inout) :: ag
  type(amg_dml_parms), intent(inout)  :: parms
  type(amg_daggr_data), intent(in)    :: ag_data
-  type(psb_zspmat_type), intent(in)   :: a
-  type(psb_desc_type), intent(in)     :: desc_a
+  type(psb_zspmat_type), intent(inout) :: a
+  type(psb_desc_type), intent(inout)   :: desc_a
  integer(psb_lpk_), allocatable, intent(out) :: ilaggr(:),nlaggr(:)
  type(psb_lzspmat_type), intent(out)  :: op_prol
  integer(psb_ipk_), intent(out)      :: info
@@ -105,7 +105,7 @@
 !
 !
 subroutine amg_zaggrmat_minnrg_bld(a,desc_a,ilaggr,nlaggr,parms,&
-     & ac,desc_ac,op_prol,op_restr,info)
+     & ac,desc_ac,op_prol,op_restr,t_prol,info)
  use psb_base_mod
  use amg_base_prec_type
  use amg_z_inner_mod, amg_protect_name => amg_zaggrmat_minnrg_bld
@@ -117,8 +117,8 @@ subroutine amg_zaggrmat_minnrg_bld(a,desc_a,ilaggr,nlaggr,parms,&
  type(psb_desc_type), intent(inout)            :: desc_a
  integer(psb_lpk_), intent(inout)              :: ilaggr(:), nlaggr(:)
  type(amg_dml_parms), intent(inout)         :: parms 
-  type(psb_lzspmat_type), intent(inout)       :: op_prol
-  type(psb_lzspmat_type), intent(out)         :: ac,op_restr
+  type(psb_lzspmat_type), intent(inout)       :: t_prol
+  type(psb_zspmat_type), intent(inout)       :: op_prol, ac,op_restr
  type(psb_desc_type), intent(inout)            :: desc_ac
  integer(psb_ipk_), intent(out)                :: info

@@ -171,6 +171,8 @@ subroutine amg_zaggrmat_minnrg_bld(a,desc_a,ilaggr,nlaggr,parms,&

  filter_mat = (parms%aggr_filter == amg_filter_mat_)

+  !NEEDS TO BE REWORKED !!
+  
  ! naggr: number of local aggregates
  ! nrow: local rows. 
  ! 
@@ -183,361 +185,361 @@ subroutine amg_zaggrmat_minnrg_bld(a,desc_a,ilaggr,nlaggr,parms,&
    goto 9999      
  end if

-  ! Get the diagonal D
-  adiag =  a%get_diag(info)
-  if (info == psb_success_) &
-       & call psb_realloc(ncol,adiag,info)    
-  if (info == psb_success_) &
-       & call psb_halo(adiag,desc_a,info)
-  if (info == psb_success_) call a%cp_to_l(la)
-  if (info /= psb_success_) then
-    call psb_errpush(psb_err_from_subroutine_,name,a_err='sp_getdiag')
-    goto 9999
-  end if
-
-  do i=1,size(adiag)
-    if (adiag(i) /= zzero) then
-      adinv(i) = zone / adiag(i)
-    else
-      adinv(i) = zone
-    end if
-  end do
-
-
-
-  ! 1. Allocate Ptilde in sparse matrix form 
-  call op_prol%mv_to(tmpcoo)
-  call ptilde%mv_from(tmpcoo)
-  call ptilde%cscnv(info,type='csr')
-
-  if (info == psb_success_) call la%cscnv(am3,info,type='csr',dupl=psb_dupl_add_)
-  if (info == psb_success_) call la%cscnv(da,info,type='csr',dupl=psb_dupl_add_)
-  if (info /= psb_success_) then
-    call psb_errpush(psb_err_from_subroutine_,name,a_err='spcnv')
-    goto 9999
-  end if
-  if (debug_level >= psb_debug_outer_) &
-       & write(debug_unit,*) me,' ',trim(name),&
-       & ' Initial copies done.'
-
-  call da%scal(adinv,info)
-
-  call psb_spspmm(da,ptilde,dap,info)
-
-  if(info /= psb_success_) then
-    call psb_errpush(psb_err_from_subroutine_,name,a_err='spspmm 1')
-    goto 9999
-  end if
-
-  call dap%clone(atmp,info)
-
-  call psb_sphalo(atmp,desc_a,am4,info,&
-       & colcnv=.false.,rowscale=.true.,outfmt='CSR  ')
-  if (info == psb_success_) call psb_rwextd(ncol,atmp,info,b=am4)      
-  if (info == psb_success_) call am4%free()
-
-  call psb_spspmm(da,atmp,dadap,info)
-  call atmp%free()
-
-  !  !$  write(0,*) 'Columns of AP',psb_sp_get_ncols(ap)
-  !  !$  write(0,*) 'Columns of ADAP',psb_sp_get_ncols(adap)
-  call dap%mv_to(csc_dap)
-  call dadap%mv_to(csc_dadap)
-
-  call csc_mat_col_prod(csc_dap,csc_dadap,omp,info)
-  call csc_mat_col_prod(csc_dadap,csc_dadap,oden,info)
-  call psb_sum(ctxt,omp)
-  call psb_sum(ctxt,oden)
-  ! !$  write(0,*) trim(name),' OMP :',omp
-  ! !$  write(0,*) trim(name),' ODEN:',oden
-
-  omp = omp/oden
-
-  ! !$  write(0,*) 'Check on output prolongator ',omp(1:min(size(omp),10))
-  if (debug_level >= psb_debug_outer_) &
-       & write(debug_unit,*) me,' ',trim(name),&
-       & 'Done NUMBMM 1'
-
-  call am3%mv_to(acsr3)
-  ! Compute omega_int
-  ommx = zzero
-  do i=1, ncol
-    if (ilaggr(i) >0) then 
-      omi(i) = omp(ilaggr(i))
-    else
-      omi(i) = zzero
-    end if
-    if(abs(omi(i)) .gt. abs(ommx)) ommx = omi(i)
-  end do
-  ! Compute omega_fine
-  do i=1, nrow
-    omf(i) = ommx
-    do j=acsr3%irp(i),acsr3%irp(i+1)-1
-      if(abs(omi(acsr3%ja(j))) .lt. abs(omf(i))) omf(i)=omi(acsr3%ja(j))
-    end do
-!!$    if(min(real(omf(i)),aimag(omf(i))) < dzero) omf(i) = zzero
-    if(psb_minreal(omf(i)) < dzero) omf(i) = zzero
-  end do
-
-  omf(1:nrow) = omf(1:nrow) * adinv(1:nrow)
-
-  if (filter_mat) then
-    !
-    ! Build the filtered matrix Af from A
-    ! 
-    call la%cscnv(acsrf,info,dupl=psb_dupl_add_)
-
-    do i=1,nrow
-      tmp = zzero
-      jd  = -1 
-      do j=acsrf%irp(i),acsrf%irp(i+1)-1
-        if (acsrf%ja(j) == i) jd = j 
-        if (abs(acsrf%val(j)) < theta*sqrt(abs(adiag(i)*adiag(acsrf%ja(j))))) then
-          tmp=tmp+acsrf%val(j)
-          acsrf%val(j)=zzero
-        endif
-      enddo
-      if (jd == -1) then 
-        write(0,*) 'Wrong input: we need the diagonal!!!!', i
-      else
-        acsrf%val(jd)=acsrf%val(jd)-tmp
-      end if
-    enddo
-    ! Take out zeroed terms 
-    call acsrf%clean_zeros(info)
-
-    !
-    ! Build the smoothed prolongator using the filtered matrix
-    ! 
-    do i=1,acsrf%get_nrows()
-      do j=acsrf%irp(i),acsrf%irp(i+1)-1
-        if (acsrf%ja(j) == i) then 
-          acsrf%val(j) = zone - omf(i)*acsrf%val(j) 
-        else
-          acsrf%val(j) = - omf(i)*acsrf%val(j) 
-        end if
-      end do
-    end do
-
-    if (debug_level >= psb_debug_outer_) &
-         & write(debug_unit,*) me,' ',trim(name),&
-         & 'Done gather, going for SYMBMM 1'
-
-    call af%mv_from(acsrf)
-    !
-    ! op_prol = (I-w*D*Af)Ptilde
-    ! Doing it this way means to consider diag(Af_i)
-    ! 
-    !
-    call psb_spspmm(af,ptilde,op_prol,info)
-    if (debug_level >= psb_debug_outer_) &
-         & write(debug_unit,*) me,' ',trim(name),&
-         & 'Done SPSPMM 1'
-  else
-    !
-    ! Build the smoothed prolongator using the original matrix
-    !
-    do i=1,acsr3%get_nrows()
-      do j=acsr3%irp(i),acsr3%irp(i+1)-1
-        if (acsr3%ja(j) == i) then 
-          acsr3%val(j) = zone - omf(i)*acsr3%val(j) 
-        else
-          acsr3%val(j) = - omf(i)*acsr3%val(j) 
-        end if
-      end do
-    end do
-
-    call am3%mv_from(acsr3)
-    if (debug_level >= psb_debug_outer_) &
-         & write(debug_unit,*) me,' ',trim(name),&
-         & 'Done gather, going for SYMBMM 1'
-    !
-    ! 
-    ! op_prol = (I-w*D*A)Ptilde
-    ! 
-    !
-    call psb_spspmm(am3,ptilde,op_prol,info)
-    if (debug_level >= psb_debug_outer_) &
-         & write(debug_unit,*) me,' ',trim(name),&
-         & 'Done NUMBMM 1'
-
-  end if
-
-
-  !
-  ! Ok, let's start over with the restrictor
-  ! 
-  call ptilde%transc(rtilde)
-  call la%cscnv(atmp,info,type='csr')
-  call psb_sphalo(atmp,desc_a,am4,info,&
-       & colcnv=.true.,rowscale=.true.)
-  nrt  = am4%get_nrows() 
-  call am4%csclip(atmp2,info,lone,nrt,lone,ncol)
-  call atmp2%cscnv(info,type='CSR')
-  if (info == psb_success_) call psb_rwextd(ncol,atmp,info,b=atmp2)      
-  call am4%free()
-  call atmp2%free()
-
-  ! This is to compute the transpose. It ONLY works if the
-  ! original A has a symmetric pattern.
-  call atmp%transc(atmp2) 
-  call atmp2%csclip(dat,info,lone,nrow,lone,ncol)
-  call dat%cscnv(info,type='csr')
-  call dat%scal(adinv,info)
-
-  ! Now for the product. 
-  call psb_spspmm(dat,ptilde,datp,info)
-
-  call datp%clone(atmp2,info)
-  call psb_sphalo(atmp2,desc_a,am4,info,&
-       & colcnv=.false.,rowscale=.true.,outfmt='CSR  ')
-  if (info == psb_success_) call psb_rwextd(ncol,atmp2,info,b=am4)      
-  if (info == psb_success_) call am4%free()
-
-
-  call psb_symbmm(dat,atmp2,datdatp,info)
-  call psb_numbmm(dat,atmp2,datdatp)
-  call atmp2%free()
-
-  call datp%mv_to(csc_datp)    
-  call datdatp%mv_to(csc_datdatp)    
-
-  call csc_mat_col_prod(csc_datp,csc_datdatp,omp,info)
-  call csc_mat_col_prod(csc_datdatp,csc_datdatp,oden,info)
-  call psb_sum(ctxt,omp)
-  call psb_sum(ctxt,oden)
-
-
-  ! !$  write(debug_unit,*) trim(name),' OMP_R :',omp
-  ! ! $  write(debug_unit,*) trim(name),' ODEN_R:',oden
-  omp = omp/oden
-  ! !$  write(0,*) 'Check on output restrictor',omp(1:min(size(omp),10))
-  ! Compute omega_int
-  ommx = zzero
-  do i=1, ncol
-    if (ilaggr(i) >0) then 
-      omi(i) = omp(ilaggr(i))
-    else
-      omi(i) = zzero
-    end if
-    if(abs(omi(i)) .gt. abs(ommx)) ommx = omi(i)
-  end do
-  ! Compute omega_fine
-  ! Going over the columns of atmp means going over the rows
-  ! of A^T. Hopefully ;-) 
-  call atmp%cp_to(acsc)
-
-  do i=1, nrow
-    omf(i) = ommx
-    do j= acsc%icp(i),acsc%icp(i+1)-1
-      if(abs(omi(acsc%ia(j))) .lt. abs(omf(i))) omf(i)=omi(acsc%ia(j))
-    end do
-!!$    if(min(real(omf(i)),aimag(omf(i))) < dzero) omf(i) = zzero
-    if(psb_minreal(omf(i)) < dzero) omf(i) = zzero
-  end do
-  omf(1:nrow) = omf(1:nrow)*adinv(1:nrow)
-  call psb_halo(omf,desc_a,info)
-  call acsc%free() 
-
-
-  call atmp%mv_to(acsr1)
-
-  do i=1,acsr1%get_nrows()
-    do j=acsr1%irp(i),acsr1%irp(i+1)-1
-      if (acsr1%ja(j) == i) then 
-        acsr1%val(j) = zone - acsr1%val(j)*omf(acsr1%ja(j))
-      else
-        acsr1%val(j) =      - acsr1%val(j)*omf(acsr1%ja(j))
-      end if
-    end do
-  end do
-  call atmp%mv_from(acsr1)
-
-  call rtilde%mv_to(tmpcoo)
-  nzl = tmpcoo%get_nzeros()
-  i=0
-  do k=1, nzl
-    if ((naggrm1 < tmpcoo%ia(k)) .and. (tmpcoo%ia(k) <= naggrp1)) then
-      i = i+1
-      tmpcoo%val(i) = tmpcoo%val(k)
-      tmpcoo%ia(i)  = tmpcoo%ia(k)
-      tmpcoo%ja(i)  = tmpcoo%ja(k)
-    end if
-  end do
-  call tmpcoo%set_nzeros(i)
-  call rtilde%mv_from(tmpcoo)
-  call rtilde%cscnv(info,type='csr')
-
-  call psb_spspmm(rtilde,atmp,op_restr,info)
-
-  !
-  ! Now we have to gather the halo of op_prol, and add it to itself
-  ! to multiply it by A,
-  !
-  call op_prol%clone(tmp_prol,info) 
-  if (info == psb_success_) call psb_sphalo(tmp_prol,desc_a,am4,info,&
-       & colcnv=.false.,rowscale=.true.)
-  if (info == psb_success_) call psb_rwextd(ncol,tmp_prol,info,b=am4)      
-  if (info == psb_success_) call am4%free()
-
-  if(info /= psb_success_) then
-    call psb_errpush(psb_err_internal_error_,name,a_err='Halo of op_prol')
-    goto 9999
-  end if
-
-  !
-  ! Now we have to fix this.  The only rows of B that are correct 
-  ! are those corresponding to "local" aggregates, i.e. indices in ilaggr(:)
-  !
-  call op_restr%mv_to(tmpcoo)
-
-  nzl = tmpcoo%get_nzeros()
-  i=0
-  do k=1, nzl
-    if ((naggrm1 < tmpcoo%ia(k)) .and. (tmpcoo%ia(k) <= naggrp1)) then
-      i = i+1
-      tmpcoo%val(i) = tmpcoo%val(k)
-      tmpcoo%ia(i)  = tmpcoo%ia(k)
-      tmpcoo%ja(i)  = tmpcoo%ja(k)
-    end if
-  end do
-  call tmpcoo%set_nzeros(i)
-  call op_restr%mv_from(tmpcoo)
-  call op_restr%cscnv(info,type='csr')
-
-
-  if (debug_level >= psb_debug_outer_) &
-       & write(debug_unit,*) me,' ',trim(name),&
-       & 'starting sphalo/ rwxtd'
-
-  call psb_spspmm(la,tmp_prol,am3,info)
-  if (debug_level >= psb_debug_outer_) &
-       & write(debug_unit,*) me,' ',trim(name),&
-       & 'Done SPSPMM 2'
-
-  call psb_sphalo(am3,desc_a,am4,info,&
-       & colcnv=.false.,rowscale=.true.)
-  if (info == psb_success_) call psb_rwextd(ncol,am3,info,b=am4)      
-  if (info == psb_success_) call am4%free()
-
-  if(info /= psb_success_) then
-    call psb_errpush(psb_err_internal_error_,name,&
-         & a_err='Extend am3')
-    goto 9999
-  end if
-  if (debug_level >= psb_debug_outer_) &
-       & write(debug_unit,*) me,' ',trim(name),&
-       & 'Done sphalo/ rwxtd'
-
-  call psb_spspmm(op_restr,am3,ac,info)
-  if (info == psb_success_) call am3%free()
-  if (info == psb_success_) call ac%cscnv(info,type='coo',dupl=psb_dupl_add_)
-
-  if (info /= psb_success_) then
-    call psb_errpush(psb_err_internal_error_,name,&
-         &a_err='Build ac = op_restr x am3')
-    goto 9999
-  end if
+!!$  ! Get the diagonal D
+!!$  adiag =  a%get_diag(info)
+!!$  if (info == psb_success_) &
+!!$       & call psb_realloc(ncol,adiag,info)    
+!!$  if (info == psb_success_) &
+!!$       & call psb_halo(adiag,desc_a,info)
+!!$  if (info == psb_success_) call a%cp_to_l(la)
+!!$  if (info /= psb_success_) then
+!!$    call psb_errpush(psb_err_from_subroutine_,name,a_err='sp_getdiag')
+!!$    goto 9999
+!!$  end if
+!!$
+!!$  do i=1,size(adiag)
+!!$    if (adiag(i) /= zzero) then
+!!$      adinv(i) = zone / adiag(i)
+!!$    else
+!!$      adinv(i) = zone
+!!$    end if
+!!$  end do
+!!$
+!!$
+!!$
+!!$  ! 1. Allocate Ptilde in sparse matrix form 
+!!$  call op_prol%mv_to(tmpcoo)
+!!$  call ptilde%mv_from(tmpcoo)
+!!$  call ptilde%cscnv(info,type='csr')
+!!$
+!!$  if (info == psb_success_) call la%cscnv(am3,info,type='csr',dupl=psb_dupl_add_)
+!!$  if (info == psb_success_) call la%cscnv(da,info,type='csr',dupl=psb_dupl_add_)
+!!$  if (info /= psb_success_) then
+!!$    call psb_errpush(psb_err_from_subroutine_,name,a_err='spcnv')
+!!$    goto 9999
+!!$  end if
+!!$  if (debug_level >= psb_debug_outer_) &
+!!$       & write(debug_unit,*) me,' ',trim(name),&
+!!$       & ' Initial copies done.'
+!!$
+!!$  call da%scal(adinv,info)
+!!$
+!!$  call psb_spspmm(da,ptilde,dap,info)
+!!$
+!!$  if(info /= psb_success_) then
+!!$    call psb_errpush(psb_err_from_subroutine_,name,a_err='spspmm 1')
+!!$    goto 9999
+!!$  end if
+!!$
+!!$  call dap%clone(atmp,info)
+!!$
+!!$  call psb_sphalo(atmp,desc_a,am4,info,&
+!!$       & colcnv=.false.,rowscale=.true.,outfmt='CSR  ')
+!!$  if (info == psb_success_) call psb_rwextd(ncol,atmp,info,b=am4)      
+!!$  if (info == psb_success_) call am4%free()
+!!$
+!!$  call psb_spspmm(da,atmp,dadap,info)
+!!$  call atmp%free()
+!!$
+!!$  !  !$  write(0,*) 'Columns of AP',psb_sp_get_ncols(ap)
+!!$  !  !$  write(0,*) 'Columns of ADAP',psb_sp_get_ncols(adap)
+!!$  call dap%mv_to(csc_dap)
+!!$  call dadap%mv_to(csc_dadap)
+!!$
+!!$  call csc_mat_col_prod(csc_dap,csc_dadap,omp,info)
+!!$  call csc_mat_col_prod(csc_dadap,csc_dadap,oden,info)
+!!$  call psb_sum(ctxt,omp)
+!!$  call psb_sum(ctxt,oden)
+!!$  ! !$  write(0,*) trim(name),' OMP :',omp
+!!$  ! !$  write(0,*) trim(name),' ODEN:',oden
+!!$
+!!$  omp = omp/oden
+!!$
+!!$  ! !$  write(0,*) 'Check on output prolongator ',omp(1:min(size(omp),10))
+!!$  if (debug_level >= psb_debug_outer_) &
+!!$       & write(debug_unit,*) me,' ',trim(name),&
+!!$       & 'Done NUMBMM 1'
+!!$
+!!$  call am3%mv_to(acsr3)
+!!$  ! Compute omega_int
+!!$  ommx = zzero
+!!$  do i=1, ncol
+!!$    if (ilaggr(i) >0) then 
+!!$      omi(i) = omp(ilaggr(i))
+!!$    else
+!!$      omi(i) = zzero
+!!$    end if
+!!$    if(abs(omi(i)) .gt. abs(ommx)) ommx = omi(i)
+!!$  end do
+!!$  ! Compute omega_fine
+!!$  do i=1, nrow
+!!$    omf(i) = ommx
+!!$    do j=acsr3%irp(i),acsr3%irp(i+1)-1
+!!$      if(abs(omi(acsr3%ja(j))) .lt. abs(omf(i))) omf(i)=omi(acsr3%ja(j))
+!!$    end do
+!!$ ! !     if(min(real(omf(i)),aimag(omf(i))) < dzero) omf(i) = zzero
+!!$    if(psb_minreal(omf(i)) < dzero) omf(i) = zzero
+!!$  end do
+!!$
+!!$  omf(1:nrow) = omf(1:nrow) * adinv(1:nrow)
+!!$
+!!$  if (filter_mat) then
+!!$    !
+!!$    ! Build the filtered matrix Af from A
+!!$    ! 
+!!$    call la%cscnv(acsrf,info,dupl=psb_dupl_add_)
+!!$
+!!$    do i=1,nrow
+!!$      tmp = zzero
+!!$      jd  = -1 
+!!$      do j=acsrf%irp(i),acsrf%irp(i+1)-1
+!!$        if (acsrf%ja(j) == i) jd = j 
+!!$        if (abs(acsrf%val(j)) < theta*sqrt(abs(adiag(i)*adiag(acsrf%ja(j))))) then
+!!$          tmp=tmp+acsrf%val(j)
+!!$          acsrf%val(j)=zzero
+!!$        endif
+!!$      enddo
+!!$      if (jd == -1) then 
+!!$        write(0,*) 'Wrong input: we need the diagonal!!!!', i
+!!$      else
+!!$        acsrf%val(jd)=acsrf%val(jd)-tmp
+!!$      end if
+!!$    enddo
+!!$    ! Take out zeroed terms 
+!!$    call acsrf%clean_zeros(info)
+!!$
+!!$    !
+!!$    ! Build the smoothed prolongator using the filtered matrix
+!!$    ! 
+!!$    do i=1,acsrf%get_nrows()
+!!$      do j=acsrf%irp(i),acsrf%irp(i+1)-1
+!!$        if (acsrf%ja(j) == i) then 
+!!$          acsrf%val(j) = zone - omf(i)*acsrf%val(j) 
+!!$        else
+!!$          acsrf%val(j) = - omf(i)*acsrf%val(j) 
+!!$        end if
+!!$      end do
+!!$    end do
+!!$
+!!$    if (debug_level >= psb_debug_outer_) &
+!!$         & write(debug_unit,*) me,' ',trim(name),&
+!!$         & 'Done gather, going for SYMBMM 1'
+!!$
+!!$    call af%mv_from(acsrf)
+!!$    !
+!!$    ! op_prol = (I-w*D*Af)Ptilde
+!!$    ! Doing it this way means to consider diag(Af_i)
+!!$    ! 
+!!$    !
+!!$    call psb_spspmm(af,ptilde,op_prol,info)
+!!$    if (debug_level >= psb_debug_outer_) &
+!!$         & write(debug_unit,*) me,' ',trim(name),&
+!!$         & 'Done SPSPMM 1'
+!!$  else
+!!$    !
+!!$    ! Build the smoothed prolongator using the original matrix
+!!$    !
+!!$    do i=1,acsr3%get_nrows()
+!!$      do j=acsr3%irp(i),acsr3%irp(i+1)-1
+!!$        if (acsr3%ja(j) == i) then 
+!!$          acsr3%val(j) = zone - omf(i)*acsr3%val(j) 
+!!$        else
+!!$          acsr3%val(j) = - omf(i)*acsr3%val(j) 
+!!$        end if
+!!$      end do
+!!$    end do
+!!$
+!!$    call am3%mv_from(acsr3)
+!!$    if (debug_level >= psb_debug_outer_) &
+!!$         & write(debug_unit,*) me,' ',trim(name),&
+!!$         & 'Done gather, going for SYMBMM 1'
+!!$    !
+!!$    ! 
+!!$    ! op_prol = (I-w*D*A)Ptilde
+!!$    ! 
+!!$    !
+!!$    call psb_spspmm(am3,ptilde,op_prol,info)
+!!$    if (debug_level >= psb_debug_outer_) &
+!!$         & write(debug_unit,*) me,' ',trim(name),&
+!!$         & 'Done NUMBMM 1'
+!!$
+!!$  end if
+!!$
+!!$
+!!$  !
+!!$  ! Ok, let's start over with the restrictor
+!!$  ! 
+!!$  call ptilde%transc(rtilde)
+!!$  call la%cscnv(atmp,info,type='csr')
+!!$  call psb_sphalo(atmp,desc_a,am4,info,&
+!!$       & colcnv=.true.,rowscale=.true.)
+!!$  nrt  = am4%get_nrows() 
+!!$  call am4%csclip(atmp2,info,lone,nrt,lone,ncol)
+!!$  call atmp2%cscnv(info,type='CSR')
+!!$  if (info == psb_success_) call psb_rwextd(ncol,atmp,info,b=atmp2)      
+!!$  call am4%free()
+!!$  call atmp2%free()
+!!$
+!!$  ! This is to compute the transpose. It ONLY works if the
+!!$  ! original A has a symmetric pattern.
+!!$  call atmp%transc(atmp2) 
+!!$  call atmp2%csclip(dat,info,lone,nrow,lone,ncol)
+!!$  call dat%cscnv(info,type='csr')
+!!$  call dat%scal(adinv,info)
+!!$
+!!$  ! Now for the product. 
+!!$  call psb_spspmm(dat,ptilde,datp,info)
+!!$
+!!$  call datp%clone(atmp2,info)
+!!$  call psb_sphalo(atmp2,desc_a,am4,info,&
+!!$       & colcnv=.false.,rowscale=.true.,outfmt='CSR  ')
+!!$  if (info == psb_success_) call psb_rwextd(ncol,atmp2,info,b=am4)      
+!!$  if (info == psb_success_) call am4%free()
+!!$
+!!$
+!!$  call psb_symbmm(dat,atmp2,datdatp,info)
+!!$  call psb_numbmm(dat,atmp2,datdatp)
+!!$  call atmp2%free()
+!!$
+!!$  call datp%mv_to(csc_datp)    
+!!$  call datdatp%mv_to(csc_datdatp)    
+!!$
+!!$  call csc_mat_col_prod(csc_datp,csc_datdatp,omp,info)
+!!$  call csc_mat_col_prod(csc_datdatp,csc_datdatp,oden,info)
+!!$  call psb_sum(ctxt,omp)
+!!$  call psb_sum(ctxt,oden)
+!!$
+!!$
+!!$  ! !$  write(debug_unit,*) trim(name),' OMP_R :',omp
+!!$  ! ! $  write(debug_unit,*) trim(name),' ODEN_R:',oden
+!!$  omp = omp/oden
+!!$  ! !$  write(0,*) 'Check on output restrictor',omp(1:min(size(omp),10))
+!!$  ! Compute omega_int
+!!$  ommx = zzero
+!!$  do i=1, ncol
+!!$    if (ilaggr(i) >0) then 
+!!$      omi(i) = omp(ilaggr(i))
+!!$    else
+!!$      omi(i) = zzero
+!!$    end if
+!!$    if(abs(omi(i)) .gt. abs(ommx)) ommx = omi(i)
+!!$  end do
+!!$  ! Compute omega_fine
+!!$  ! Going over the columns of atmp means going over the rows
+!!$  ! of A^T. Hopefully ;-) 
+!!$  call atmp%cp_to(acsc)
+!!$
+!!$  do i=1, nrow
+!!$    omf(i) = ommx
+!!$    do j= acsc%icp(i),acsc%icp(i+1)-1
+!!$      if(abs(omi(acsc%ia(j))) .lt. abs(omf(i))) omf(i)=omi(acsc%ia(j))
+!!$    end do
+!!$ ! !    if(min(real(omf(i)),aimag(omf(i))) < dzero) omf(i) = zzero
+!!$    if(psb_minreal(omf(i)) < dzero) omf(i) = zzero
+!!$  end do
+!!$  omf(1:nrow) = omf(1:nrow)*adinv(1:nrow)
+!!$  call psb_halo(omf,desc_a,info)
+!!$  call acsc%free() 
+!!$
+!!$
+!!$  call atmp%mv_to(acsr1)
+!!$
+!!$  do i=1,acsr1%get_nrows()
+!!$    do j=acsr1%irp(i),acsr1%irp(i+1)-1
+!!$      if (acsr1%ja(j) == i) then 
+!!$        acsr1%val(j) = zone - acsr1%val(j)*omf(acsr1%ja(j))
+!!$      else
+!!$        acsr1%val(j) =      - acsr1%val(j)*omf(acsr1%ja(j))
+!!$      end if
+!!$    end do
+!!$  end do
+!!$  call atmp%mv_from(acsr1)
+!!$
+!!$  call rtilde%mv_to(tmpcoo)
+!!$  nzl = tmpcoo%get_nzeros()
+!!$  i=0
+!!$  do k=1, nzl
+!!$    if ((naggrm1 < tmpcoo%ia(k)) .and. (tmpcoo%ia(k) <= naggrp1)) then
+!!$      i = i+1
+!!$      tmpcoo%val(i) = tmpcoo%val(k)
+!!$      tmpcoo%ia(i)  = tmpcoo%ia(k)
+!!$      tmpcoo%ja(i)  = tmpcoo%ja(k)
+!!$    end if
+!!$  end do
+!!$  call tmpcoo%set_nzeros(i)
+!!$  call rtilde%mv_from(tmpcoo)
+!!$  call rtilde%cscnv(info,type='csr')
+!!$
+!!$  call psb_spspmm(rtilde,atmp,op_restr,info)
+!!$
+!!$  !
+!!$  ! Now we have to gather the halo of op_prol, and add it to itself
+!!$  ! to multiply it by A,
+!!$  !
+!!$  call op_prol%clone(tmp_prol,info) 
+!!$  if (info == psb_success_) call psb_sphalo(tmp_prol,desc_a,am4,info,&
+!!$       & colcnv=.false.,rowscale=.true.)
+!!$  if (info == psb_success_) call psb_rwextd(ncol,tmp_prol,info,b=am4)      
+!!$  if (info == psb_success_) call am4%free()
+!!$
+!!$  if(info /= psb_success_) then
+!!$    call psb_errpush(psb_err_internal_error_,name,a_err='Halo of op_prol')
+!!$    goto 9999
+!!$  end if
+!!$
+!!$  !
+!!$  ! Now we have to fix this.  The only rows of B that are correct 
+!!$  ! are those corresponding to "local" aggregates, i.e. indices in ilaggr(:)
+!!$  !
+!!$  call op_restr%mv_to(tmpcoo)
+!!$
+!!$  nzl = tmpcoo%get_nzeros()
+!!$  i=0
+!!$  do k=1, nzl
+!!$    if ((naggrm1 < tmpcoo%ia(k)) .and. (tmpcoo%ia(k) <= naggrp1)) then
+!!$      i = i+1
+!!$      tmpcoo%val(i) = tmpcoo%val(k)
+!!$      tmpcoo%ia(i)  = tmpcoo%ia(k)
+!!$      tmpcoo%ja(i)  = tmpcoo%ja(k)
+!!$    end if
+!!$  end do
+!!$  call tmpcoo%set_nzeros(i)
+!!$  call op_restr%mv_from(tmpcoo)
+!!$  call op_restr%cscnv(info,type='csr')
+!!$
+!!$
+!!$  if (debug_level >= psb_debug_outer_) &
+!!$       & write(debug_unit,*) me,' ',trim(name),&
+!!$       & 'starting sphalo/ rwxtd'
+!!$
+!!$  call psb_spspmm(la,tmp_prol,am3,info)
+!!$  if (debug_level >= psb_debug_outer_) &
+!!$       & write(debug_unit,*) me,' ',trim(name),&
+!!$       & 'Done SPSPMM 2'
+!!$
+!!$  call psb_sphalo(am3,desc_a,am4,info,&
+!!$       & colcnv=.false.,rowscale=.true.)
+!!$  if (info == psb_success_) call psb_rwextd(ncol,am3,info,b=am4)      
+!!$  if (info == psb_success_) call am4%free()
+!!$
+!!$  if(info /= psb_success_) then
+!!$    call psb_errpush(psb_err_internal_error_,name,&
+!!$         & a_err='Extend am3')
+!!$    goto 9999
+!!$  end if
+!!$  if (debug_level >= psb_debug_outer_) &
+!!$       & write(debug_unit,*) me,' ',trim(name),&
+!!$       & 'Done sphalo/ rwxtd'
+!!$
+!!$  call psb_spspmm(op_restr,am3,ac,info)
+!!$  if (info == psb_success_) call am3%free()
+!!$  if (info == psb_success_) call ac%cscnv(info,type='coo',dupl=psb_dupl_add_)
+!!$
+!!$  if (info /= psb_success_) then
+!!$    call psb_errpush(psb_err_internal_error_,name,&
+!!$         &a_err='Build ac = op_restr x am3')
+!!$    goto 9999
+!!$  end if



@@ -116,7 +116,7 @@ subroutine amg_zaggrmat_smth_bld(a,desc_a,ilaggr,nlaggr,parms,&
  type(psb_desc_type), intent(inout)       :: desc_a
  integer(psb_lpk_), intent(inout)         :: ilaggr(:), nlaggr(:)
  type(amg_dml_parms), intent(inout)      :: parms 
-  type(psb_zspmat_type), intent(out)      :: op_prol,ac,op_restr
+  type(psb_zspmat_type), intent(inout)      :: op_prol,ac,op_restr
  type(psb_lzspmat_type), intent(inout)    :: t_prol
  type(psb_desc_type), intent(inout)       :: desc_ac
  integer(psb_ipk_), intent(out)           :: info
@@ -571,7 +571,6 @@ subroutine amg_ccprecsetc(p,what,string,info,ilev,ilmax,pos,idx)
              type(amg_c_krm_solver_type)          :: krm_slv
              call p%precv(nlev_)%set('SMOOTHER_TYPE',amg_krm_,info,pos=pos)
              call p%precv(nlev_)%set(krm_slv,info)
-              call p%precv(nlev_)%default()
              call p%precv(nlev_)%set('COARSE_MAT',amg_distr_mat_,info,pos=pos)
            end block
          end select
@@ -729,7 +728,6 @@ subroutine amg_ccprecsetc(p,what,string,info,ilev,ilmax,pos,idx)
              type(amg_c_krm_solver_type)          :: krm_slv
              call p%precv(nlev_)%set('SMOOTHER_TYPE',amg_krm_,info,pos=pos)
              call p%precv(nlev_)%set(krm_slv,info)
-              call p%precv(nlev_)%default()
              call p%precv(nlev_)%set('COARSE_MAT',amg_distr_mat_,info,pos=pos)
            end block
        end select
@@ -45,7 +45,7 @@ subroutine amg_cprecsetsm(p,val,info,ilev,ilmax,pos)
  implicit none

  ! Arguments
-  class(amg_cprec_type), intent(inout)         :: p
+  class(amg_cprec_type), target, intent(inout):: p
  class(amg_c_base_smoother_type), intent(in) :: val
  integer(psb_ipk_), intent(out)              :: info
  integer(psb_ipk_), optional, intent(in)     :: ilev,ilmax
@@ -599,7 +599,6 @@ subroutine amg_dcprecsetc(p,what,string,info,ilev,ilmax,pos,idx)
              type(amg_d_krm_solver_type)          :: krm_slv
              call p%precv(nlev_)%set('SMOOTHER_TYPE',amg_krm_,info,pos=pos)
              call p%precv(nlev_)%set(krm_slv,info)
-              call p%precv(nlev_)%default()
              call p%precv(nlev_)%set('COARSE_MAT',amg_distr_mat_,info,pos=pos)
            end block
          end select
@@ -773,7 +772,6 @@ subroutine amg_dcprecsetc(p,what,string,info,ilev,ilmax,pos,idx)
              type(amg_d_krm_solver_type)          :: krm_slv
              call p%precv(nlev_)%set('SMOOTHER_TYPE',amg_krm_,info,pos=pos)
              call p%precv(nlev_)%set(krm_slv,info)
-              call p%precv(nlev_)%default()
              call p%precv(nlev_)%set('COARSE_MAT',amg_distr_mat_,info,pos=pos)
            end block
        end select
@@ -45,7 +45,7 @@ subroutine amg_dprecsetsm(p,val,info,ilev,ilmax,pos)
  implicit none

  ! Arguments
-  class(amg_dprec_type), intent(inout)         :: p
+  class(amg_dprec_type), target, intent(inout):: p
  class(amg_d_base_smoother_type), intent(in) :: val
  integer(psb_ipk_), intent(out)              :: info
  integer(psb_ipk_), optional, intent(in)     :: ilev,ilmax
@@ -94,7 +94,7 @@ SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

 #define HANDLE_SIZE  8

-#if defined(SLUD_VERSION_63)
+#if (SLUD_VERSION_>=63) 
 typedef struct {
  SuperMatrix *A;
  dLUstruct_t *LUstruct;
@@ -135,7 +135,7 @@ int amg_dsludist_fact(int n, int nl, int nnzl, int ffstr,
    SuperMatrix *A;
    NRformat_loc *Astore;

-#if defined(SLUD_VERSION_63)
+#if (SLUD_VERSION_>=63)
    dScalePermstruct_t *ScalePermstruct;
    dLUstruct_t *LUstruct;
    dSOLVEstruct_t SOLVEstruct;
@@ -148,9 +148,9 @@ int amg_dsludist_fact(int n, int nl, int nnzl, int ffstr,
    int      i, panel_size, permc_spec, relax, info;
    trans_t  trans;
    double   drop_tol = 0.0, b[1], berr[1];
-#if defined(SLUD_VERSION_63) || defined(SLUD_VERSION_6)||defined(SLUD_VERSION_5)
+#if (SLUD_VERSION_>=50)
    superlu_dist_options_t options;
-#elif defined(SLUD_VERSION_4)||defined(SLUD_VERSION_3)
+#elif (SLUD_VERSION_>=30)
    superlu_options_t options;
 #else
    choke_on_me;
@@ -174,7 +174,7 @@ int amg_dsludist_fact(int n, int nl, int nnzl, int ffstr,
 				   SLU_NR_loc, SLU_D, SLU_GE);
    
    /* Initialize ScalePermstruct and LUstruct. */
-#if defined(SLUD_VERSION_63)
+#if (SLUD_VERSION_>=63) 
    ScalePermstruct = (dScalePermstruct_t *) SUPERLU_MALLOC(sizeof(dScalePermstruct_t));
    LUstruct = (dLUstruct_t *) SUPERLU_MALLOC(sizeof(dLUstruct_t));
    dScalePermstructInit(n,n, ScalePermstruct);
@@ -183,11 +183,11 @@ int amg_dsludist_fact(int n, int nl, int nnzl, int ffstr,
    LUstruct = (LUstruct_t *) SUPERLU_MALLOC(sizeof(LUstruct_t));
    ScalePermstructInit(n,n, ScalePermstruct);
 #endif
-#if defined(SLUD_VERSION_63)
+#if (SLUD_VERSION_>=63) 
    dLUstructInit(n, LUstruct);
-#elif defined(SLUD_VERSION_4) || defined(SLUD_VERSION_5) || defined(SLUD_VERSION_6)
+#elif (SLUD_VERSION_>=40) 
    LUstructInit(n, LUstruct);
-#elif defined(SLUD_VERSION_3) 
+#elif (SLUD_VERSION_>=30)
    LUstructInit(n,n, LUstruct);
 #else
    choke_on_me;
@@ -245,7 +245,7 @@ int amg_dsludist_solve(int itrans, int n, int nrhs,
 */
 #ifdef Have_SLUDist_ 
    SuperMatrix *A;
-#if defined(SLUD_VERSION_63)
+#if (SLUD_VERSION_>=63) 
    dScalePermstruct_t *ScalePermstruct;
    dLUstruct_t *LUstruct;
    dSOLVEstruct_t SOLVEstruct;
@@ -259,9 +259,9 @@ int amg_dsludist_solve(int itrans, int n, int nrhs,
    trans_t  trans;
    double   drop_tol = 0.0;
    double *berr;
-#if defined(SLUD_VERSION_63) || defined(SLUD_VERSION_6) ||defined(SLUD_VERSION_5)
+#if (SLUD_VERSION_>=50) 
    superlu_dist_options_t options;
-#elif defined(SLUD_VERSION_4)|| defined(SLUD_VERSION_3)
+#elif (SLUD_VERSION_>=30)
    superlu_options_t options;
 #else
    choke_on_me;
@@ -331,7 +331,7 @@ int amg_dsludist_free(void *f_factors)
 */
 #ifdef Have_SLUDist_ 
    SuperMatrix *A;
-#if defined(SLUD_VERSION_63)
+#if (SLUD_VERSION_>=63)
    dScalePermstruct_t *ScalePermstruct;
    dLUstruct_t *LUstruct;
    dSOLVEstruct_t SOLVEstruct;
@@ -345,9 +345,9 @@ int amg_dsludist_free(void *f_factors)
    trans_t  trans;
    double   drop_tol = 0.0;
    double *berr;
-#if defined(SLUD_VERSION_63)||defined(SLUD_VERSION_6)||defined(SLUD_VERSION_5)
+#if (SLUD_VERSION_>=50) 
    superlu_dist_options_t options;
-#elif defined(SLUD_VERSION_4)||defined(SLUD_VERSION_3)
+#elif (SLUD_VERSION_>=30)
    superlu_options_t options;
 #else
    choke_on_me;
@@ -368,7 +368,7 @@ int amg_dsludist_free(void *f_factors)
    // we either have a leak or a segfault here.
    // To be investigated further. 
    //Destroy_CompRowLoc_Matrix_dist(A);
-#if defined(SLUD_VERSION_63)
+#if (SLUD_VERSION_>=63)
    dScalePermstructFree(ScalePermstruct);
    dLUstructFree(LUstruct);
 #else
@@ -571,7 +571,6 @@ subroutine amg_scprecsetc(p,what,string,info,ilev,ilmax,pos,idx)
              type(amg_s_krm_solver_type)          :: krm_slv
              call p%precv(nlev_)%set('SMOOTHER_TYPE',amg_krm_,info,pos=pos)
              call p%precv(nlev_)%set(krm_slv,info)
-              call p%precv(nlev_)%default()
              call p%precv(nlev_)%set('COARSE_MAT',amg_distr_mat_,info,pos=pos)
            end block
          end select
@@ -729,7 +728,6 @@ subroutine amg_scprecsetc(p,what,string,info,ilev,ilmax,pos,idx)
              type(amg_s_krm_solver_type)          :: krm_slv
              call p%precv(nlev_)%set('SMOOTHER_TYPE',amg_krm_,info,pos=pos)
              call p%precv(nlev_)%set(krm_slv,info)
-              call p%precv(nlev_)%default()
              call p%precv(nlev_)%set('COARSE_MAT',amg_distr_mat_,info,pos=pos)
            end block
        end select
@@ -45,7 +45,7 @@ subroutine amg_sprecsetsm(p,val,info,ilev,ilmax,pos)
  implicit none

  ! Arguments
-  class(amg_sprec_type), intent(inout)         :: p
+  class(amg_sprec_type), target, intent(inout):: p
  class(amg_s_base_smoother_type), intent(in) :: val
  integer(psb_ipk_), intent(out)              :: info
  integer(psb_ipk_), optional, intent(in)     :: ilev,ilmax
@@ -599,7 +599,6 @@ subroutine amg_zcprecsetc(p,what,string,info,ilev,ilmax,pos,idx)
              type(amg_z_krm_solver_type)          :: krm_slv
              call p%precv(nlev_)%set('SMOOTHER_TYPE',amg_krm_,info,pos=pos)
              call p%precv(nlev_)%set(krm_slv,info)
-              call p%precv(nlev_)%default()
              call p%precv(nlev_)%set('COARSE_MAT',amg_distr_mat_,info,pos=pos)
            end block
          end select
@@ -773,7 +772,6 @@ subroutine amg_zcprecsetc(p,what,string,info,ilev,ilmax,pos,idx)
              type(amg_z_krm_solver_type)          :: krm_slv
              call p%precv(nlev_)%set('SMOOTHER_TYPE',amg_krm_,info,pos=pos)
              call p%precv(nlev_)%set(krm_slv,info)
-              call p%precv(nlev_)%default()
              call p%precv(nlev_)%set('COARSE_MAT',amg_distr_mat_,info,pos=pos)
            end block
        end select
@@ -45,7 +45,7 @@ subroutine amg_zprecsetsm(p,val,info,ilev,ilmax,pos)
  implicit none

  ! Arguments
-  class(amg_zprec_type), intent(inout)         :: p
+  class(amg_zprec_type), target, intent(inout):: p
  class(amg_z_base_smoother_type), intent(in) :: val
  integer(psb_ipk_), intent(out)              :: info
  integer(psb_ipk_), optional, intent(in)     :: ilev,ilmax
@@ -94,7 +94,7 @@ SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

 #define HANDLE_SIZE  8

-#if defined(SLUD_VERSION_63)
+#if (SLUD_VERSION_>=63)
 typedef struct {
  SuperMatrix *A;
  zLUstruct_t *LUstruct;
@@ -142,7 +142,7 @@ int amg_zsludist_fact(int n, int nl, int nnzl, int ffstr,
    SuperMatrix *A;
    NRformat_loc *Astore;

-#if defined(SLUD_VERSION_63)
+#if (SLUD_VERSION_>=63)
    zScalePermstruct_t *ScalePermstruct;
    zLUstruct_t *LUstruct;
    zSOLVEstruct_t SOLVEstruct;
@@ -155,9 +155,9 @@ int amg_zsludist_fact(int n, int nl, int nnzl, int ffstr,
    int      i, panel_size, permc_spec, relax, info;
    trans_t  trans;
    double   drop_tol = 0.0,berr[1];
-#if defined(SLUD_VERSION_63) || defined(SLUD_VERSION_6)||defined(SLUD_VERSION_5)
+#if (SLUD_VERSION_>=50) 
    superlu_dist_options_t options;
-#elif defined(SLUD_VERSION_4)||defined(SLUD_VERSION_3)
+#elif (SLUD_VERSION_>=30)
    superlu_options_t options;
 #else
    choke_on_me;
@@ -181,7 +181,7 @@ int amg_zsludist_fact(int n, int nl, int nnzl, int ffstr,
 				   SLU_NR_loc, SLU_Z, SLU_GE);
    
    /* Initialize ScalePermstruct and LUstruct. */
-#if defined(SLUD_VERSION_63)
+#if (SLUD_VERSION_>=63)
    ScalePermstruct = (zScalePermstruct_t *) SUPERLU_MALLOC(sizeof(zScalePermstruct_t));
    LUstruct = (zLUstruct_t *) SUPERLU_MALLOC(sizeof(zLUstruct_t));
    zScalePermstructInit(n,n, ScalePermstruct);
@@ -190,11 +190,11 @@ int amg_zsludist_fact(int n, int nl, int nnzl, int ffstr,
    LUstruct = (LUstruct_t *) SUPERLU_MALLOC(sizeof(LUstruct_t));
    ScalePermstructInit(n,n, ScalePermstruct);
 #endif
-#if defined(SLUD_VERSION_63)
+#if (SLUD_VERSION_>=63)
    zLUstructInit(n, LUstruct);
-#elif defined(SLUD_VERSION_4) || defined(SLUD_VERSION_5) || defined(SLUD_VERSION_6)
+#elif (SLUD_VERSION_>=40) 
    LUstructInit(n, LUstruct);
-#elif defined(SLUD_VERSION_3)
+#elif (SLUD_VERSION_>=30)
    LUstructInit(n,n, LUstruct);
 #else
    choke_on_me;
@@ -257,7 +257,7 @@ int amg_zsludist_solve(int itrans, int n, int nrhs,
 */
 #ifdef Have_SLUDist_ 
    SuperMatrix *A;
-#if defined(SLUD_VERSION_63)
+#if (SLUD_VERSION_>=63) 
    zScalePermstruct_t *ScalePermstruct;
    zLUstruct_t *LUstruct;
    zSOLVEstruct_t SOLVEstruct;
@@ -271,9 +271,9 @@ int amg_zsludist_solve(int itrans, int n, int nrhs,
    trans_t  trans;
    double   drop_tol = 0.0;
    double *berr;
-#if defined(SLUD_VERSION_63) || defined(SLUD_VERSION_6) ||defined(SLUD_VERSION_5)
+#if (SLUD_VERSION_>=50) 
    superlu_dist_options_t options;
-#elif defined(SLUD_VERSION_4)|| defined(SLUD_VERSION_3)
+#elif (SLUD_VERSION_>=30)
    superlu_options_t options;
 #else
    choke_on_me;
@@ -343,7 +343,7 @@ int amg_zsludist_free(void *f_factors)
 */
 #ifdef Have_SLUDist_ 
    SuperMatrix *A;
-#if defined(SLUD_VERSION_63)
+#if (SLUD_VERSION_>=63) 
    zScalePermstruct_t *ScalePermstruct;
    zLUstruct_t *LUstruct;
    zSOLVEstruct_t SOLVEstruct;
@@ -357,9 +357,9 @@ int amg_zsludist_free(void *f_factors)
    trans_t  trans;
    double   drop_tol = 0.0;
    double *berr;
-#if defined(SLUD_VERSION_63)||defined(SLUD_VERSION_6)||defined(SLUD_VERSION_5)
+#if (SLUD_VERSION_>=50) 
    superlu_dist_options_t options;
-#elif defined(SLUD_VERSION_4)||defined(SLUD_VERSION_3)
+#elif (SLUD_VERSION_>=30)
    superlu_options_t options;
 #else
    choke_on_me;
@@ -380,7 +380,7 @@ int amg_zsludist_free(void *f_factors)
    // we either have a leak or a segfault here.
    // To be investigated further. 
    //Destroy_CompRowLoc_Matrix_dist(A);
-#if defined(SLUD_VERSION_63)
+#if (SLUD_VERSION_>=63) 
    zScalePermstructFree(ScalePermstruct);
    zLUstructFree(LUstruct);
 #else
@@ -83,6 +83,8 @@ subroutine amg_c_base_onelev_descr(lv,il,nl,ilmin,info,iout,verbosity)
  write(iout_,*) 
  if (il == ilmin) then 
    call lv%parms%mlcycledsc(iout_,info)
+  end if
+  if (((ilmin==1).and.(il==2)).or.((ilmin>1).and.(il==ilmin))) then 
    if (allocated(lv%aggr)) then
      call lv%aggr%descr(lv%parms,iout_,info)
    else
@@ -101,7 +101,13 @@ subroutine amg_c_base_onelev_dump(lv,level,info,prefix,head,ac,rp,&
  end if

  if (global_num_) then
-    if (level >= 2) then 
+    if (level == 1) then
+      if (ac_) then
+        ivr = lv%base_desc%get_global_indices(owned=.false.)
+        write(fname(lname+1:),'(a,i3.3,a)')'_l',level,'_ac.mtx'
+        call lv%base_a%print(fname,head=head,iv=ivr)
+      end if
+    else if (level >= 2) then 
      if (ac_) then
        ivr = lv%desc_ac%get_global_indices(owned=.false.)
        write(fname(lname+1:),'(a,i3.3,a)')'_l',level,'_ac.mtx'
@@ -126,7 +132,12 @@ subroutine amg_c_base_onelev_dump(lv,level,info,prefix,head,ac,rp,&
      end if
    end if
  else
-    if (level >= 2) then 
+    if (level == 1) then
+      if (ac_) then
+        write(fname(lname+1:),'(a,i3.3,a)')'_l',level,'_ac.mtx'
+        call lv%base_a%print(fname,head=head)
+      end if
+    else  if (level >= 2) then 
      if (ac_) then 
        write(fname(lname+1:),'(a,i3.3,a)')'_l',level,'_ac.mtx'
        call lv%ac%print(fname,head=head)
@@ -146,16 +157,7 @@ subroutine amg_c_base_onelev_dump(lv,level,info,prefix,head,ac,rp,&
    end if
  end if
  
-  if (level >= 2) then 
-    if (allocated(lv%sm)) then
-      call lv%sm%dump(lv%desc_ac,level,info,smoother=smoother, &
-           & solver=solver,prefix=trim(prefix_)//"_sm",global_num=global_num)
-    end if
-    if (allocated(lv%sm2a)) then
-      call lv%sm2a%dump(lv%desc_ac,level,info,smoother=smoother, &
-           & solver=solver,prefix=trim(prefix_)//"_sm2a",global_num=global_num)
-    end if
-  else
+  if (level >= 1) then 
    if (allocated(lv%sm)) then
      call lv%sm%dump(lv%base_desc,level,info,smoother=smoother, &
           & solver=solver,prefix=trim(prefix_)//"_sm",global_num=global_num)
@@ -83,6 +83,8 @@ subroutine amg_d_base_onelev_descr(lv,il,nl,ilmin,info,iout,verbosity)
  write(iout_,*) 
  if (il == ilmin) then 
    call lv%parms%mlcycledsc(iout_,info)
+  end if
+  if (((ilmin==1).and.(il==2)).or.((ilmin>1).and.(il==ilmin))) then 
    if (allocated(lv%aggr)) then
      call lv%aggr%descr(lv%parms,iout_,info)
    else
@@ -101,7 +101,13 @@ subroutine amg_d_base_onelev_dump(lv,level,info,prefix,head,ac,rp,&
  end if

  if (global_num_) then
-    if (level >= 2) then 
+    if (level == 1) then
+      if (ac_) then
+        ivr = lv%base_desc%get_global_indices(owned=.false.)
+        write(fname(lname+1:),'(a,i3.3,a)')'_l',level,'_ac.mtx'
+        call lv%base_a%print(fname,head=head,iv=ivr)
+      end if
+    else if (level >= 2) then 
      if (ac_) then
        ivr = lv%desc_ac%get_global_indices(owned=.false.)
        write(fname(lname+1:),'(a,i3.3,a)')'_l',level,'_ac.mtx'
@@ -126,7 +132,12 @@ subroutine amg_d_base_onelev_dump(lv,level,info,prefix,head,ac,rp,&
      end if
    end if
  else
-    if (level >= 2) then 
+    if (level == 1) then
+      if (ac_) then
+        write(fname(lname+1:),'(a,i3.3,a)')'_l',level,'_ac.mtx'
+        call lv%base_a%print(fname,head=head)
+      end if
+    else  if (level >= 2) then 
      if (ac_) then 
        write(fname(lname+1:),'(a,i3.3,a)')'_l',level,'_ac.mtx'
        call lv%ac%print(fname,head=head)
@@ -146,16 +157,7 @@ subroutine amg_d_base_onelev_dump(lv,level,info,prefix,head,ac,rp,&
    end if
  end if
  
-  if (level >= 2) then 
-    if (allocated(lv%sm)) then
-      call lv%sm%dump(lv%desc_ac,level,info,smoother=smoother, &
-           & solver=solver,prefix=trim(prefix_)//"_sm",global_num=global_num)
-    end if
-    if (allocated(lv%sm2a)) then
-      call lv%sm2a%dump(lv%desc_ac,level,info,smoother=smoother, &
-           & solver=solver,prefix=trim(prefix_)//"_sm2a",global_num=global_num)
-    end if
-  else
+  if (level >= 1) then 
    if (allocated(lv%sm)) then
      call lv%sm%dump(lv%base_desc,level,info,smoother=smoother, &
           & solver=solver,prefix=trim(prefix_)//"_sm",global_num=global_num)
@@ -83,6 +83,8 @@ subroutine amg_s_base_onelev_descr(lv,il,nl,ilmin,info,iout,verbosity)
  write(iout_,*) 
  if (il == ilmin) then 
    call lv%parms%mlcycledsc(iout_,info)
+  end if
+  if (((ilmin==1).and.(il==2)).or.((ilmin>1).and.(il==ilmin))) then 
    if (allocated(lv%aggr)) then
      call lv%aggr%descr(lv%parms,iout_,info)
    else
@@ -101,7 +101,13 @@ subroutine amg_s_base_onelev_dump(lv,level,info,prefix,head,ac,rp,&
  end if

  if (global_num_) then
-    if (level >= 2) then 
+    if (level == 1) then
+      if (ac_) then
+        ivr = lv%base_desc%get_global_indices(owned=.false.)
+        write(fname(lname+1:),'(a,i3.3,a)')'_l',level,'_ac.mtx'
+        call lv%base_a%print(fname,head=head,iv=ivr)
+      end if
+    else if (level >= 2) then 
      if (ac_) then
        ivr = lv%desc_ac%get_global_indices(owned=.false.)
        write(fname(lname+1:),'(a,i3.3,a)')'_l',level,'_ac.mtx'
@@ -126,7 +132,12 @@ subroutine amg_s_base_onelev_dump(lv,level,info,prefix,head,ac,rp,&
      end if
    end if
  else
-    if (level >= 2) then 
+    if (level == 1) then
+      if (ac_) then
+        write(fname(lname+1:),'(a,i3.3,a)')'_l',level,'_ac.mtx'
+        call lv%base_a%print(fname,head=head)
+      end if
+    else  if (level >= 2) then 
      if (ac_) then 
        write(fname(lname+1:),'(a,i3.3,a)')'_l',level,'_ac.mtx'
        call lv%ac%print(fname,head=head)
@@ -146,16 +157,7 @@ subroutine amg_s_base_onelev_dump(lv,level,info,prefix,head,ac,rp,&
    end if
  end if
  
-  if (level >= 2) then 
-    if (allocated(lv%sm)) then
-      call lv%sm%dump(lv%desc_ac,level,info,smoother=smoother, &
-           & solver=solver,prefix=trim(prefix_)//"_sm",global_num=global_num)
-    end if
-    if (allocated(lv%sm2a)) then
-      call lv%sm2a%dump(lv%desc_ac,level,info,smoother=smoother, &
-           & solver=solver,prefix=trim(prefix_)//"_sm2a",global_num=global_num)
-    end if
-  else
+  if (level >= 1) then 
    if (allocated(lv%sm)) then
      call lv%sm%dump(lv%base_desc,level,info,smoother=smoother, &
           & solver=solver,prefix=trim(prefix_)//"_sm",global_num=global_num)
@@ -83,6 +83,8 @@ subroutine amg_z_base_onelev_descr(lv,il,nl,ilmin,info,iout,verbosity)
  write(iout_,*) 
  if (il == ilmin) then 
    call lv%parms%mlcycledsc(iout_,info)
+  end if
+  if (((ilmin==1).and.(il==2)).or.((ilmin>1).and.(il==ilmin))) then 
    if (allocated(lv%aggr)) then
      call lv%aggr%descr(lv%parms,iout_,info)
    else
@@ -101,7 +101,13 @@ subroutine amg_z_base_onelev_dump(lv,level,info,prefix,head,ac,rp,&
  end if

  if (global_num_) then
-    if (level >= 2) then 
+    if (level == 1) then
+      if (ac_) then
+        ivr = lv%base_desc%get_global_indices(owned=.false.)
+        write(fname(lname+1:),'(a,i3.3,a)')'_l',level,'_ac.mtx'
+        call lv%base_a%print(fname,head=head,iv=ivr)
+      end if
+    else if (level >= 2) then 
      if (ac_) then
        ivr = lv%desc_ac%get_global_indices(owned=.false.)
        write(fname(lname+1:),'(a,i3.3,a)')'_l',level,'_ac.mtx'
@@ -126,7 +132,12 @@ subroutine amg_z_base_onelev_dump(lv,level,info,prefix,head,ac,rp,&
      end if
    end if
  else
-    if (level >= 2) then 
+    if (level == 1) then
+      if (ac_) then
+        write(fname(lname+1:),'(a,i3.3,a)')'_l',level,'_ac.mtx'
+        call lv%base_a%print(fname,head=head)
+      end if
+    else  if (level >= 2) then 
      if (ac_) then 
        write(fname(lname+1:),'(a,i3.3,a)')'_l',level,'_ac.mtx'
        call lv%ac%print(fname,head=head)
@@ -146,16 +157,7 @@ subroutine amg_z_base_onelev_dump(lv,level,info,prefix,head,ac,rp,&
    end if
  end if
  
-  if (level >= 2) then 
-    if (allocated(lv%sm)) then
-      call lv%sm%dump(lv%desc_ac,level,info,smoother=smoother, &
-           & solver=solver,prefix=trim(prefix_)//"_sm",global_num=global_num)
-    end if
-    if (allocated(lv%sm2a)) then
-      call lv%sm2a%dump(lv%desc_ac,level,info,smoother=smoother, &
-           & solver=solver,prefix=trim(prefix_)//"_sm2a",global_num=global_num)
-    end if
-  else
+  if (level >= 1) then 
    if (allocated(lv%sm)) then
      call lv%sm%dump(lv%base_desc,level,info,smoother=smoother, &
           & solver=solver,prefix=trim(prefix_)//"_sm",global_num=global_num)
@@ -61,6 +61,7 @@ subroutine amg_c_as_smoother_free(sm,info)
    end if
  end if
  call sm%nd%free()
+  call sm%desc_data%free(info)

  call psb_erractionrestore(err_act)
  return
@@ -61,6 +61,7 @@ subroutine amg_d_as_smoother_free(sm,info)
    end if
  end if
  call sm%nd%free()
+  call sm%desc_data%free(info)

  call psb_erractionrestore(err_act)
  return
@@ -61,6 +61,7 @@ subroutine amg_s_as_smoother_free(sm,info)
    end if
  end if
  call sm%nd%free()
+  call sm%desc_data%free(info)

  call psb_erractionrestore(err_act)
  return
@@ -61,6 +61,7 @@ subroutine amg_z_as_smoother_free(sm,info)
    end if
  end if
  call sm%nd%free()
+  call sm%desc_data%free(info)

  call psb_erractionrestore(err_act)
  return
@@ -42,7 +42,7 @@ subroutine amg_c_base_solver_csetr(sv,what,val,info,idx)
  Implicit None
  ! Arguments
  class(amg_c_base_solver_type), intent(inout) :: sv 
-  integer(psb_ipk_), intent(in)                  :: what 
+  character(len=*), intent(in)                   :: what 
  real(psb_spk_), intent(in)                      :: val
  integer(psb_ipk_), intent(out)                 :: info
  integer(psb_ipk_), intent(in), optional        :: idx
@@ -44,7 +44,7 @@ subroutine amg_c_bwgs_solver_bld(a,desc_a,sv,info,b,amold,vmold,imold)

  ! Arguments
  type(psb_cspmat_type), intent(in), target           :: a
-  Type(psb_desc_type), Intent(in)                     :: desc_a 
+  Type(psb_desc_type), Intent(inout)                  :: desc_a 
  class(amg_c_bwgs_solver_type), intent(inout)         :: sv
  integer(psb_ipk_), intent(out)                      :: info
  type(psb_cspmat_type), intent(in), target, optional :: b
@@ -77,7 +77,7 @@
 !  This is the implementation file corresponding to amg_c_krm_solver_mod.
 !
 !
-subroutine amg_c_krm_solver_bld(a,desc_a,sv,info,b,amold,vmold)
+subroutine amg_c_krm_solver_bld(a,desc_a,sv,info,b,amold,vmold,imold)

  use psb_base_mod
  use amg_c_krm_solver, amg_protect_name => amg_c_krm_solver_bld
@@ -85,13 +85,14 @@ subroutine amg_c_krm_solver_bld(a,desc_a,sv,info,b,amold,vmold)
  Implicit None

  ! Arguments
-  type(psb_cspmat_type), intent(inout), target        :: a
+  type(psb_cspmat_type), intent(in), target        :: a
  Type(psb_desc_type), Intent(inout)                  :: desc_a
  class(amg_c_krm_solver_type), intent(inout)         :: sv
  integer(psb_ipk_), intent(out)                      :: info
  type(psb_cspmat_type), intent(in), target, optional :: b
  class(psb_c_base_sparse_mat), intent(in), optional  :: amold
  class(psb_c_base_vect_type), intent(in), optional   :: vmold
+  class(psb_i_base_vect_type), intent(in), optional   :: imold
  ! Local variables
  integer(psb_ipk_) :: n_row,n_col, nrow_a, nztota
  integer(psb_lpk_) :: lnr
@@ -42,7 +42,7 @@ subroutine amg_d_base_solver_csetr(sv,what,val,info,idx)
  Implicit None
  ! Arguments
  class(amg_d_base_solver_type), intent(inout) :: sv 
-  integer(psb_ipk_), intent(in)                  :: what 
+  character(len=*), intent(in)                   :: what 
  real(psb_dpk_), intent(in)                      :: val
  integer(psb_ipk_), intent(out)                 :: info
  integer(psb_ipk_), intent(in), optional        :: idx
@@ -44,7 +44,7 @@ subroutine amg_d_bwgs_solver_bld(a,desc_a,sv,info,b,amold,vmold,imold)

  ! Arguments
  type(psb_dspmat_type), intent(in), target           :: a
-  Type(psb_desc_type), Intent(in)                     :: desc_a 
+  Type(psb_desc_type), Intent(inout)                  :: desc_a 
  class(amg_d_bwgs_solver_type), intent(inout)         :: sv
  integer(psb_ipk_), intent(out)                      :: info
  type(psb_dspmat_type), intent(in), target, optional :: b
@@ -77,7 +77,7 @@
 !  This is the implementation file corresponding to amg_d_krm_solver_mod.
 !
 !
-subroutine amg_d_krm_solver_bld(a,desc_a,sv,info,b,amold,vmold)
+subroutine amg_d_krm_solver_bld(a,desc_a,sv,info,b,amold,vmold,imold)

  use psb_base_mod
  use amg_d_krm_solver, amg_protect_name => amg_d_krm_solver_bld
@@ -85,13 +85,14 @@ subroutine amg_d_krm_solver_bld(a,desc_a,sv,info,b,amold,vmold)
  Implicit None

  ! Arguments
-  type(psb_dspmat_type), intent(inout), target        :: a
+  type(psb_dspmat_type), intent(in), target        :: a
  Type(psb_desc_type), Intent(inout)                  :: desc_a
  class(amg_d_krm_solver_type), intent(inout)         :: sv
  integer(psb_ipk_), intent(out)                      :: info
  type(psb_dspmat_type), intent(in), target, optional :: b
  class(psb_d_base_sparse_mat), intent(in), optional  :: amold
  class(psb_d_base_vect_type), intent(in), optional   :: vmold
+  class(psb_i_base_vect_type), intent(in), optional   :: imold
  ! Local variables
  integer(psb_ipk_) :: n_row,n_col, nrow_a, nztota
  integer(psb_lpk_) :: lnr
@@ -42,7 +42,7 @@ subroutine amg_s_base_solver_csetr(sv,what,val,info,idx)
  Implicit None
  ! Arguments
  class(amg_s_base_solver_type), intent(inout) :: sv 
-  integer(psb_ipk_), intent(in)                  :: what 
+  character(len=*), intent(in)                   :: what 
  real(psb_spk_), intent(in)                      :: val
  integer(psb_ipk_), intent(out)                 :: info
  integer(psb_ipk_), intent(in), optional        :: idx
@@ -44,7 +44,7 @@ subroutine amg_s_bwgs_solver_bld(a,desc_a,sv,info,b,amold,vmold,imold)

  ! Arguments
  type(psb_sspmat_type), intent(in), target           :: a
-  Type(psb_desc_type), Intent(in)                     :: desc_a 
+  Type(psb_desc_type), Intent(inout)                  :: desc_a 
  class(amg_s_bwgs_solver_type), intent(inout)         :: sv
  integer(psb_ipk_), intent(out)                      :: info
  type(psb_sspmat_type), intent(in), target, optional :: b
@@ -77,7 +77,7 @@
 !  This is the implementation file corresponding to amg_s_krm_solver_mod.
 !
 !
-subroutine amg_s_krm_solver_bld(a,desc_a,sv,info,b,amold,vmold)
+subroutine amg_s_krm_solver_bld(a,desc_a,sv,info,b,amold,vmold,imold)

  use psb_base_mod
  use amg_s_krm_solver, amg_protect_name => amg_s_krm_solver_bld
@@ -85,13 +85,14 @@ subroutine amg_s_krm_solver_bld(a,desc_a,sv,info,b,amold,vmold)
  Implicit None

  ! Arguments
-  type(psb_sspmat_type), intent(inout), target        :: a
+  type(psb_sspmat_type), intent(in), target        :: a
  Type(psb_desc_type), Intent(inout)                  :: desc_a
  class(amg_s_krm_solver_type), intent(inout)         :: sv
  integer(psb_ipk_), intent(out)                      :: info
  type(psb_sspmat_type), intent(in), target, optional :: b
  class(psb_s_base_sparse_mat), intent(in), optional  :: amold
  class(psb_s_base_vect_type), intent(in), optional   :: vmold
+  class(psb_i_base_vect_type), intent(in), optional   :: imold
  ! Local variables
  integer(psb_ipk_) :: n_row,n_col, nrow_a, nztota
  integer(psb_lpk_) :: lnr
@@ -42,7 +42,7 @@ subroutine amg_z_base_solver_csetr(sv,what,val,info,idx)
  Implicit None
  ! Arguments
  class(amg_z_base_solver_type), intent(inout) :: sv 
-  integer(psb_ipk_), intent(in)                  :: what 
+  character(len=*), intent(in)                   :: what 
  real(psb_dpk_), intent(in)                      :: val
  integer(psb_ipk_), intent(out)                 :: info
  integer(psb_ipk_), intent(in), optional        :: idx
@@ -44,7 +44,7 @@ subroutine amg_z_bwgs_solver_bld(a,desc_a,sv,info,b,amold,vmold,imold)

  ! Arguments
  type(psb_zspmat_type), intent(in), target           :: a
-  Type(psb_desc_type), Intent(in)                     :: desc_a 
+  Type(psb_desc_type), Intent(inout)                  :: desc_a 
  class(amg_z_bwgs_solver_type), intent(inout)         :: sv
  integer(psb_ipk_), intent(out)                      :: info
  type(psb_zspmat_type), intent(in), target, optional :: b
@@ -77,7 +77,7 @@
 !  This is the implementation file corresponding to amg_z_krm_solver_mod.
 !
 !
-subroutine amg_z_krm_solver_bld(a,desc_a,sv,info,b,amold,vmold)
+subroutine amg_z_krm_solver_bld(a,desc_a,sv,info,b,amold,vmold,imold)

  use psb_base_mod
  use amg_z_krm_solver, amg_protect_name => amg_z_krm_solver_bld
@@ -85,13 +85,14 @@ subroutine amg_z_krm_solver_bld(a,desc_a,sv,info,b,amold,vmold)
  Implicit None

  ! Arguments
-  type(psb_zspmat_type), intent(inout), target        :: a
+  type(psb_zspmat_type), intent(in), target        :: a
  Type(psb_desc_type), Intent(inout)                  :: desc_a
  class(amg_z_krm_solver_type), intent(inout)         :: sv
  integer(psb_ipk_), intent(out)                      :: info
  type(psb_zspmat_type), intent(in), target, optional :: b
  class(psb_z_base_sparse_mat), intent(in), optional  :: amold
  class(psb_z_base_vect_type), intent(in), optional   :: vmold
+  class(psb_i_base_vect_type), intent(in), optional   :: imold
  ! Local variables
  integer(psb_ipk_) :: n_row,n_col, nrow_a, nztota
  integer(psb_lpk_) :: lnr
@@ -852,6 +852,7 @@ if test "x$pac_sludist_header_ok" == "xyes" ; then
     dnl Maybe lib?
     SLUDIST_LIBS="$amg4psblas_cv_superludist  -L$amg4psblas_cv_superludistdir/lib";
     LIBS="$SLUDIST_LIBS -lm $save_LIBS";
+     AC_MSG_CHECKING([for superlu_malloc_dist in $SLUDIST_LIBS])
     AC_TRY_LINK_FUNC(superlu_malloc_dist, 
 		     [amg4psblas_cv_have_superludist=yes;pac_sludist_lib_ok=yes;],
 		     [amg4psblas_cv_have_superludist=no;pac_sludist_lib_ok=no; 
@@ -861,12 +862,13 @@ if test "x$pac_sludist_header_ok" == "xyes" ; then
     dnl Maybe lib64?
     SLUDIST_LIBS="$amg4psblas_cv_superludist  -L$amg4psblas_cv_superludistdir/lib64";
     LIBS="$SLUDIST_LIBS -lm $save_LIBS";
+     AC_MSG_CHECKING([for superlu_malloc_dist in $SLUDIST_LIBS])
     AC_TRY_LINK_FUNC(superlu_malloc_dist, 
 		     [amg4psblas_cv_have_superludist=yes;pac_sludist_lib_ok=yes;],
 		     [amg4psblas_cv_have_superludist=no;pac_sludist_lib_ok=no; 
 		      SLUDIST_LIBS="";SLUDIST_INCLUDES=""])
  fi
-  AC_MSG_RESULT($pac_sludist_lib_ok)
+  AC_MSG_RESULT([$pac_sludist_lib_ok $SLUDIST_LIBS])
 fi

 if test "x$pac_sludist_lib_ok" == "xyes" ; then
@@ -133,6 +133,9 @@ FCFLAGS="$save_FCFLAGS";
 save_CFLAGS="$CFLAGS";
 AC_PROG_CC([cc xlc pgcc icc gcc ])
 CFLAGS="$save_CFLAGS";
+save_CXXFLAGS="$CXXFLAGS";
+AC_PROG_CXX([CC xlc++  icpc g++])
+CXXFLAGS="$save_CXXFLAGS";
 dnl AC_PROG_CXX

 dnl AC_PROG_F90 doesn't exist, at the time of writing this !
@@ -158,6 +161,16 @@ if eval "$FC -qversion 2>&1 | grep XL 2>/dev/null" ; then
 # since (as far as it is known to us) -WF, is not used in earlier versions.
 # More problems could be undocumented yet.
 fi 
+
+if test "X$CC" == "X" ; then
+	AC_MSG_ERROR([Problem : No C compiler specified nor found!])
+fi
+AC_PROG_CC_STDC()
+if test "x$ac_cv_prog_cc_stdc" == "xno" ; then
+   AC_MSG_ERROR([Problem : Need a C99 compiler ! ])
+else
+   C99OPT="$ac_cv_prog_cc_stdc";
+fi	
 ###############################################################################
 #			Suitable MPI compilers detection
 ###############################################################################
@@ -171,6 +184,7 @@ if test x"$pac_cv_serial_mpi" == x"yes" ; then
   FAKEMPI="fakempi.o";
   MPIFC="$FC";
   MPICC="$CC";
+   MPICXX="$CXX";	
   CXXDEFINES="-DSERIAL_MPI $CXXDEFINES";

 else 
@@ -190,9 +204,17 @@ if test "X$MPIFC" = "X" ; then
 fi

 ACX_MPI([], [AC_MSG_ERROR([[Cannot find any suitable MPI implementation for Fortran]])])
+AC_LANG([C++])
+if test "X$MPICXX" = "X" ; then
+    # This is our MPICC compiler preference: it will override ACX_MPI's first try.
+    AC_CHECK_PROGS([MPICXX],[mpxlc++ mpiicpc  mpicxx])
+fi
+ACX_MPI([], [AC_MSG_ERROR([[Cannot find any suitable MPI implementation for C++]])])
+AC_LANG([Fortran])

 FC="$MPIFC" ;
 CC="$MPICC";
+CXX="$MPICXX";
 fi

 AC_LANG([C])
@@ -217,10 +239,12 @@ fi

 dnl NOTE : no spaces before the comma, and no brackets before the second argument!
 PAC_ARG_WITH_FLAGS(ccopt,CCOPT)
+PAC_ARG_WITH_FLAGS(cxxopt,CXXOPT)
 PAC_ARG_WITH_FLAGS(fcopt,FCOPT)
 PAC_ARG_WITH_LIBS
 PAC_ARG_WITH_FLAGS(clibs,CLIBS)
 PAC_ARG_WITH_FLAGS(flibs,FLIBS)
+PAC_ARG_WITH_FLAGS(cxxlibs,CXXLIBS)

 dnl candidates for removal:
 PAC_ARG_WITH_FLAGS(library-path,LIBRARYPATH)
@@ -230,6 +254,20 @@ PAC_ARG_WITH_FLAGS(module-path,MODULE_PATH)
 # we just gave the user the chance to append values to these variables
 PAC_ARG_WITH_EXTRA_LIBS

+# Check if we need extra libs (e.g. for OpenMPI)
+AC_MSG_CHECKING([Extra mpicxx libs?])
+xtrlibs=`mpicxx --showme:link 2>/dev/null`;
+if (( $? == 0 ))
+then
+  EXTRA_LIBS="$EXTRA_LIBS $xtrlibs";
+  AC_MSG_RESULT([$xtrlibs])
+else
+  AC_MSG_RESULT([none])
+fi
+
+
+ 
+
 ###############################################################################
 # Sanity checks, although redundant (useful when debugging this configure.ac)!
 ###############################################################################
@@ -395,6 +433,36 @@ if test "X$CCOPT" == "X" ; then
 fi
 fi
 #CFLAGS="${CCOPT}"
+if test "X$CXXOPT" == "X" ; then
+  CXXOPT="$CXXFLAGS";
+fi	
+if test "X$CXXOPT" == "X" ; then
+ if test "X$psblas_cv_fc" == "Xgcc" ; then
+ # note that no space should be placed around the equality symbol in assignements
+ # Note : 'native' is valid _only_ on GCC/x86  (32/64 bits)
+   CXXOPT="-g -O3  $CXXOPT"
+
+ elif test "X$psblas_cv_fc" == X"xlf" ; then
+   # XL compiler : consider using -qarch=auto
+     CXXOPT="-O3 -qarch=auto $CXXOPT"
+ elif test "X$psblas_cv_fc" == X"ifc" ; then
+   # other compilers ..
+     CXXOPT="-O3 $CXXOPT"
+ elif test "X$psblas_cv_fc" == X"pg" ; then
+   # other compilers ..
+     CXXCOPT="-fast $CXXOPT"
+  # NOTE : PG & Sun use -fast instead -O3
+ elif test "X$psblas_cv_fc" == X"sun" ; then
+   # other compilers ..
+     CXXOPT="-fast $CXXOPT"
+ elif test "X$psblas_cv_fc" == X"cray" ; then
+     CXXOPT="-O3 $CXXOPT"
+     MPICXX="CC"
+ else 
+     CXXOPT="-g -O3 $CXXOPT"
+ fi
+fi
+

 # Honor FCFLAGS if they were specified explicitly, but --with-fcopt take precedence
 if test "X$FCOPT" == "X" ; then
@@ -443,7 +511,8 @@ fi
 ##############################################################################
 FC=${FC}
 CC=${CC}
-MPCC=${MPICC}
+CXX=${CXX}
+CCOPT="$CCOPT $C99OPT"


 ##############################################################################
@@ -478,6 +547,30 @@ fi
 if test "X$FLINK" == "X" ; then
   FLINK=${MPF90}
 fi
+# Custom test : do we have a module or include for MPI Fortran interface?
+if test x"$pac_cv_serial_mpi" == x"yes" ; then
+   FDEFINES="$psblas_cv_define_prepend-DSERIAL_MPI $psblas_cv_define_prepend-DMPI_MOD $FDEFINES";
+else 
+ PAC_FORTRAN_CHECK_HAVE_MPI_MOD_F08()
+ if test x"$pac_cv_mpi_f08" == x"yes" ; then
+dnl    FDEFINES="$psblas_cv_define_prepend-DMPI_MOD_F08 $FDEFINES";
+    FDEFINES="$psblas_cv_define_prepend-DMPI_MOD $FDEFINES";
+  else					     	     
+    PAC_FORTRAN_CHECK_HAVE_MPI_MOD(
+ 	  [FDEFINES="$psblas_cv_define_prepend-DMPI_MOD $FDEFINES"],
+	  [FDEFINES="$psblas_cv_define_prepend-DMPI_H $FDEFINES"])	  
+ fi
+fi
+
+FLINK="$MPIFC"
+PAC_ARG_OPENMP()
+if test x"$pac_cv_openmp" == x"yes" ; then
+   FDEFINES="$psblas_cv_define_prepend-DOPENMP $FDEFINES";
+   CDEFINES="-DOPENMP $CDEFINES";	
+   FCOPT="$FCOPT $pac_cv_openmp_fcopt";
+   CCOPT="$CCOPT $pac_cv_openmp_ccopt";
+   FLINK="$FLINK $pac_cv_openmp_fcopt";
+fi

 PAC_FORTRAN_HAVE_PSBLAS([AC_MSG_RESULT([yes.])],
 	[AC_MSG_ERROR([no.  Could not find working version of PSBLAS.])])
@@ -693,14 +786,10 @@ fi
 PAC_CHECK_SUPERLUDIST()

 if test "x$amg4psblas_cv_have_superludist" == "xyes" ; then
-   pac_sludist_version="$amg4psblas_cv_superludist_major";      
-   if (($amg4psblas_cv_superludist_major==6)); then
-      if (($amg4psblas_cv_superludist_minor>=3)); then
-        pac_sludist_version="63";
-      fi
-   fi	
+   pac_sludist_version="$amg4psblas_cv_superludist_major$amg4psblas_cv_superludist_minor";  
+   AC_MSG_NOTICE([Configuring with SuperLU_DIST version flag $pac_sludist_version])
   SLUDIST_FLAGS=""	       
-   SLUDIST_FLAGS="-DHave_SLUDist_ -DSLUD_VERSION_$pac_sludist_version $SLUDIST_INCLUDES"
+   SLUDIST_FLAGS="-DHave_SLUDist_ -DSLUD_VERSION_="$pac_sludist_version" $SLUDIST_INCLUDES"
   FDEFINES="$amg_cv_define_prepend-DHAVE_SLUDIST_ $FDEFINES"
 else
   SLUDIST_FLAGS=""
@@ -753,6 +753,7 @@ infodir
 docdir
 oldincludedir
 includedir
+runstatedir
 localstatedir
 sharedstatedir
 sysconfdir
@@ -877,6 +878,7 @@ datadir='${datarootdir}'
 sysconfdir='${prefix}/etc'
 sharedstatedir='${prefix}/com'
 localstatedir='${prefix}/var'
+runstatedir='${localstatedir}/run'
 includedir='${prefix}/include'
 oldincludedir='/usr/include'
 docdir='${datarootdir}/doc/${PACKAGE_TARNAME}'
@@ -1129,6 +1131,15 @@ do
  | -silent | --silent | --silen | --sile | --sil)
    silent=yes ;;

+  -runstatedir | --runstatedir | --runstatedi | --runstated \
+  | --runstate | --runstat | --runsta | --runst | --runs \
+  | --run | --ru | --r)
+    ac_prev=runstatedir ;;
+  -runstatedir=* | --runstatedir=* | --runstatedi=* | --runstated=* \
+  | --runstate=* | --runstat=* | --runsta=* | --runst=* | --runs=* \
+  | --run=* | --ru=* | --r=*)
+    runstatedir=$ac_optarg ;;
+
  -sbindir | --sbindir | --sbindi | --sbind | --sbin | --sbi | --sb)
    ac_prev=sbindir ;;
  -sbindir=* | --sbindir=* | --sbindi=* | --sbind=* | --sbin=* \
@@ -1266,7 +1277,7 @@ fi
 for ac_var in	exec_prefix prefix bindir sbindir libexecdir datarootdir \
 		datadir sysconfdir sharedstatedir localstatedir includedir \
 		oldincludedir docdir infodir htmldir dvidir pdfdir psdir \
-		libdir localedir mandir
+		libdir localedir mandir runstatedir
 do
  eval ac_val=\$$ac_var
  # Remove trailing slashes.
@@ -1419,6 +1430,7 @@ Fine tuning of the installation directories:
  --sysconfdir=DIR        read-only single-machine data [PREFIX/etc]
  --sharedstatedir=DIR    modifiable architecture-independent data [PREFIX/com]
  --localstatedir=DIR     modifiable single-machine data [PREFIX/var]
+  --runstatedir=DIR       modifiable per-process data [LOCALSTATEDIR/run]
  --libdir=DIR            object code libraries [EPREFIX/lib]
  --includedir=DIR        C header files [PREFIX/include]
  --oldincludedir=DIR     C header files for non-gcc [/usr/include]
@@ -8087,6 +8099,44 @@ ac_compiler_gnu=$ac_cv_c_compiler_gnu



+{ $as_echo "$as_me:${as_lineno-$LINENO}: checking support for Fortran ISO_C_BINDING module" >&5
+$as_echo_n "checking support for Fortran ISO_C_BINDING module... " >&6; }
+ ac_ext=${ac_fc_srcext-f}
+ac_compile='$FC -c $FCFLAGS $ac_fcflags_srcext conftest.$ac_ext >&5'
+ac_link='$FC -o conftest$ac_exeext $FCFLAGS $LDFLAGS $ac_fcflags_srcext conftest.$ac_ext $LIBS >&5'
+ac_compiler_gnu=$ac_cv_fc_compiler_gnu
+
+ ac_exeext=''
+ ac_ext='f90'
+  ac_fc=${MPIFC-$FC};
+ cat > conftest.$ac_ext <<_ACEOF
+
+program conftest
+  use iso_c_binding
+end program conftest
+_ACEOF
+if ac_fn_fc_try_compile "$LINENO"; then :
+    { $as_echo "$as_me:${as_lineno-$LINENO}: result: yes" >&5
+$as_echo "yes" >&6; }
+		     :
+else
+    { $as_echo "$as_me:${as_lineno-$LINENO}: result: no" >&5
+$as_echo "no" >&6; }
+		     echo "configure: failed program was:" >&5
+		     cat conftest.$ac_ext >&5
+		      as_fn_error $? "Sorry, cannot build PSBLAS without support for ISO_C_BINDING.
+ Please get a Fortran compiler that supports it, e.g. GNU Fortran 4.8." "$LINENO" 5
+
+fi
+rm -f core conftest.err conftest.$ac_objext conftest.$ac_ext
+ac_ext=c
+ac_cpp='$CPP $CPPFLAGS'
+ac_compile='$CC -c $CFLAGS $CPPFLAGS conftest.$ac_ext >&5'
+ac_link='$CC -o conftest$ac_exeext $CFLAGS $CPPFLAGS $LDFLAGS conftest.$ac_ext $LIBS >&5'
+ac_compiler_gnu=$ac_cv_c_compiler_gnu
+
+
+
 { $as_echo "$as_me:${as_lineno-$LINENO}: checking support for ISO_FORTRAN_ENV" >&5
 $as_echo_n "checking support for ISO_FORTRAN_ENV... " >&6; }
 ac_ext=${ac_fc_srcext-f}
@@ -10706,6 +10756,8 @@ rm -f core conftest.err conftest.$ac_objext \
  if test "x$pac_sludist_lib_ok" == "xno" ; then
          SLUDIST_LIBS="$amg4psblas_cv_superludist  -L$amg4psblas_cv_superludistdir/lib";
     LIBS="$SLUDIST_LIBS -lm $save_LIBS";
+     { $as_echo "$as_me:${as_lineno-$LINENO}: checking for superlu_malloc_dist in $SLUDIST_LIBS" >&5
+$as_echo_n "checking for superlu_malloc_dist in $SLUDIST_LIBS... " >&6; }
     cat confdefs.h - <<_ACEOF >conftest.$ac_ext
 /* end confdefs.h.  */

@@ -10736,6 +10788,8 @@ rm -f core conftest.err conftest.$ac_objext \
  if test "x$pac_sludist_lib_ok" == "xno" ; then
          SLUDIST_LIBS="$amg4psblas_cv_superludist  -L$amg4psblas_cv_superludistdir/lib64";
     LIBS="$SLUDIST_LIBS -lm $save_LIBS";
+     { $as_echo "$as_me:${as_lineno-$LINENO}: checking for superlu_malloc_dist in $SLUDIST_LIBS" >&5
+$as_echo_n "checking for superlu_malloc_dist in $SLUDIST_LIBS... " >&6; }
     cat confdefs.h - <<_ACEOF >conftest.$ac_ext
 /* end confdefs.h.  */

@@ -10763,8 +10817,8 @@ fi
 rm -f core conftest.err conftest.$ac_objext \
    conftest$ac_exeext conftest.$ac_ext
  fi
-  { $as_echo "$as_me:${as_lineno-$LINENO}: result: $pac_sludist_lib_ok" >&5
-$as_echo "$pac_sludist_lib_ok" >&6; }
+  { $as_echo "$as_me:${as_lineno-$LINENO}: result: $pac_sludist_lib_ok $SLUDIST_LIBS" >&5
+$as_echo "$pac_sludist_lib_ok $SLUDIST_LIBS" >&6; }
 fi

 if test "x$pac_sludist_lib_ok" == "xyes" ; then
@@ -10906,14 +10960,11 @@ ac_compiler_gnu=$ac_cv_c_compiler_gnu


 if test "x$amg4psblas_cv_have_superludist" == "xyes" ; then
-   pac_sludist_version="$amg4psblas_cv_superludist_major";
-   if (($amg4psblas_cv_superludist_major==6)); then
-      if (($amg4psblas_cv_superludist_minor>=3)); then
-        pac_sludist_version="63";
-      fi
-   fi
+   pac_sludist_version="$amg4psblas_cv_superludist_major$amg4psblas_cv_superludist_minor";
+   { $as_echo "$as_me:${as_lineno-$LINENO}: Configuring with SuperLU_DIST version flag $pac_sludist_version" >&5
+$as_echo "$as_me: Configuring with SuperLU_DIST version flag $pac_sludist_version" >&6;}
   SLUDIST_FLAGS=""
-   SLUDIST_FLAGS="-DHave_SLUDist_ -DSLUD_VERSION_$pac_sludist_version $SLUDIST_INCLUDES"
+   SLUDIST_FLAGS="-DHave_SLUDist_ -DSLUD_VERSION_="$pac_sludist_version" $SLUDIST_INCLUDES"
   FDEFINES="$amg_cv_define_prepend-DHAVE_SLUDIST_ $FDEFINES"
 else
   SLUDIST_FLAGS=""
@@ -12275,7 +12326,9 @@ $as_echo X/"$am_mf" |
    { { $as_echo "$as_me:${as_lineno-$LINENO}: error: in \`$ac_pwd':" >&5
 $as_echo "$as_me: error: in \`$ac_pwd':" >&2;}
 as_fn_error $? "Something went wrong bootstrapping makefile fragments
-    for automatic dependency tracking.  Try re-running configure with the
+    for automatic dependency tracking.  If GNU make was not used, consider
+    re-running the configure script with MAKE=\"gmake\" (or whatever is
+    necessary).  You can also try re-running configure with the
    '--disable-dependency-tracking' option to at least be able to build
    the package (albeit without support for automatic dependency tracking).
 See \`config.log' for more details" "$LINENO" 5; }
@@ -673,6 +673,12 @@ PAC_FORTRAN_TEST_VOLATILE(
        [AC_MSG_ERROR([Sorry, cannot build PSBLAS without support for	VOLATILE])]	
 )

+PAC_FORTRAN_TEST_ISO_C_BIND( 
+	[],
+        [AC_MSG_ERROR([Sorry, cannot build PSBLAS without support for ISO_C_BINDING. 
+ Please get a Fortran compiler that supports it, e.g. GNU Fortran 4.8.])]
+)
+
 PAC_FORTRAN_TEST_ISO_FORTRAN_ENV(
 	[],
        [AC_MSG_ERROR([Sorry, cannot build PSBLAS without support for	ISO_FORTRAN_ENV])]	
@@ -856,14 +862,10 @@ fi
 PAC_CHECK_SUPERLUDIST()

 if test "x$amg4psblas_cv_have_superludist" == "xyes" ; then
-   pac_sludist_version="$amg4psblas_cv_superludist_major";      
-   if (($amg4psblas_cv_superludist_major==6)); then
-      if (($amg4psblas_cv_superludist_minor>=3)); then
-        pac_sludist_version="63";
-      fi
-   fi	
+   pac_sludist_version="$amg4psblas_cv_superludist_major$amg4psblas_cv_superludist_minor";  
+   AC_MSG_NOTICE([Configuring with SuperLU_DIST version flag $pac_sludist_version])
   SLUDIST_FLAGS=""	       
-   SLUDIST_FLAGS="-DHave_SLUDist_ -DSLUD_VERSION_$pac_sludist_version $SLUDIST_INCLUDES"
+   SLUDIST_FLAGS="-DHave_SLUDist_ -DSLUD_VERSION_="$pac_sludist_version" $SLUDIST_INCLUDES"
   FDEFINES="$amg_cv_define_prepend-DHAVE_SLUDIST_ $FDEFINES"
 else
   SLUDIST_FLAGS=""
@@ -10,7 +10,7 @@
 <link rel="stylesheet" type="text/css" href="userhtml.css"> 
 </head><body 
 >
-   <!--l. 678--><div class="crosslinks"><p class="noindent"><span 
+   <!--l. 685--><div class="crosslinks"><p class="noindent"><span 
 class="cmr-12">[</span><a 
 href="userhtmlsu11.html" ><span 
 class="cmr-12">next</span></a><span 
@@ -34,8 +34,8 @@ class="cmr-12">Method hierarchy</span><span
 class="cmr-12">_build</span></h4>
 <div class="center" 
 >
-<!--l. 680--><p class="noindent" >
-<!--l. 681--><p class="noindent" ><span class="lstinline"></span><span 
+<!--l. 687--><p class="noindent" >
+<!--l. 688--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">call</span><span 
 class="cmtt-12">&#x00A0;</span><span 
 class="cmtt-12">p</span><span 
@@ -49,7 +49,7 @@ class="cmtt-12">,</span><span
 class="cmtt-12">info</span><span 
 class="cmtt-12">)</span><br />
 </div>
-<!--l. 684--><p class="noindent" ><span 
+<!--l. 691--><p class="noindent" ><span 
 class="cmr-12">This method builds the hierarchy of matrices and restriction/prolongation operators for</span>
 <span 
 class="cmr-12">the multilevel preconditioner </span><span class="lstinline"></span><span 
@@ -61,7 +61,7 @@ class="cmtt-12">init</span> <span
 class="cmr-12">and </span><span class="lstinline"></span><span 
 class="cmtt-12">set</span><span 
 class="cmr-12">.</span>
-<!--l. 689--><p class="noindent" ><span 
+<!--l. 696--><p class="noindent" ><span 
 class="cmbx-12">Arguments</span>
   <div class="tabular"> <table id="TBL-11" class="tabular" 
 cellspacing="0" cellpadding="0"  
@@ -69,9 +69,9 @@ cellspacing="0" cellpadding="0"
 id="TBL-11-1"><col 
 id="TBL-11-2"></colgroup><tr  
 style="vertical-align:baseline;" id="TBL-11-1-"><td  style="white-space:normal; text-align:left;" id="TBL-11-1-1"  
-class="td11"><!--l. 692--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 699--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">a</span>     </td><td  style="white-space:normal; text-align:left;" id="TBL-11-1-2"  
-class="td11"><!--l. 692--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 699--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">type</span><span 
 class="cmtt-12">(</span><span 
 class="cmtt-12">psb_</span><span 
@@ -86,8 +86,8 @@ class="cmtt-12">)</span><span
 class="cmr-12">.</span>                                  </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-11-2-"><td  style="white-space:normal; text-align:left;" id="TBL-11-2-1"  
-class="td11"><!--l. 693--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-11-2-2"  
-class="td11"><!--l. 693--><p class="noindent" ><span 
+class="td11"><!--l. 700--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-11-2-2"  
+class="td11"><!--l. 700--><p class="noindent" ><span 
 class="cmr-12">The sparse matrix structure containing the local part of the matrix</span>
 <span 
 class="cmr-12">to be preconditioned. Note that </span><span 
@@ -107,9 +107,9 @@ class="cmr-12">]</span></span><span
 class="cmr-12">.</span>            </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-11-3-"><td  style="white-space:normal; text-align:left;" id="TBL-11-3-1"  
-class="td11"><!--l. 698--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 705--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">desc_a</span></td><td  style="white-space:normal; text-align:left;" id="TBL-11-3-2"  
-class="td11"><!--l. 698--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 705--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">type</span><span 
 class="cmtt-12">(</span><span 
 class="cmtt-12">psb_desc_type</span><span 
@@ -122,8 +122,8 @@ class="cmtt-12">)</span><span
 class="cmr-12">.</span>                                     </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-11-4-"><td  style="white-space:normal; text-align:left;" id="TBL-11-4-1"  
-class="td11"><!--l. 699--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-11-4-2"  
-class="td11"><!--l. 699--><p class="noindent" ><span 
+class="td11"><!--l. 706--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-11-4-2"  
+class="td11"><!--l. 706--><p class="noindent" ><span 
 class="cmr-12">The communication descriptor of </span><span class="lstinline"></span><span 
 class="cmtt-12">a</span><span 
 class="cmr-12">. See the PSBLAS User&#8217;s Guide</span>
@@ -136,9 +136,9 @@ class="cmr-12">]</span></span><span
 class="cmr-12">.</span>                                                                    </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-11-5-"><td  style="white-space:normal; text-align:left;" id="TBL-11-5-1"  
-class="td11"><!--l. 701--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 708--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">info</span>  </td><td  style="white-space:normal; text-align:left;" id="TBL-11-5-2"  
-class="td11"><!--l. 701--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 708--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">integer</span><span 
 class="cmtt-12">,</span><span 
 class="cmtt-12">&#x00A0;</span><span 
@@ -149,8 +149,8 @@ class="cmtt-12">)</span><span
 class="cmr-12">.</span>                                                       </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-11-6-"><td  style="white-space:normal; text-align:left;" id="TBL-11-6-1"  
-class="td11"><!--l. 702--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-11-6-2"  
-class="td11"><!--l. 702--><p class="noindent" ><span 
+class="td11"><!--l. 709--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-11-6-2"  
+class="td11"><!--l. 709--><p class="noindent" ><span 
 class="cmr-12">Error code. If no error, 0 is returned. See Section</span><span 
 class="cmr-12">&#x00A0;</span><a 
 href="userhtmlse7.html#x28-320007"><span 
@@ -165,7 +165,7 @@ class="td11">      </td> </tr></table></div>
                                                                               

                                                                               
-   <!--l. 708--><div class="crosslinks"><p class="noindent"><span 
+   <!--l. 715--><div class="crosslinks"><p class="noindent"><span 
 class="cmr-12">[</span><a 
 href="userhtmlsu11.html" ><span 
 class="cmr-12">next</span></a><span 
@@ -182,6 +182,6 @@ class="cmr-12">] [</span><a
 href="userhtmlse5.html#userhtmlsu10.html" ><span 
 class="cmr-12">up</span></a><span 
 class="cmr-12">] </span></p></div>
-<!--l. 708--><p class="indent" >   <a 
+<!--l. 715--><p class="indent" >   <a 
 id="tailuserhtmlsu10.html"></a>  
 </body></html> 
@@ -10,7 +10,7 @@
 <link rel="stylesheet" type="text/css" href="userhtml.css"> 
 </head><body 
 >
-   <!--l. 708--><div class="crosslinks"><p class="noindent"><span 
+   <!--l. 715--><div class="crosslinks"><p class="noindent"><span 
 class="cmr-12">[</span><a 
 href="userhtmlsu12.html" ><span 
 class="cmr-12">next</span></a><span 
@@ -34,8 +34,8 @@ class="cmr-12">Method smoothers</span><span
 class="cmr-12">_build</span></h4>
 <div class="center" 
 >
-<!--l. 711--><p class="noindent" >
-<!--l. 712--><p class="noindent" ><span class="lstinline"></span><span 
+<!--l. 718--><p class="noindent" >
+<!--l. 719--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">call</span><span 
 class="cmtt-12">&#x00A0;</span><span 
 class="cmtt-12">p</span><span 
@@ -57,7 +57,7 @@ class="cmtt-12">,</span><span
 class="cmtt-12">imold</span><span 
 class="cmtt-12">])</span><br />
 </div>
-<!--l. 715--><p class="noindent" ><span 
+<!--l. 722--><p class="noindent" ><span 
 class="cmr-12">This method builds the smoothers and the coarsest-level solvers for the multilevel</span>
 <span 
 class="cmr-12">preconditioner </span><span class="lstinline"></span><span 
@@ -75,7 +75,7 @@ class="cmr-12">&#x00A0;</span><a
 href="userhtmlsu10.html#x20-190005.3"><span 
 class="cmr-12">5.3</span><!--tex4ht:ref: sec:hier_bld --></a><span 
 class="cmr-12">).</span>
-<!--l. 722--><p class="noindent" ><span 
+<!--l. 729--><p class="noindent" ><span 
 class="cmbx-12">Arguments</span>
   <div class="tabular"> <table id="TBL-12" class="tabular" 
 cellspacing="0" cellpadding="0"  
@@ -83,9 +83,9 @@ cellspacing="0" cellpadding="0"
 id="TBL-12-1"><col 
 id="TBL-12-2"></colgroup><tr  
 style="vertical-align:baseline;" id="TBL-12-1-"><td  style="white-space:normal; text-align:left;" id="TBL-12-1-1"  
-class="td11"><!--l. 725--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 732--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">a</span>     </td><td  style="white-space:normal; text-align:left;" id="TBL-12-1-2"  
-class="td11"><!--l. 725--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 732--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">type</span><span 
 class="cmtt-12">(</span><span 
 class="cmtt-12">psb_</span><span 
@@ -100,8 +100,8 @@ class="cmtt-12">)</span><span
 class="cmr-12">.</span>                                  </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-12-2-"><td  style="white-space:normal; text-align:left;" id="TBL-12-2-1"  
-class="td11"><!--l. 726--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-12-2-2"  
-class="td11"><!--l. 726--><p class="noindent" ><span 
+class="td11"><!--l. 733--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-12-2-2"  
+class="td11"><!--l. 733--><p class="noindent" ><span 
 class="cmr-12">The sparse matrix structure containing the local part of the matrix</span>
 <span 
 class="cmr-12">to be preconditioned. Note that </span><span 
@@ -121,9 +121,9 @@ class="cmr-12">]</span></span><span
 class="cmr-12">.</span>            </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-12-3-"><td  style="white-space:normal; text-align:left;" id="TBL-12-3-1"  
-class="td11"><!--l. 730--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 737--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">desc_a</span></td><td  style="white-space:normal; text-align:left;" id="TBL-12-3-2"  
-class="td11"><!--l. 730--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 737--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">type</span><span 
 class="cmtt-12">(</span><span 
 class="cmtt-12">psb_desc_type</span><span 
@@ -136,8 +136,8 @@ class="cmtt-12">)</span><span
 class="cmr-12">.</span>                                     </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-12-4-"><td  style="white-space:normal; text-align:left;" id="TBL-12-4-1"  
-class="td11"><!--l. 731--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-12-4-2"  
-class="td11"><!--l. 731--><p class="noindent" ><span 
+class="td11"><!--l. 738--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-12-4-2"  
+class="td11"><!--l. 738--><p class="noindent" ><span 
 class="cmr-12">The communication descriptor of </span><span class="lstinline"></span><span 
 class="cmtt-12">a</span><span 
 class="cmr-12">. See the PSBLAS User&#8217;s Guide</span>
@@ -150,9 +150,9 @@ class="cmr-12">]</span></span><span
 class="cmr-12">.</span>                                                                    </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-12-5-"><td  style="white-space:normal; text-align:left;" id="TBL-12-5-1"  
-class="td11"><!--l. 733--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 740--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">info</span>  </td><td  style="white-space:normal; text-align:left;" id="TBL-12-5-2"  
-class="td11"><!--l. 733--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 740--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">integer</span><span 
 class="cmtt-12">,</span><span 
 class="cmtt-12">&#x00A0;</span><span 
@@ -163,8 +163,8 @@ class="cmtt-12">)</span><span
 class="cmr-12">.</span>                                                       </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-12-6-"><td  style="white-space:normal; text-align:left;" id="TBL-12-6-1"  
-class="td11"><!--l. 734--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-12-6-2"  
-class="td11"><!--l. 734--><p class="noindent" ><span 
+class="td11"><!--l. 741--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-12-6-2"  
+class="td11"><!--l. 741--><p class="noindent" ><span 
 class="cmr-12">Error code. If no error, 0 is returned. See Section</span><span 
 class="cmr-12">&#x00A0;</span><a 
 href="userhtmlse7.html#x28-320007"><span 
@@ -172,9 +172,9 @@ class="cmr-12">7</span><!--tex4ht:ref: sec:errors --></a> <span
 class="cmr-12">for details.</span>     </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-12-7-"><td  style="white-space:normal; text-align:left;" id="TBL-12-7-1"  
-class="td11"><!--l. 735--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 742--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">amold</span> </td><td  style="white-space:normal; text-align:left;" id="TBL-12-7-2"  
-class="td11"><!--l. 735--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 742--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">class</span><span 
 class="cmtt-12">(</span><span 
 class="cmtt-12">psb_</span><span 
@@ -191,8 +191,8 @@ class="cmtt-12">optional</span><span
 class="cmr-12">.</span>       </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-12-8-"><td  style="white-space:normal; text-align:left;" id="TBL-12-8-1"  
-class="td11"><!--l. 736--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-12-8-2"  
-class="td11"><!--l. 736--><p class="noindent" ><span 
+class="td11"><!--l. 743--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-12-8-2"  
+class="td11"><!--l. 743--><p class="noindent" ><span 
 class="cmr-12">The  desired  dynamic  type  for  internal  matrix  components;  this</span>
 <span 
 class="cmr-12">allows  e.g.  running  on  GPUs;  it  needs  not  be  the  same  on  all</span>
@@ -205,9 +205,9 @@ class="cmr-12">]</span></span><span
 class="cmr-12">.</span>             </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-12-9-"><td  style="white-space:normal; text-align:left;" id="TBL-12-9-1"  
-class="td11"><!--l. 740--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 747--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">vmold</span> </td><td  style="white-space:normal; text-align:left;" id="TBL-12-9-2"  
-class="td11"><!--l. 740--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 747--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">class</span><span 
 class="cmtt-12">(</span><span 
 class="cmtt-12">psb_</span><span 
@@ -224,16 +224,16 @@ class="cmtt-12">optional</span><span
 class="cmr-12">.</span>         </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-12-10-"><td  style="white-space:normal; text-align:left;" id="TBL-12-10-1"  
-class="td11"><!--l. 741--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-12-10-2"  
-class="td11"><!--l. 741--><p class="noindent" ><span 
+class="td11"><!--l. 748--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-12-10-2"  
+class="td11"><!--l. 748--><p class="noindent" ><span 
 class="cmr-12">The  desired  dynamic  type  for  internal  vector  components;  this</span>
 <span 
 class="cmr-12">allows e.g. running on GPUs.</span>                                                 </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-12-11-"><td  style="white-space:normal; text-align:left;" id="TBL-12-11-1"  
-class="td11"><!--l. 743--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 750--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">imold</span> </td><td  style="white-space:normal; text-align:left;" id="TBL-12-11-2"  
-class="td11"><!--l. 743--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 750--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">class</span><span 
 class="cmtt-12">(</span><span 
 class="cmtt-12">psb_i_base_vect_type</span><span 
@@ -248,8 +248,8 @@ class="cmtt-12">optional</span><span
 class="cmr-12">.</span>         </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-12-12-"><td  style="white-space:normal; text-align:left;" id="TBL-12-12-1"  
-class="td11"><!--l. 744--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-12-12-2"  
-class="td11"><!--l. 744--><p class="noindent" ><span 
+class="td11"><!--l. 751--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-12-12-2"  
+class="td11"><!--l. 751--><p class="noindent" ><span 
 class="cmr-12">The desired dynamic type for internal integer vector components;</span>
 <span 
 class="cmr-12">this allows e.g. running on GPUs.</span>                                           </td>
@@ -262,7 +262,7 @@ class="td11">      </td> </tr></table></div>
                                                                               

                                                                               
-   <!--l. 750--><div class="crosslinks"><p class="noindent"><span 
+   <!--l. 757--><div class="crosslinks"><p class="noindent"><span 
 class="cmr-12">[</span><a 
 href="userhtmlsu12.html" ><span 
 class="cmr-12">next</span></a><span 
@@ -279,6 +279,6 @@ class="cmr-12">] [</span><a
 href="userhtmlse5.html#userhtmlsu11.html" ><span 
 class="cmr-12">up</span></a><span 
 class="cmr-12">] </span></p></div>
-<!--l. 750--><p class="indent" >   <a 
+<!--l. 757--><p class="indent" >   <a 
 id="tailuserhtmlsu11.html"></a>  
 </body></html> 
@@ -10,7 +10,7 @@
 <link rel="stylesheet" type="text/css" href="userhtml.css"> 
 </head><body 
 >
-   <!--l. 750--><div class="crosslinks"><p class="noindent"><span 
+   <!--l. 757--><div class="crosslinks"><p class="noindent"><span 
 class="cmr-12">[</span><a 
 href="userhtmlsu13.html" ><span 
 class="cmr-12">next</span></a><span 
@@ -33,8 +33,8 @@ class="cmr-12">5.5   </span></span> <a
 class="cmr-12">Method build</span></h4>
 <div class="center" 
 >
-<!--l. 752--><p class="noindent" >
-<!--l. 753--><p class="noindent" ><span class="lstinline"></span><span 
+<!--l. 759--><p class="noindent" >
+<!--l. 760--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">call</span><span 
 class="cmtt-12">&#x00A0;</span><span 
 class="cmtt-12">p</span><span 
@@ -54,7 +54,7 @@ class="cmtt-12">,</span><span
 class="cmtt-12">imold</span><span 
 class="cmtt-12">])</span><br />
 </div>
-<!--l. 756--><p class="noindent" ><span 
+<!--l. 763--><p class="noindent" ><span 
 class="cmr-12">This method builds the preconditioner </span><span class="lstinline"></span><span 
 class="cmtt-12">p</span> <span 
 class="cmr-12">according to the requirements made by the</span>
@@ -83,7 +83,7 @@ class="cmtt-12">smoothers_build</span><span
 class="cmr-12">, whose nomenclature would however be somewhat unnatural when</span>
 <span 
 class="cmr-12">dealing with simple one-level preconditioners.</span>
-<!--l. 766--><p class="noindent" ><span 
+<!--l. 773--><p class="noindent" ><span 
 class="cmbx-12">Arguments</span>
   <div class="tabular"> <table id="TBL-13" class="tabular" 
 cellspacing="0" cellpadding="0"  
@@ -91,9 +91,9 @@ cellspacing="0" cellpadding="0"
 id="TBL-13-1"><col 
 id="TBL-13-2"></colgroup><tr  
 style="vertical-align:baseline;" id="TBL-13-1-"><td  style="white-space:normal; text-align:left;" id="TBL-13-1-1"  
-class="td11"><!--l. 769--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 776--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">a</span>     </td><td  style="white-space:normal; text-align:left;" id="TBL-13-1-2"  
-class="td11"><!--l. 769--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 776--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">type</span><span 
 class="cmtt-12">(</span><span 
 class="cmtt-12">psb_</span><span 
@@ -108,8 +108,8 @@ class="cmtt-12">)</span><span
 class="cmr-12">.</span>                                  </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-13-2-"><td  style="white-space:normal; text-align:left;" id="TBL-13-2-1"  
-class="td11"><!--l. 770--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-13-2-2"  
-class="td11"><!--l. 770--><p class="noindent" ><span 
+class="td11"><!--l. 777--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-13-2-2"  
+class="td11"><!--l. 777--><p class="noindent" ><span 
 class="cmr-12">The sparse matrix structure containing the local part of the matrix</span>
 <span 
 class="cmr-12">to be preconditioned. Note that </span><span 
@@ -129,9 +129,9 @@ class="cmr-12">]</span></span><span
 class="cmr-12">.</span>            </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-13-3-"><td  style="white-space:normal; text-align:left;" id="TBL-13-3-1"  
-class="td11"><!--l. 774--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 781--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">desc_a</span></td><td  style="white-space:normal; text-align:left;" id="TBL-13-3-2"  
-class="td11"><!--l. 774--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 781--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">type</span><span 
 class="cmtt-12">(</span><span 
 class="cmtt-12">psb_desc_type</span><span 
@@ -144,8 +144,8 @@ class="cmtt-12">)</span><span
 class="cmr-12">.</span>                                     </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-13-4-"><td  style="white-space:normal; text-align:left;" id="TBL-13-4-1"  
-class="td11"><!--l. 775--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-13-4-2"  
-class="td11"><!--l. 775--><p class="noindent" ><span 
+class="td11"><!--l. 782--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-13-4-2"  
+class="td11"><!--l. 782--><p class="noindent" ><span 
 class="cmr-12">The communication descriptor of </span><span class="lstinline"></span><span 
 class="cmtt-12">a</span><span 
 class="cmr-12">. See the PSBLAS User&#8217;s Guide</span>
@@ -158,9 +158,9 @@ class="cmr-12">]</span></span><span
 class="cmr-12">.</span>                                                                    </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-13-5-"><td  style="white-space:normal; text-align:left;" id="TBL-13-5-1"  
-class="td11"><!--l. 777--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 784--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">info</span>  </td><td  style="white-space:normal; text-align:left;" id="TBL-13-5-2"  
-class="td11"><!--l. 777--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 784--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">integer</span><span 
 class="cmtt-12">,</span><span 
 class="cmtt-12">&#x00A0;</span><span 
@@ -171,8 +171,8 @@ class="cmtt-12">)</span><span
 class="cmr-12">.</span>                                                       </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-13-6-"><td  style="white-space:normal; text-align:left;" id="TBL-13-6-1"  
-class="td11"><!--l. 778--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-13-6-2"  
-class="td11"><!--l. 778--><p class="noindent" ><span 
+class="td11"><!--l. 785--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-13-6-2"  
+class="td11"><!--l. 785--><p class="noindent" ><span 
 class="cmr-12">Error code. If no error, 0 is returned. See Section</span><span 
 class="cmr-12">&#x00A0;</span><a 
 href="userhtmlse7.html#x28-320007"><span 
@@ -180,9 +180,9 @@ class="cmr-12">7</span><!--tex4ht:ref: sec:errors --></a> <span
 class="cmr-12">for details.</span>     </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-13-7-"><td  style="white-space:normal; text-align:left;" id="TBL-13-7-1"  
-class="td11"><!--l. 779--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 786--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">amold</span> </td><td  style="white-space:normal; text-align:left;" id="TBL-13-7-2"  
-class="td11"><!--l. 779--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 786--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">class</span><span 
 class="cmtt-12">(</span><span 
 class="cmtt-12">psb_</span><span 
@@ -199,8 +199,8 @@ class="cmtt-12">optional</span><span
 class="cmr-12">.</span>       </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-13-8-"><td  style="white-space:normal; text-align:left;" id="TBL-13-8-1"  
-class="td11"><!--l. 780--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-13-8-2"  
-class="td11"><!--l. 780--><p class="noindent" ><span 
+class="td11"><!--l. 787--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-13-8-2"  
+class="td11"><!--l. 787--><p class="noindent" ><span 
 class="cmr-12">The  desired  dynamic  type  for  internal  matrix  components;  this</span>
 <span 
 class="cmr-12">allows  e.g.  running  on  GPUs;  it  needs  not  be  the  same  on  all</span>
@@ -213,9 +213,9 @@ class="cmr-12">]</span></span><span
 class="cmr-12">.</span>             </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-13-9-"><td  style="white-space:normal; text-align:left;" id="TBL-13-9-1"  
-class="td11"><!--l. 784--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 791--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">vmold</span> </td><td  style="white-space:normal; text-align:left;" id="TBL-13-9-2"  
-class="td11"><!--l. 784--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 791--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">class</span><span 
 class="cmtt-12">(</span><span 
 class="cmtt-12">psb_</span><span 
@@ -232,16 +232,16 @@ class="cmtt-12">optional</span><span
 class="cmr-12">.</span>         </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-13-10-"><td  style="white-space:normal; text-align:left;" id="TBL-13-10-1"  
-class="td11"><!--l. 785--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-13-10-2"  
-class="td11"><!--l. 785--><p class="noindent" ><span 
+class="td11"><!--l. 792--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-13-10-2"  
+class="td11"><!--l. 792--><p class="noindent" ><span 
 class="cmr-12">The  desired  dynamic  type  for  internal  vector  components;  this</span>
 <span 
 class="cmr-12">allows e.g. running on GPUs.</span>                                                 </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-13-11-"><td  style="white-space:normal; text-align:left;" id="TBL-13-11-1"  
-class="td11"><!--l. 787--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 794--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">imold</span> </td><td  style="white-space:normal; text-align:left;" id="TBL-13-11-2"  
-class="td11"><!--l. 787--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 794--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">class</span><span 
 class="cmtt-12">(</span><span 
 class="cmtt-12">psb_i_base_vect_type</span><span 
@@ -256,15 +256,15 @@ class="cmtt-12">optional</span><span
 class="cmr-12">.</span>         </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-13-12-"><td  style="white-space:normal; text-align:left;" id="TBL-13-12-1"  
-class="td11"><!--l. 788--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-13-12-2"  
-class="td11"><!--l. 788--><p class="noindent" ><span 
+class="td11"><!--l. 795--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-13-12-2"  
+class="td11"><!--l. 795--><p class="noindent" ><span 
 class="cmr-12">The desired dynamic type for internal integer vector components;</span>
 <span 
 class="cmr-12">this allows e.g. running on GPUs.</span>                                           </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-13-13-"><td  style="white-space:normal; text-align:left;" id="TBL-13-13-1"  
 class="td11">      </td> </tr></table></div>
-<!--l. 793--><p class="noindent" ><span 
+<!--l. 800--><p class="noindent" ><span 
 class="cmr-12">The method can be used to build multilevel preconditioners too.</span>
                                                                               

@@ -272,7 +272,7 @@ class="cmr-12">The method can be used to build multilevel preconditioners too.</
                                                                               

                                                                               
-   <!--l. 798--><div class="crosslinks"><p class="noindent"><span 
+   <!--l. 805--><div class="crosslinks"><p class="noindent"><span 
 class="cmr-12">[</span><a 
 href="userhtmlsu13.html" ><span 
 class="cmr-12">next</span></a><span 
@@ -289,6 +289,6 @@ class="cmr-12">] [</span><a
 href="userhtmlse5.html#userhtmlsu12.html" ><span 
 class="cmr-12">up</span></a><span 
 class="cmr-12">] </span></p></div>
-<!--l. 798--><p class="indent" >   <a 
+<!--l. 805--><p class="indent" >   <a 
 id="tailuserhtmlsu12.html"></a>  
 </body></html> 
@@ -10,7 +10,7 @@
 <link rel="stylesheet" type="text/css" href="userhtml.css"> 
 </head><body 
 >
-   <!--l. 798--><div class="crosslinks"><p class="noindent"><span 
+   <!--l. 805--><div class="crosslinks"><p class="noindent"><span 
 class="cmr-12">[</span><a 
 href="userhtmlsu14.html" ><span 
 class="cmr-12">next</span></a><span 
@@ -33,8 +33,8 @@ class="cmr-12">5.6   </span></span> <a
 class="cmr-12">Method apply</span></h4>
 <div class="center" 
 >
-<!--l. 800--><p class="noindent" >
-<!--l. 801--><p class="noindent" ><span class="lstinline"></span><span 
+<!--l. 807--><p class="noindent" >
+<!--l. 808--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">call</span><span 
 class="cmtt-12">&#x00A0;</span><span 
 class="cmtt-12">p</span><span 
@@ -55,7 +55,7 @@ class="cmtt-12">,</span><span
 class="cmtt-12">work</span><span 
 class="cmtt-12">])</span><br />
 </div>
-<!--l. 804--><p class="noindent" ><span 
+<!--l. 811--><p class="noindent" ><span 
 class="cmr-12">This method computes </span><span 
 class="cmmi-12">y </span><span 
 class="cmr-12">= </span><span 
@@ -89,7 +89,7 @@ class="cmtt-12">psb_krylov</span> <span
 class="cmr-12">and hence it is</span>
 <span 
 class="cmr-12">completely transparent to the user.</span>
-<!--l. 813--><p class="noindent" ><span 
+<!--l. 820--><p class="noindent" ><span 
 class="cmbx-12">Arguments</span>
   <div class="tabular"> <table id="TBL-14" class="tabular" 
 cellspacing="0" cellpadding="0"  
@@ -97,9 +97,9 @@ cellspacing="0" cellpadding="0"
 id="TBL-14-1"><col 
 id="TBL-14-2"></colgroup><tr  
 style="vertical-align:baseline;" id="TBL-14-1-"><td  style="white-space:normal; text-align:left;" id="TBL-14-1-1"  
-class="td11"><!--l. 816--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 823--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">x</span>     </td><td  style="white-space:normal; text-align:left;" id="TBL-14-1-2"  
-class="td11"><!--l. 816--><p class="noindent" ><span 
+class="td11"><!--l. 823--><p class="noindent" ><span 
 class="cmti-12">type</span><span class="lstinline"></span><span 
 class="cmtt-12">(</span><span 
 class="cmtt-12">kind_parameter</span><span 
@@ -107,8 +107,8 @@ class="cmtt-12">)</span><span
 class="cmr-12">, dimension(:), intent(in)&#8212;.</span>                    </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-14-2-"><td  style="white-space:normal; text-align:left;" id="TBL-14-2-1"  
-class="td11"><!--l. 817--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-14-2-2"  
-class="td11"><!--l. 817--><p class="noindent" ><span 
+class="td11"><!--l. 824--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-14-2-2"  
+class="td11"><!--l. 824--><p class="noindent" ><span 
 class="cmr-12">The local part of the vector </span><span 
 class="cmmi-12">x</span><span 
 class="cmr-12">. Note that </span><span 
@@ -126,9 +126,9 @@ class="cmr-12">,  single/double</span>
 class="cmr-12">precision version of AMG4PSBLAS under use.</span>                          </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-14-3-"><td  style="white-space:normal; text-align:left;" id="TBL-14-3-1"  
-class="td11"><!--l. 820--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 827--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">y</span>     </td><td  style="white-space:normal; text-align:left;" id="TBL-14-3-2"  
-class="td11"><!--l. 820--><p class="noindent" ><span 
+class="td11"><!--l. 827--><p class="noindent" ><span 
 class="cmti-12">type</span><span class="lstinline"></span><span 
 class="cmtt-12">(</span><span 
 class="cmtt-12">kind_parameter</span><span 
@@ -136,8 +136,8 @@ class="cmtt-12">)</span><span
 class="cmr-12">, dimension(:), intent(out)&#8212;.</span>                  </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-14-4-"><td  style="white-space:normal; text-align:left;" id="TBL-14-4-1"  
-class="td11"><!--l. 821--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-14-4-2"  
-class="td11"><!--l. 821--><p class="noindent" ><span 
+class="td11"><!--l. 828--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-14-4-2"  
+class="td11"><!--l. 828--><p class="noindent" ><span 
 class="cmr-12">The local part of the vector </span><span 
 class="cmmi-12">y</span><span 
 class="cmr-12">. Note that </span><span 
@@ -155,9 +155,9 @@ class="cmr-12">,  single/double</span>
 class="cmr-12">precision version of AMG4PSBLAS under use.</span>                          </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-14-5-"><td  style="white-space:normal; text-align:left;" id="TBL-14-5-1"  
-class="td11"><!--l. 824--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 831--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">desc_a</span></td><td  style="white-space:normal; text-align:left;" id="TBL-14-5-2"  
-class="td11"><!--l. 824--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 831--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">type</span><span 
 class="cmtt-12">(</span><span 
 class="cmtt-12">psb_desc_type</span><span 
@@ -170,16 +170,16 @@ class="cmtt-12">)</span><span
 class="cmr-12">.</span>                                     </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-14-6-"><td  style="white-space:normal; text-align:left;" id="TBL-14-6-1"  
-class="td11"><!--l. 825--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-14-6-2"  
-class="td11"><!--l. 825--><p class="noindent" ><span 
+class="td11"><!--l. 832--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-14-6-2"  
+class="td11"><!--l. 832--><p class="noindent" ><span 
 class="cmr-12">The  communication  descriptor  associated  to  the  matrix  to  be</span>
 <span 
 class="cmr-12">preconditioned.</span>                                                                   </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-14-7-"><td  style="white-space:normal; text-align:left;" id="TBL-14-7-1"  
-class="td11"><!--l. 827--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 834--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">info</span>  </td><td  style="white-space:normal; text-align:left;" id="TBL-14-7-2"  
-class="td11"><!--l. 827--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 834--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">integer</span><span 
 class="cmtt-12">,</span><span 
 class="cmtt-12">&#x00A0;</span><span 
@@ -190,8 +190,8 @@ class="cmtt-12">)</span><span
 class="cmr-12">.</span>                                                       </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-14-8-"><td  style="white-space:normal; text-align:left;" id="TBL-14-8-1"  
-class="td11"><!--l. 828--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-14-8-2"  
-class="td11"><!--l. 828--><p class="noindent" ><span 
+class="td11"><!--l. 835--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-14-8-2"  
+class="td11"><!--l. 835--><p class="noindent" ><span 
 class="cmr-12">Error code. If no error, 0 is returned. See Section</span><span 
 class="cmr-12">&#x00A0;</span><a 
 href="userhtmlse7.html#x28-320007"><span 
@@ -199,9 +199,9 @@ class="cmr-12">7</span><!--tex4ht:ref: sec:errors --></a> <span
 class="cmr-12">for details.</span>     </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-14-9-"><td  style="white-space:normal; text-align:left;" id="TBL-14-9-1"  
-class="td11"><!--l. 829--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 836--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">trans</span> </td><td  style="white-space:normal; text-align:left;" id="TBL-14-9-2"  
-class="td11"><!--l. 829--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 836--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">character</span><span 
 class="cmtt-12">(</span><span 
 class="cmtt-12">len</span><span 
@@ -216,8 +216,8 @@ class="cmtt-12">in</span><span
 class="cmtt-12">).</span>                </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-14-10-"><td  style="white-space:normal; text-align:left;" id="TBL-14-10-1"  
-class="td11"><!--l. 830--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-14-10-2"  
-class="td11"><!--l. 830--><p class="noindent" ><span 
+class="td11"><!--l. 837--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-14-10-2"  
+class="td11"><!--l. 837--><p class="noindent" ><span 
 class="cmr-12">If </span><span class="lstinline"></span><span 
 class="cmtt-12">trans</span> <span 
 class="cmr-12">= </span><span class="lstinline"></span><span 
@@ -291,9 +291,9 @@ class="cmr-8">1</span></sup><span
 class="cmr-12">).</span>                           </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-14-11-"><td  style="white-space:normal; text-align:left;" id="TBL-14-11-1"  
-class="td11"><!--l. 834--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 841--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">work</span>  </td><td  style="white-space:normal; text-align:left;" id="TBL-14-11-2"  
-class="td11"><!--l. 834--><p class="noindent" ><span 
+class="td11"><!--l. 841--><p class="noindent" ><span 
 class="cmti-12">type</span><span class="lstinline"></span><span 
 class="cmtt-12">(</span><span 
 class="cmtt-12">kind_parameter</span><span 
@@ -301,8 +301,8 @@ class="cmtt-12">)</span><span
 class="cmr-12">, dimension(:), optional, target&#8212;.</span>            </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-14-12-"><td  style="white-space:normal; text-align:left;" id="TBL-14-12-1"  
-class="td11"><!--l. 835--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-14-12-2"  
-class="td11"><!--l. 835--><p class="noindent" ><span 
+class="td11"><!--l. 842--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-14-12-2"  
+class="td11"><!--l. 842--><p class="noindent" ><span 
 class="cmr-12">Workspace.  Its  size  should  be  at  least  </span><span class="lstinline"></span><span 
 class="cmtt-12">4</span><span 
 class="cmtt-12">&#x00A0;</span><span 
@@ -339,7 +339,7 @@ class="td11">      </td> </tr></table></div>
                                                                               

                                                                               
-   <!--l. 844--><div class="crosslinks"><p class="noindent"><span 
+   <!--l. 851--><div class="crosslinks"><p class="noindent"><span 
 class="cmr-12">[</span><a 
 href="userhtmlsu14.html" ><span 
 class="cmr-12">next</span></a><span 
@@ -356,6 +356,6 @@ class="cmr-12">] [</span><a
 href="userhtmlse5.html#userhtmlsu13.html" ><span 
 class="cmr-12">up</span></a><span 
 class="cmr-12">] </span></p></div>
-<!--l. 844--><p class="indent" >   <a 
+<!--l. 851--><p class="indent" >   <a 
 id="tailuserhtmlsu13.html"></a>  
 </body></html> 
@@ -10,7 +10,7 @@
 <link rel="stylesheet" type="text/css" href="userhtml.css"> 
 </head><body 
 >
-   <!--l. 844--><div class="crosslinks"><p class="noindent"><span 
+   <!--l. 851--><div class="crosslinks"><p class="noindent"><span 
 class="cmr-12">[</span><a 
 href="userhtmlsu15.html" ><span 
 class="cmr-12">next</span></a><span 
@@ -33,8 +33,8 @@ class="cmr-12">5.7   </span></span> <a
 class="cmr-12">Method free</span></h4>
 <div class="center" 
 >
-<!--l. 846--><p class="noindent" >
-<!--l. 847--><p class="noindent" ><span class="lstinline"></span><span 
+<!--l. 853--><p class="noindent" >
+<!--l. 854--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">call</span><span 
 class="cmtt-12">&#x00A0;</span><span 
 class="cmtt-12">p</span><span 
@@ -46,11 +46,11 @@ class="cmtt-12">,</span><span
 class="cmtt-12">info</span><span 
 class="cmtt-12">)</span><br />
 </div>
-<!--l. 850--><p class="noindent" ><span 
+<!--l. 857--><p class="noindent" ><span 
 class="cmr-12">This method deallocates the preconditioner data structure </span><span class="lstinline"></span><span 
 class="cmtt-12">p</span><span 
 class="cmr-12">.</span>
-<!--l. 853--><p class="noindent" ><span 
+<!--l. 860--><p class="noindent" ><span 
 class="cmbx-12">Arguments</span>
   <div class="tabular"> <table id="TBL-15" class="tabular" 
 cellspacing="0" cellpadding="0"  
@@ -58,9 +58,9 @@ cellspacing="0" cellpadding="0"
 id="TBL-15-1"><col 
 id="TBL-15-2"></colgroup><tr  
 style="vertical-align:baseline;" id="TBL-15-1-"><td  style="white-space:normal; text-align:left;" id="TBL-15-1-1"  
-class="td11"><!--l. 856--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 863--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">info</span>  </td><td  style="white-space:normal; text-align:left;" id="TBL-15-1-2"  
-class="td11"><!--l. 856--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 863--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">integer</span><span 
 class="cmtt-12">,</span><span 
 class="cmtt-12">&#x00A0;</span><span 
@@ -71,8 +71,8 @@ class="cmtt-12">)</span><span
 class="cmr-12">.</span>                                            </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-15-2-"><td  style="white-space:normal; text-align:left;" id="TBL-15-2-1"  
-class="td11"><!--l. 857--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-15-2-2"  
-class="td11"><!--l. 857--><p class="noindent" ><span 
+class="td11"><!--l. 864--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-15-2-2"  
+class="td11"><!--l. 864--><p class="noindent" ><span 
 class="cmr-12">Error code. If no error, 0 is returned. See Section</span><span 
 class="cmr-12">&#x00A0;</span><a 
 href="userhtmlse7.html#x28-320007"><span 
@@ -89,7 +89,7 @@ class="td11">      </td> </tr></table></div>
                                                                               

                                                                               
-   <!--l. 863--><div class="crosslinks"><p class="noindent"><span 
+   <!--l. 870--><div class="crosslinks"><p class="noindent"><span 
 class="cmr-12">[</span><a 
 href="userhtmlsu15.html" ><span 
 class="cmr-12">next</span></a><span 
@@ -106,6 +106,6 @@ class="cmr-12">] [</span><a
 href="userhtmlse5.html#userhtmlsu14.html" ><span 
 class="cmr-12">up</span></a><span 
 class="cmr-12">] </span></p></div>
-<!--l. 863--><p class="indent" >   <a 
+<!--l. 870--><p class="indent" >   <a 
 id="tailuserhtmlsu14.html"></a>  
 </body></html> 
@@ -10,7 +10,7 @@
 <link rel="stylesheet" type="text/css" href="userhtml.css"> 
 </head><body 
 >
-   <!--l. 863--><div class="crosslinks"><p class="noindent"><span 
+   <!--l. 870--><div class="crosslinks"><p class="noindent"><span 
 class="cmr-12">[</span><a 
 href="userhtmlsu16.html" ><span 
 class="cmr-12">next</span></a><span 
@@ -33,8 +33,8 @@ class="cmr-12">5.8   </span></span> <a
 class="cmr-12">Method descr</span></h4>
 <div class="center" 
 >
-<!--l. 865--><p class="noindent" >
-<!--l. 866--><p class="noindent" ><span class="lstinline"></span><span 
+<!--l. 872--><p class="noindent" >
+<!--l. 873--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">call</span><span 
 class="cmtt-12">&#x00A0;</span><span 
 class="cmtt-12">p</span><span 
@@ -54,7 +54,7 @@ class="cmtt-12">&#x00A0;</span><span
 class="cmtt-12">verbosity</span><span 
 class="cmtt-12">])</span><br />
 </div>
-<!--l. 869--><p class="noindent" ><span 
+<!--l. 876--><p class="noindent" ><span 
 class="cmr-12">This method prints a description of the preconditioner </span><span class="lstinline"></span><span 
 class="cmtt-12">p</span> <span 
 class="cmr-12">to the standard output or to a</span>
@@ -68,7 +68,7 @@ class="cmtt-12">build</span><span
 class="cmr-12">, have</span>
 <span 
 class="cmr-12">been called.</span>
-<!--l. 874--><p class="noindent" ><span 
+<!--l. 881--><p class="noindent" ><span 
 class="cmbx-12">Arguments</span>
   <div class="tabular"> <table id="TBL-16" class="tabular" 
 cellspacing="0" cellpadding="0"  
@@ -76,9 +76,9 @@ cellspacing="0" cellpadding="0"
 id="TBL-16-1"><col 
 id="TBL-16-2"></colgroup><tr  
 style="vertical-align:baseline;" id="TBL-16-1-"><td  style="white-space:normal; text-align:left;" id="TBL-16-1-1"  
-class="td11"><!--l. 877--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 884--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">info</span>      </td><td  style="white-space:normal; text-align:left;" id="TBL-16-1-2"  
-class="td11"><!--l. 877--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 884--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">integer</span><span 
 class="cmtt-12">,</span><span 
 class="cmtt-12">&#x00A0;</span><span 
@@ -89,8 +89,8 @@ class="cmtt-12">)</span><span
 class="cmr-12">.</span>                                                </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-16-2-"><td  style="white-space:normal; text-align:left;" id="TBL-16-2-1"  
-class="td11"><!--l. 878--><p class="noindent" >           </td><td  style="white-space:normal; text-align:left;" id="TBL-16-2-2"  
-class="td11"><!--l. 878--><p class="noindent" ><span 
+class="td11"><!--l. 885--><p class="noindent" >           </td><td  style="white-space:normal; text-align:left;" id="TBL-16-2-2"  
+class="td11"><!--l. 885--><p class="noindent" ><span 
 class="cmr-12">Error code. If no error, 0 is returned. See Section</span><span 
 class="cmr-12">&#x00A0;</span><a 
 href="userhtmlse7.html#x28-320007"><span 
@@ -98,9 +98,9 @@ class="cmr-12">7</span><!--tex4ht:ref: sec:errors --></a> <span
 class="cmr-12">for details.</span></td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-16-3-"><td  style="white-space:normal; text-align:left;" id="TBL-16-3-1"  
-class="td11"><!--l. 879--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 886--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">iout</span>       </td><td  style="white-space:normal; text-align:left;" id="TBL-16-3-2"  
-class="td11"><!--l. 879--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 886--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">integer</span><span 
 class="cmtt-12">,</span><span 
 class="cmtt-12">&#x00A0;</span><span 
@@ -113,16 +113,16 @@ class="cmtt-12">optional</span><span
 class="cmr-12">.</span>                                 </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-16-4-"><td  style="white-space:normal; text-align:left;" id="TBL-16-4-1"  
-class="td11"><!--l. 880--><p class="noindent" >           </td><td  style="white-space:normal; text-align:left;" id="TBL-16-4-2"  
-class="td11"><!--l. 880--><p class="noindent" ><span 
+class="td11"><!--l. 887--><p class="noindent" >           </td><td  style="white-space:normal; text-align:left;" id="TBL-16-4-2"  
+class="td11"><!--l. 887--><p class="noindent" ><span 
 class="cmr-12">The id of the file where the preconditioner description will be</span>
 <span 
 class="cmr-12">printed; the default is the standard output.</span>                       </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-16-5-"><td  style="white-space:normal; text-align:left;" id="TBL-16-5-1"  
-class="td11"><!--l. 882--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 889--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">root</span>       </td><td  style="white-space:normal; text-align:left;" id="TBL-16-5-2"  
-class="td11"><!--l. 882--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 889--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">integer</span><span 
 class="cmtt-12">,</span><span 
 class="cmtt-12">&#x00A0;</span><span 
@@ -135,8 +135,8 @@ class="cmtt-12">optional</span><span
 class="cmr-12">.</span>                                 </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-16-6-"><td  style="white-space:normal; text-align:left;" id="TBL-16-6-1"  
-class="td11"><!--l. 883--><p class="noindent" >           </td><td  style="white-space:normal; text-align:left;" id="TBL-16-6-2"  
-class="td11"><!--l. 883--><p class="noindent" ><span 
+class="td11"><!--l. 890--><p class="noindent" >           </td><td  style="white-space:normal; text-align:left;" id="TBL-16-6-2"  
+class="td11"><!--l. 890--><p class="noindent" ><span 
 class="cmr-12">The id of the process where the preconditioner description</span>
 <span 
 class="cmr-12">will be printed; the default is </span><span class="lstinline"></span><span 
@@ -144,9 +144,9 @@ class="cmtt-12">psb_root_</span><span
 class="cmr-12">.</span>                          </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-16-7-"><td  style="white-space:normal; text-align:left;" id="TBL-16-7-1"  
-class="td11"><!--l. 885--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 892--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">verbosity</span>  </td><td  style="white-space:normal; text-align:left;" id="TBL-16-7-2"  
-class="td11"><!--l. 885--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 892--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">integer</span><span 
 class="cmtt-12">,</span><span 
 class="cmtt-12">&#x00A0;</span><span 
@@ -159,8 +159,8 @@ class="cmtt-12">optional</span><span
 class="cmr-12">.</span>                                 </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-16-8-"><td  style="white-space:normal; text-align:left;" id="TBL-16-8-1"  
-class="td11"><!--l. 886--><p class="noindent" >           </td><td  style="white-space:normal; text-align:left;" id="TBL-16-8-2"  
-class="td11"><!--l. 886--><p class="noindent" ><span 
+class="td11"><!--l. 893--><p class="noindent" >           </td><td  style="white-space:normal; text-align:left;" id="TBL-16-8-2"  
+class="td11"><!--l. 893--><p class="noindent" ><span 
 class="cmr-12">The verbosity level of the description. Default value is 0. For</span>
 <span 
 class="cmr-12">values higher than 0, it prints out further information, e.g., for</span>
@@ -174,7 +174,7 @@ class="td11">           </td> </tr></table></div>
                                                                               

                                                                               
-   <!--l. 893--><div class="crosslinks"><p class="noindent"><span 
+   <!--l. 900--><div class="crosslinks"><p class="noindent"><span 
 class="cmr-12">[</span><a 
 href="userhtmlsu16.html" ><span 
 class="cmr-12">next</span></a><span 
@@ -191,6 +191,6 @@ class="cmr-12">] [</span><a
 href="userhtmlse5.html#userhtmlsu15.html" ><span 
 class="cmr-12">up</span></a><span 
 class="cmr-12">] </span></p></div>
-<!--l. 893--><p class="indent" >   <a 
+<!--l. 900--><p class="indent" >   <a 
 id="tailuserhtmlsu15.html"></a>  
 </body></html> 
@@ -10,7 +10,7 @@
 <link rel="stylesheet" type="text/css" href="userhtml.css"> 
 </head><body 
 >
-   <!--l. 893--><div class="crosslinks"><p class="noindent"><span 
+   <!--l. 900--><div class="crosslinks"><p class="noindent"><span 
 class="cmr-12">[</span><a 
 href="userhtmlsu15.html" ><span 
 class="cmr-12">prev</span></a><span 
@@ -28,19 +28,19 @@ class="cmr-12">] </span></p></div>
 class="cmr-12">5.9   </span></span> <a 
 id="x26-250005.9"></a><span 
 class="cmr-12">Auxiliary Methods</span></h4>
-<!--l. 894--><p class="noindent" ><span 
+<!--l. 901--><p class="noindent" ><span 
 class="cmr-12">Various functionalities are implemented as additional methods of the preconditioner</span>
 <span 
 class="cmr-12">object.</span>
-<!--l. 897--><p class="noindent" >
+<!--l. 904--><p class="noindent" >
   <h5 class="subsubsectionHead"><span class="titlemark"><span 
 class="cmr-12">5.9.1   </span></span> <a 
 id="x26-260005.9.1"></a><span 
 class="cmr-12">Method: dump</span></h5>
 <div class="center" 
 >
-<!--l. 899--><p class="noindent" >
-<!--l. 900--><p class="noindent" ><span class="lstinline"></span><span 
+<!--l. 906--><p class="noindent" >
+<!--l. 907--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">call</span><span 
 class="cmtt-12">&#x00A0;</span><span 
 class="cmtt-12">p</span><span 
@@ -68,9 +68,9 @@ class="cmtt-12">,</span><span
 class="cmtt-12">global_num</span><span 
 class="cmtt-12">])</span><br />
 </div>
-<!--l. 903--><p class="noindent" ><span 
+<!--l. 910--><p class="noindent" ><span 
 class="cmr-12">Dump on file.</span>
-<!--l. 906--><p class="noindent" ><span 
+<!--l. 913--><p class="noindent" ><span 
 class="cmbx-12">Arguments</span>
   <div class="tabular"> <table id="TBL-17" class="tabular" 
 cellspacing="0" cellpadding="0"  
@@ -78,9 +78,9 @@ cellspacing="0" cellpadding="0"
 id="TBL-17-1"><col 
 id="TBL-17-2"></colgroup><tr  
 style="vertical-align:baseline;" id="TBL-17-1-"><td  style="white-space:normal; text-align:left;" id="TBL-17-1-1"  
-class="td11"><!--l. 909--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 916--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">info</span>  </td><td  style="white-space:normal; text-align:left;" id="TBL-17-1-2"  
-class="td11"><!--l. 909--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 916--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">integer</span><span 
 class="cmtt-12">,</span><span 
 class="cmtt-12">&#x00A0;</span><span 
@@ -91,8 +91,8 @@ class="cmtt-12">)</span><span
 class="cmr-12">.</span>                                                       </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-17-2-"><td  style="white-space:normal; text-align:left;" id="TBL-17-2-1"  
-class="td11"><!--l. 910--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-17-2-2"  
-class="td11"><!--l. 910--><p class="noindent" ><span 
+class="td11"><!--l. 917--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-17-2-2"  
+class="td11"><!--l. 917--><p class="noindent" ><span 
 class="cmr-12">Error code. If no error, 0 is returned. See Section</span><span 
 class="cmr-12">&#x00A0;</span><a 
 href="userhtmlse7.html#x28-320007"><span 
@@ -100,9 +100,9 @@ class="cmr-12">7</span><!--tex4ht:ref: sec:errors --></a> <span
 class="cmr-12">for details.</span>     </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-17-3-"><td  style="white-space:normal; text-align:left;" id="TBL-17-3-1"  
-class="td11"><!--l. 911--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 918--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">amold</span> </td><td  style="white-space:normal; text-align:left;" id="TBL-17-3-2"  
-class="td11"><!--l. 911--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 918--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">class</span><span 
 class="cmtt-12">(</span><span 
 class="cmtt-12">psb_</span><span 
@@ -119,8 +119,8 @@ class="cmtt-12">optional</span><span
 class="cmr-12">.</span>       </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-17-4-"><td  style="white-space:normal; text-align:left;" id="TBL-17-4-1"  
-class="td11"><!--l. 912--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-17-4-2"  
-class="td11"><!--l. 912--><p class="noindent" ><span 
+class="td11"><!--l. 919--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-17-4-2"  
+class="td11"><!--l. 919--><p class="noindent" ><span 
 class="cmr-12">The  desired  dynamic  type  for  internal  matrix  components;  this</span>
 <span 
 class="cmr-12">allows  e.g.  running  on  GPUs;  it  needs  not  be  the  same  on  all</span>
@@ -134,15 +134,15 @@ class="cmr-12">.</span>             </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-17-5-"><td  style="white-space:normal; text-align:left;" id="TBL-17-5-1"  
 class="td11">      </td> </tr></table></div>
-<!--l. 919--><p class="noindent" >
+<!--l. 926--><p class="noindent" >
   <h5 class="subsubsectionHead"><span class="titlemark"><span 
 class="cmr-12">5.9.2   </span></span> <a 
 id="x26-270005.9.2"></a><span 
 class="cmr-12">Method: clone</span></h5>
 <div class="center" 
 >
-<!--l. 921--><p class="noindent" >
-<!--l. 922--><p class="noindent" ><span class="lstinline"></span><span 
+<!--l. 928--><p class="noindent" >
+<!--l. 929--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">call</span><span 
 class="cmtt-12">&#x00A0;</span><span 
 class="cmtt-12">p</span><span 
@@ -157,9 +157,9 @@ class="cmtt-12">)</span><br />
                                                                               

                                                                               
-<!--l. 925--><p class="noindent" ><span 
+<!--l. 932--><p class="noindent" ><span 
 class="cmr-12">Create a (deep) copy of the preconditioner object.</span>
-<!--l. 928--><p class="noindent" ><span 
+<!--l. 935--><p class="noindent" ><span 
 class="cmbx-12">Arguments</span>
   <div class="tabular"> <table id="TBL-18" class="tabular" 
 cellspacing="0" cellpadding="0"  
@@ -167,9 +167,9 @@ cellspacing="0" cellpadding="0"
 id="TBL-18-1"><col 
 id="TBL-18-2"></colgroup><tr  
 style="vertical-align:baseline;" id="TBL-18-1-"><td  style="white-space:normal; text-align:left;" id="TBL-18-1-1"  
-class="td11"><!--l. 931--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 938--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">pout</span>  </td><td  style="white-space:normal; text-align:left;" id="TBL-18-1-2"  
-class="td11"><!--l. 931--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 938--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">type</span><span 
 class="cmtt-12">(</span><span 
 class="cmtt-12">amg_</span><span 
@@ -184,8 +184,8 @@ class="cmtt-12">)</span><span
 class="cmr-12">.</span>                                  </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-18-2-"><td  style="white-space:normal; text-align:left;" id="TBL-18-2-1"  
-class="td11"><!--l. 932--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-18-2-2"  
-class="td11"><!--l. 932--><p class="noindent" ><span 
+class="td11"><!--l. 939--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-18-2-2"  
+class="td11"><!--l. 939--><p class="noindent" ><span 
 class="cmr-12">The copy of the preconditioner data structure. Note that </span><span 
 class="cmti-12">x </span><span 
 class="cmr-12">must</span>
@@ -199,9 +199,9 @@ class="cmr-12">, single/double precision</span>
 class="cmr-12">version of AMG4PSBLAS under use.</span>                                       </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-18-3-"><td  style="white-space:normal; text-align:left;" id="TBL-18-3-1"  
-class="td11"><!--l. 935--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 942--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">info</span>  </td><td  style="white-space:normal; text-align:left;" id="TBL-18-3-2"  
-class="td11"><!--l. 935--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 942--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">integer</span><span 
 class="cmtt-12">,</span><span 
 class="cmtt-12">&#x00A0;</span><span 
@@ -212,8 +212,8 @@ class="cmtt-12">)</span><span
 class="cmr-12">.</span>                                                       </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-18-4-"><td  style="white-space:normal; text-align:left;" id="TBL-18-4-1"  
-class="td11"><!--l. 936--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-18-4-2"  
-class="td11"><!--l. 936--><p class="noindent" ><span 
+class="td11"><!--l. 943--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-18-4-2"  
+class="td11"><!--l. 943--><p class="noindent" ><span 
 class="cmr-12">Error code. If no error, 0 is returned. See Section</span><span 
 class="cmr-12">&#x00A0;</span><a 
 href="userhtmlse7.html#x28-320007"><span 
@@ -222,15 +222,15 @@ class="cmr-12">for details.</span>     </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-18-5-"><td  style="white-space:normal; text-align:left;" id="TBL-18-5-1"  
 class="td11">      </td> </tr></table></div>
-<!--l. 941--><p class="noindent" >
+<!--l. 948--><p class="noindent" >
   <h5 class="subsubsectionHead"><span class="titlemark"><span 
 class="cmr-12">5.9.3   </span></span> <a 
 id="x26-280005.9.3"></a><span 
 class="cmr-12">Method: sizeof</span></h5>
 <div class="center" 
 >
-<!--l. 943--><p class="noindent" >
-<!--l. 944--><p class="noindent" ><span class="lstinline"></span><span 
+<!--l. 950--><p class="noindent" >
+<!--l. 951--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">sz</span><span 
 class="cmtt-12">&#x00A0;</span><span 
 class="cmtt-12">=</span><span 
@@ -248,9 +248,9 @@ cellspacing="0" cellpadding="0"
 id="TBL-19-1"><col 
 id="TBL-19-2"></colgroup><tr  
 style="vertical-align:baseline;" id="TBL-19-1-"><td  style="white-space:normal; text-align:left;" id="TBL-19-1-1"  
-class="td11"><!--l. 948--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 955--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">global</span> </td> <td  style="white-space:normal; text-align:left;" id="TBL-19-1-2"  
-class="td11"><!--l. 948--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 955--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">logical</span><span 
 class="cmtt-12">,</span><span 
 class="cmtt-12">&#x00A0;</span><span 
@@ -258,8 +258,8 @@ class="cmtt-12">optional</span><span
 class="cmr-12">.</span>                                                            </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-19-2-"><td  style="white-space:normal; text-align:left;" id="TBL-19-2-1"  
-class="td11"><!--l. 949--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-19-2-2"  
-class="td11"><!--l. 949--><p class="noindent" ><span 
+class="td11"><!--l. 956--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-19-2-2"  
+class="td11"><!--l. 956--><p class="noindent" ><span 
 class="cmr-12">Whether the global or local preconditioner memory occupation is</span>
 <span 
 class="cmr-12">desired. Default: </span><span class="lstinline"></span><span 
@@ -272,7 +272,7 @@ class="cmr-12">.</span>                                                     </td
 class="td11">      </td> </tr></table>
 </div> <span 
 class="cmr-12">Return memory footprint in bytes.</span>
-<!--l. 956--><p class="noindent" >
+<!--l. 963--><p class="noindent" >
   <h5 class="subsubsectionHead"><span class="titlemark"><span 
 class="cmr-12">5.9.4   </span></span> <a 
 id="x26-290005.9.4"></a><span 
@@ -280,8 +280,8 @@ class="cmr-12">Method: allocate</span><span
 class="cmr-12">_wrk</span></h5>
 <div class="center" 
 >
-<!--l. 958--><p class="noindent" >
-<!--l. 959--><p class="noindent" ><span class="lstinline"></span><span 
+<!--l. 965--><p class="noindent" >
+<!--l. 966--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">call</span><span 
 class="cmtt-12">&#x00A0;</span><span 
 class="cmtt-12">p</span><span 
@@ -294,7 +294,7 @@ class="cmtt-12">&#x00A0;</span><span
 class="cmtt-12">vmold</span><span 
 class="cmtt-12">])</span><br />
 </div>
-<!--l. 962--><p class="noindent" ><span 
+<!--l. 969--><p class="noindent" ><span 
 class="cmr-12">Allocate internal work vectors. Each application of the preconditioner uses a number of</span>
 <span 
 class="cmr-12">work vectors which are allocated internally as necessary; therefore allocation and</span>
@@ -326,7 +326,7 @@ class="cmr-12">argument to the </span><span class="lstinline"></span><span
 class="cmtt-12">apply</span>
 <span 
 class="cmr-12">method.</span>
-<!--l. 978--><p class="noindent" ><span 
+<!--l. 985--><p class="noindent" ><span 
 class="cmbx-12">Arguments</span>
   <div class="tabular"> <table id="TBL-20" class="tabular" 
 cellspacing="0" cellpadding="0"  
@@ -334,9 +334,9 @@ cellspacing="0" cellpadding="0"
 id="TBL-20-1"><col 
 id="TBL-20-2"></colgroup><tr  
 style="vertical-align:baseline;" id="TBL-20-1-"><td  style="white-space:normal; text-align:left;" id="TBL-20-1-1"  
-class="td11"><!--l. 981--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 988--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">info</span>  </td><td  style="white-space:normal; text-align:left;" id="TBL-20-1-2"  
-class="td11"><!--l. 981--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 988--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">integer</span><span 
 class="cmtt-12">,</span><span 
 class="cmtt-12">&#x00A0;</span><span 
@@ -347,8 +347,8 @@ class="cmtt-12">)</span><span
 class="cmr-12">.</span>                                                       </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-20-2-"><td  style="white-space:normal; text-align:left;" id="TBL-20-2-1"  
-class="td11"><!--l. 982--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-20-2-2"  
-class="td11"><!--l. 982--><p class="noindent" ><span 
+class="td11"><!--l. 989--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-20-2-2"  
+class="td11"><!--l. 989--><p class="noindent" ><span 
 class="cmr-12">Error code. If no error, 0 is returned. See Section</span><span 
 class="cmr-12">&#x00A0;</span><a 
 href="userhtmlse7.html#x28-320007"><span 
@@ -356,9 +356,9 @@ class="cmr-12">7</span><!--tex4ht:ref: sec:errors --></a> <span
 class="cmr-12">for details.</span>     </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-20-3-"><td  style="white-space:normal; text-align:left;" id="TBL-20-3-1"  
-class="td11"><!--l. 983--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 990--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">vmold</span> </td><td  style="white-space:normal; text-align:left;" id="TBL-20-3-2"  
-class="td11"><!--l. 983--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 990--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">class</span><span 
 class="cmtt-12">(</span><span 
 class="cmtt-12">psb_</span><span 
@@ -375,15 +375,15 @@ class="cmtt-12">optional</span><span
 class="cmr-12">.</span>         </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-20-4-"><td  style="white-space:normal; text-align:left;" id="TBL-20-4-1"  
-class="td11"><!--l. 984--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-20-4-2"  
-class="td11"><!--l. 984--><p class="noindent" ><span 
+class="td11"><!--l. 991--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-20-4-2"  
+class="td11"><!--l. 991--><p class="noindent" ><span 
 class="cmr-12">The  desired  dynamic  type  for  internal  vector  components;  this</span>
 <span 
 class="cmr-12">allows e.g. running on GPUs.</span>                                                 </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-20-5-"><td  style="white-space:normal; text-align:left;" id="TBL-20-5-1"  
 class="td11">      </td> </tr></table></div>
-<!--l. 990--><p class="noindent" >
+<!--l. 997--><p class="noindent" >
   <h5 class="subsubsectionHead"><span class="titlemark"><span 
 class="cmr-12">5.9.5   </span></span> <a 
 id="x26-300005.9.5"></a><span 
@@ -391,8 +391,8 @@ class="cmr-12">Method: free</span><span
 class="cmr-12">_wrk</span></h5>
 <div class="center" 
 >
-<!--l. 992--><p class="noindent" >
-<!--l. 993--><p class="noindent" ><span class="lstinline"></span><span 
+<!--l. 999--><p class="noindent" >
+<!--l. 1000--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">call</span><span 
 class="cmtt-12">&#x00A0;</span><span 
 class="cmtt-12">p</span><span 
@@ -402,9 +402,9 @@ class="cmtt-12">(</span><span
 class="cmtt-12">info</span><span 
 class="cmtt-12">)</span><br />
 </div>
-<!--l. 996--><p class="noindent" ><span 
+<!--l. 1003--><p class="noindent" ><span 
 class="cmr-12">Deallocate internal work vectors.</span>
-<!--l. 999--><p class="noindent" ><span 
+<!--l. 1006--><p class="noindent" ><span 
 class="cmbx-12">Arguments</span>
   <div class="tabular"> <table id="TBL-21" class="tabular" 
 cellspacing="0" cellpadding="0"  
@@ -412,9 +412,9 @@ cellspacing="0" cellpadding="0"
 id="TBL-21-1"><col 
 id="TBL-21-2"></colgroup><tr  
 style="vertical-align:baseline;" id="TBL-21-1-"><td  style="white-space:normal; text-align:left;" id="TBL-21-1-1"  
-class="td11"><!--l. 1002--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 1009--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">info</span>  </td><td  style="white-space:normal; text-align:left;" id="TBL-21-1-2"  
-class="td11"><!--l. 1002--><p class="noindent" ><span class="lstinline"></span><span 
+class="td11"><!--l. 1009--><p class="noindent" ><span class="lstinline"></span><span 
 class="cmtt-12">integer</span><span 
 class="cmtt-12">,</span><span 
 class="cmtt-12">&#x00A0;</span><span 
@@ -425,8 +425,8 @@ class="cmtt-12">)</span><span
 class="cmr-12">.</span>                                                       </td>
 </tr><tr  
 style="vertical-align:baseline;" id="TBL-21-2-"><td  style="white-space:normal; text-align:left;" id="TBL-21-2-1"  
-class="td11"><!--l. 1003--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-21-2-2"  
-class="td11"><!--l. 1003--><p class="noindent" ><span 
+class="td11"><!--l. 1010--><p class="noindent" >      </td><td  style="white-space:normal; text-align:left;" id="TBL-21-2-2"  
+class="td11"><!--l. 1010--><p class="noindent" ><span 
 class="cmr-12">Error code. If no error, 0 is returned. See Section</span><span 
 class="cmr-12">&#x00A0;</span><a 
 href="userhtmlse7.html#x28-320007"><span 
@@ -4228,9 +4228,9 @@ class="cmtt-12">github</span><span
 class="cmtt-12">.</span><span 
 class="cmtt-12">com</span><span 
 class="cmtt-12">/</span><span 
-class="cmtt-12">sfilippone</span><span 
+class="cmtt-12">psctoolkit</span><span 
 class="cmtt-12">/</span><span 
-class="cmtt-12">amg4psblas</span><span 
+class="cmtt-12">psctoolkit</span><span 
 class="cmtt-12">/</span><span 
 class="cmtt-12">issues</span><span 
 class="cmtt-12">&#x003E;.</span>
@@ -36,8 +36,8 @@ class="cmr-12">If you find any bugs in our codes, please report them through our
 <span 
 class="cmr-12">on</span><br 
 class="newline" /> <a 
-href="https://github.com/psctoolkit/amg4psblas/issues" class="url" ><span 
-class="cmtt-12">https://github.com/psctoolkit/amg4psblas/issues</span></a><br 
+href="https://github.com/psctoolkit/psctoolkit/issues" class="url" ><span 
+class="cmtt-12">https://github.com/psctoolkit/psctoolkit/issues</span></a><br 
 class="newline" />
 <!--l. 195--><p class="indent" >   <span 
 class="cmr-12">To enable us to track the bug, please provide a log from the failing application, the</span>
@@ -29,40 +29,41 @@ class="cmr-12">3.5   </span></span> <a
 id="x13-120003.5"></a><span 
 class="cmr-12">Example and test programs</span></h4>
 <!--l. 200--><p class="noindent" ><span 
-class="cmr-12">The package contains the </span><span class="obeylines-h"><span class="verb"><span 
-class="cmtt-12">examples</span></span></span> <span 
-class="cmr-12">and </span><span class="obeylines-h"><span class="verb"><span 
-class="cmtt-12">tests</span></span></span> <span 
-class="cmr-12">directories; both of them are</span>
-<span 
-class="cmr-12">further divided into </span><span class="obeylines-h"><span class="verb"><span 
+class="cmr-12">The package contains a </span><span class="obeylines-h"><span class="verb"><span 
+class="cmtt-12">samples</span></span></span> <span 
+class="cmr-12">directory, divided in two subdirs </span><span class="obeylines-h"><span class="verb"><span 
+class="cmtt-12">simple</span></span></span> <span 
+class="cmr-12">and</span>
+<span class="obeylines-h"><span class="verb"><span 
+class="cmtt-12">advanced</span></span></span><span 
+class="cmr-12">; both of them are further divided into </span><span class="obeylines-h"><span class="verb"><span 
 class="cmtt-12">fileread</span></span></span> <span 
 class="cmr-12">and </span><span class="obeylines-h"><span class="verb"><span 
 class="cmtt-12">pdegen</span></span></span> <span 
-class="cmr-12">subdirectories. Their purpose is as</span>
+class="cmr-12">subdirectories.</span>
 <span 
-class="cmr-12">follows:</span>
+class="cmr-12">Their purpose is as follows:</span>
     <dl class="description"><dt class="description">
 <span 
-class="cmtt-12">examples</span> </dt><dd 
+class="cmtt-12">simple</span> </dt><dd 
 class="description"><span 
-class="cmr-12">contains a set of simple example programs with a predefined choice</span>
+class="cmr-12">contains a set of simple example programs with a predefined choice of</span>
     <span 
-class="cmr-12">of preconditioners, selectable via integer values. These are intended to get</span>
+class="cmr-12">preconditioners,  selectable  via  integer  values.  These  are  intended  to  get</span>
     <span 
 class="cmr-12">acquainted with the multilevel preconditioners available in AMG4PSBLAS.</span>
     </dd><dt class="description">
 <span 
-class="cmtt-12">tests</span> </dt><dd 
+class="cmtt-12">advanced</span> </dt><dd 
 class="description"><span 
-class="cmr-12">contains a set of more sophisticated examples that will allow the user, via</span>
+class="cmr-12">contains a set of more sophisticated examples that will allow the user,</span>
     <span 
-class="cmr-12">the input files in the </span><span class="obeylines-h"><span class="verb"><span 
+class="cmr-12">via the input files in the </span><span class="obeylines-h"><span class="verb"><span 
 class="cmtt-12">runs</span></span></span> <span 
-class="cmr-12">subdirectories, to experiment with the full range</span>
+class="cmr-12">subdirectories, to experiment with the full</span>
     <span 
-class="cmr-12">of preconditioners implemented in the package.</span></dd></dl>
-<!--l. 213--><p class="noindent" ><span 
+class="cmr-12">range of preconditioners implemented in the package.</span></dd></dl>
+<!--l. 214--><p class="noindent" ><span 
 class="cmr-12">The </span><span class="obeylines-h"><span class="verb"><span 
 class="cmtt-12">fileread</span></span></span> <span 
 class="cmr-12">directories contain sample programs that read sparse matrices from files,</span>
@@ -71,38 +71,39 @@ class="cmr-12">The part of the code dealing with reading and assembling the spar
 <span 
 class="cmr-12">right-hand side vector and the deallocation of the relevant data structures, performed</span>
 <span 
-class="cmr-12">through the PSBLAS routines for sparse matrix and vector management, is not</span>
+class="cmr-12">through the PSBLAS routines for sparse matrix and vector management,</span>
 <span 
-class="cmr-12">reported here for the sake of conciseness. The complete code can be found in the</span>
+class="cmr-12">is not reported here for the sake of conciseness. The complete code can be</span>
 <span 
-class="cmr-12">example program file </span><span class="obeylines-h"><span class="verb"><span 
+class="cmr-12">found in the example program file </span><span class="obeylines-h"><span class="verb"><span 
 class="cmtt-12">amg_dexample_ml.f90</span></span></span><span 
-class="cmr-12">, in the directory </span><span class="obeylines-h"><span class="verb"><span 
-class="cmtt-12">examples/fileread</span></span></span> <span 
-class="cmr-12">of</span>
-<span 
-class="cmr-12">the AMG4PSBLAS implementation (see Section</span><span 
+class="cmr-12">, in the directory</span>
+<span class="obeylines-h"><span class="verb"><span 
+class="cmtt-12">samples/simple/file</span></span></span><span class="obeylines-h"><span class="verb"><span 
+class="cmtt-12">read</span></span></span> <span 
+class="cmr-12">of the AMG4PSBLAS implementation (see Section</span><span 
 class="cmr-12">&#x00A0;</span><a 
 href="userhtmlsu5.html#x13-120003.5"><span 
 class="cmr-12">3.5</span><!--tex4ht:ref: sec:ex_and_test --></a><span 
-class="cmr-12">). A sample test problem along</span>
+class="cmr-12">). A</span>
 <span 
-class="cmr-12">with the relevant input data is available in </span><span class="obeylines-h"><span class="verb"><span 
-class="cmtt-12">examples/fileread/runs</span></span></span><span 
-class="cmr-12">. For details on</span>
+class="cmr-12">sample test problem along with the relevant input data is available in</span>
+<span class="obeylines-h"><span class="verb"><span 
+class="cmtt-12">samples/simple/fileread/runs</span></span></span><span 
+class="cmr-12">. For details on the use of the PSBLAS routines, see</span>
 <span 
-class="cmr-12">the use of the PSBLAS routines, see the PSBLAS User&#8217;s Guide</span><span 
+class="cmr-12">the PSBLAS User&#8217;s Guide</span><span 
 class="cmr-12">&#x00A0;</span><span class="cite"><span 
 class="cmr-12">[</span><a 
 href="userhtmlli5.html#XPSBLASGUIDE"><span 
 class="cmr-12">20</span></a><span 
 class="cmr-12">]</span></span><span 
 class="cmr-12">.</span>
-<!--l. 138--><p class="indent" >   <span 
-class="cmr-12">The setup and application of the default multilevel preconditioner for the real single</span>
                                                                               

                                                                               
+<!--l. 138--><p class="indent" >   <span 
+class="cmr-12">The setup and application of the default multilevel preconditioner for the real single</span>
 <span 
 class="cmr-12">precision and the complex, single and double precision, versions are obtained</span>
 <span 
@@ -114,7 +115,8 @@ class="cmr-12">for</span>
 <span 
 class="cmr-12">details). If these versions are installed, the corresponding codes are available in</span>
 <span class="obeylines-h"><span class="verb"><span 
-class="cmtt-12">examples/fileread/</span></span></span><span 
+class="cmtt-12">samples/simple/file</span></span></span><span class="obeylines-h"><span class="verb"><span 
+class="cmtt-12">read</span></span></span><span 
 class="cmr-12">.</span>
                                                                               

@@ -160,7 +162,7 @@ class="cmr-12">.</span>
 !&#x00A0;with&#x00A0;basic&#x00A0;smoothed&#x00A0;aggregation,&#x00A0;1&#x00A0;hybrid&#x00A0;forward/backward
 !&#x00A0;GS&#x00A0;sweep&#x00A0;as&#x00A0;pre/post-smoother&#x00A0;and&#x00A0;UMFPACK&#x00A0;as&#x00A0;coarsest-level
 !&#x00A0;solver
-&#x00A0;&#x00A0;call&#x00A0;P%init(&#8217;ML&#8217;,info)
+&#x00A0;&#x00A0;call&#x00A0;P%init(ctxt,&#8217;ML&#8217;,info)
 !
 !&#x00A0;build&#x00A0;the&#x00A0;preconditioner
 &#x00A0;&#x00A0;call&#x00A0;P%hierarchy_build(A,desc_A,info)
@@ -169,8 +171,8 @@ class="cmr-12">.</span>
 !&#x00A0;set&#x00A0;the&#x00A0;solver&#x00A0;parameters&#x00A0;and&#x00A0;the&#x00A0;initial&#x00A0;guess
 &#x00A0;&#x00A0;...&#x00A0;...
 !
-!&#x00A0;solve&#x00A0;Ax=b&#x00A0;with&#x00A0;preconditioned&#x00A0;CG
-&#x00A0;&#x00A0;call&#x00A0;psb_krylov(&#8217;CG&#8217;,A,P,b,x,tol,desc_A,info)
+!&#x00A0;solve&#x00A0;Ax=b&#x00A0;with&#x00A0;preconditioned&#x00A0;FCG
+&#x00A0;&#x00A0;call&#x00A0;psb_krylov(&#8217;FCG&#8217;,A,P,b,x,tol,desc_A,info)
 &#x00A0;&#x00A0;...&#x00A0;...
 !
 !&#x00A0;deallocate&#x00A0;the&#x00A0;preconditioner
@@ -277,7 +279,7 @@ class="cmr-12">For all the previous preconditioners, example programs where the
 class="cmr-12">and the right-hand side are generated by discretizing a PDE with Dirichlet</span>
 <span 
 class="cmr-12">boundary conditions are also available in the directory </span><span class="obeylines-h"><span class="verb"><span 
-class="cmtt-12">examples/pdegen</span></span></span><span 
+class="cmtt-12">samples/simple/pdegen</span></span></span><span 
 class="cmr-12">.</span>
                                                                               

@@ -296,7 +298,7 @@ class="cmr-12">.</span>
 !&#x00A0;build&#x00A0;a&#x00A0;V-cycle&#x00A0;preconditioner&#x00A0;with&#x00A0;1&#x00A0;block-Jacobi&#x00A0;sweep&#x00A0;(with
 !&#x00A0;ILU(0)&#x00A0;on&#x00A0;the&#x00A0;blocks)&#x00A0;as&#x00A0;pre-&#x00A0;and&#x00A0;post-smoother,&#x00A0;and&#x00A0;8&#x00A0;&#x00A0;block-Jacobi
 !&#x00A0;sweeps&#x00A0;(with&#x00A0;ILU(0)&#x00A0;on&#x00A0;the&#x00A0;blocks)&#x00A0;as&#x00A0;coarsest-level&#x00A0;solver
-&#x00A0;&#x00A0;call&#x00A0;P%init(&#8217;ML&#8217;,info)
+&#x00A0;&#x00A0;call&#x00A0;P%init(ctxt,&#8217;ML&#8217;,info)
 &#x00A0;&#x00A0;call&#x00A0;P%set(&#8217;SMOOTHER_TYPE&#8217;,&#8217;BJAC&#8217;,info)
 &#x00A0;&#x00A0;call&#x00A0;P%set(&#8217;COARSE_SOLVE&#8217;,&#8217;BJAC&#8217;,info)
 &#x00A0;&#x00A0;call&#x00A0;P%set(&#8217;COARSE_SWEEPS&#8217;,8,info)
@@ -323,25 +325,27 @@ class="content">setup of a multilevel preconditioner based on the default decoup
                                                                               
 <div class="center" 
 >
-<!--l. 360--><p class="noindent" >
+<!--l. 362--><p class="noindent" >
 <div class="minipage"><pre class="verbatim" id="verbatim-8">
 ...&#x00A0;...
 !&#x00A0;build&#x00A0;a&#x00A0;W-cycle&#x00A0;preconditioner&#x00A0;with&#x00A0;2&#x00A0;hybrid&#x00A0;Gauss-Seidel&#x00A0;sweeps
 !&#x00A0;as&#x00A0;pre-&#x00A0;and&#x00A0;post-smoother,&#x00A0;a&#x00A0;distributed&#x00A0;coarsest
 !&#x00A0;matrix,&#x00A0;and&#x00A0;MUMPS&#x00A0;as&#x00A0;coarsest-level&#x00A0;solver
-&#x00A0;&#x00A0;call&#x00A0;P%init(&#8217;ML&#8217;,info)
+&#x00A0;&#x00A0;call&#x00A0;P%init(ctxt,&#8217;ML&#8217;,info)
 &#x00A0;&#x00A0;call&#x00A0;P%set(&#8217;PAR_AGGR_ALG&#8217;,&#8217;COUPLED&#8217;,info)
-call&#x00A0;P%set(&#8217;AGGR_TYPE&#8217;,&#8217;MATCHBOXP&#8217;,info)
-call&#x00A0;P%set(&#8217;AGGR_SIZE&#8217;,8,info)
+&#x00A0;&#x00A0;call&#x00A0;P%set(&#8217;AGGR_TYPE&#8217;,&#8217;MATCHBOXP&#8217;,info)
+&#x00A0;&#x00A0;call&#x00A0;P%set(&#8217;AGGR_SIZE&#8217;,8,info)
 &#x00A0;&#x00A0;call&#x00A0;P%set(&#8217;ML_CYCLE&#8217;,&#8217;WCYCLE&#8217;,info)
 &#x00A0;&#x00A0;call&#x00A0;P%set(&#8217;SMOOTHER_TYPE&#8217;,&#8217;FBGS&#8217;,info)
 &#x00A0;&#x00A0;call&#x00A0;P%set(&#8217;SMOOTHER_SWEEPS&#8217;,2,info)
 &#x00A0;&#x00A0;call&#x00A0;P%set(&#8217;COARSE_SOLVE&#8217;,&#8217;KRM&#8217;,info)
+&#x00A0;&#x00A0;call&#x00A0;P%set(&#8217;COARSE_MAT&#8217;,&#8217;DIST&#8217;,info)
+&#x00A0;&#x00A0;call&#x00A0;P%set(&#8217;KRM_METHOD&#8217;,&#8217;FCG&#8217;,info)
 &#x00A0;&#x00A0;call&#x00A0;P%hierarchy_build(A,desc_A,info)
 &#x00A0;&#x00A0;call&#x00A0;P%smoothers_build(A,desc_A,info)
 ...&#x00A0;...
 </pre>
-<!--l. 379--><p class="nopar" ></div></div>
+<!--l. 383--><p class="nopar" ></div></div>
 <br /> <div class="caption" 
 ><span class="id">Listing 3: </span><span  
 class="content">setup of a multilevel preconditioner based on the coupled coarsening using
@@ -353,7 +357,7 @@ weighted matching</span></div><!--tex4ht:label?: x15-14003r3 -->
                                                                               

                                                                               
-<!--l. 386--><p class="indent" >   <a 
+<!--l. 390--><p class="indent" >   <a 
 id="x15-14004r4"></a><hr class="float"><div class="float" 
 >
                                                                               
@@ -361,18 +365,18 @@ weighted matching</span></div><!--tex4ht:label?: x15-14003r3 -->
                                                                               
 <div class="center" 
 >
-<!--l. 398--><p class="noindent" >
+<!--l. 402--><p class="noindent" >
 <div class="minipage"><pre class="verbatim" id="verbatim-9">
 ...&#x00A0;...
 !&#x00A0;set&#x00A0;RAS&#x00A0;with&#x00A0;overlap&#x00A0;2&#x00A0;and&#x00A0;ILU(0)&#x00A0;on&#x00A0;the&#x00A0;local&#x00A0;blocks
-&#x00A0;&#x00A0;call&#x00A0;P%init(&#8217;AS&#8217;,info)
+&#x00A0;&#x00A0;call&#x00A0;P%init(ctxt,&#8217;AS&#8217;,info)
 &#x00A0;&#x00A0;call&#x00A0;P%set(&#8217;SUB_OVR&#8217;,2,info)
 &#x00A0;&#x00A0;call&#x00A0;P%bld(A,desc_A,info)
 ...&#x00A0;...
 !&#x00A0;solve&#x00A0;Ax=b&#x00A0;with&#x00A0;preconditioned&#x00A0;BiCGSTAB
 &#x00A0;&#x00A0;call&#x00A0;psb_krylov(&#8217;BICGSTAB&#8217;,A,P,b,x,tol,desc_A,info)
 </pre>
-<!--l. 410--><p class="nopar" ></div></div>
+<!--l. 414--><p class="nopar" ></div></div>
 <br /> <div class="caption" 
 ><span class="id">Listing 4: </span><span  
 class="content">setup of a one-level Schwarz preconditioner.</span></div><!--tex4ht:label?: x15-14004r4 -->
@@ -383,7 +387,7 @@ class="content">setup of a one-level Schwarz preconditioner.</span></div><!--tex
                                                                               

                                                                               
-   <!--l. 420--><div class="crosslinks"><p class="noindent"><span 
+   <!--l. 424--><div class="crosslinks"><p class="noindent"><span 
 class="cmr-12">[</span><a 
 href="userhtmlsu7.html" ><span 
 class="cmr-12">next</span></a><span 
@@ -394,6 +398,6 @@ class="cmr-12">] [</span><a
 href="userhtmlse4.html#userhtmlsu6.html" ><span 
 class="cmr-12">up</span></a><span 
 class="cmr-12">] </span></p></div>
-<!--l. 420--><p class="indent" >   <a 
+<!--l. 424--><p class="indent" >   <a 
 id="tailuserhtmlsu6.html"></a>  
 </body></html> 
@@ -10,7 +10,7 @@
 <link rel="stylesheet" type="text/css" href="userhtml.css"> 
 </head><body 
 >
-   <!--l. 420--><div class="crosslinks"><p class="noindent"><span 
+   <!--l. 424--><div class="crosslinks"><p class="noindent"><span 
 class="cmr-12">[</span><a 
 href="userhtmlsu6.html" ><span 
 class="cmr-12">prev</span></a><span 
@@ -28,7 +28,7 @@ class="cmr-12">] </span></p></div>
 class="cmr-12">4.2   </span></span> <a 
 id="x16-150004.2"></a><span 
 class="cmr-12">GPU example</span></h4>
-<!--l. 422--><p class="noindent" ><span 
+<!--l. 426--><p class="noindent" ><span 
 class="cmr-12">The code discussed here shows how to set up a program exploiting the combined GPU</span>
 <span 
 class="cmr-12">capabilities of PSBLAS and AMG4PSBLAS. The code example is available in the</span>
@@ -36,14 +36,14 @@ class="cmr-12">capabilities of PSBLAS and AMG4PSBLAS. The code example is availa
 class="cmr-12">source distribution directory </span><span class="obeylines-h"><span class="verb"><span 
 class="cmtt-12">amg4psblas/examples/gpu</span></span></span><span 
 class="cmr-12">.</span>
-<!--l. 427--><p class="indent" >   <span 
+<!--l. 431--><p class="indent" >   <span 
 class="cmr-12">First of all, we need to include the appropriate modules and declare some auxiliary</span>
 <span 
 class="cmr-12">variables:</span>
                                                                               

                                                                               
-<!--l. 429--><p class="indent" >   <a 
+<!--l. 433--><p class="indent" >   <a 
 id="x16-15001r5"></a><hr class="float"><div class="float" 
 >
                                                                               
@@ -51,7 +51,7 @@ class="cmr-12">variables:</span>
                                                                               
 <div class="center" 
 >
-<!--l. 448--><p class="noindent" >
+<!--l. 452--><p class="noindent" >
 <div class="minipage"><pre class="verbatim" id="verbatim-10">
 program&#x00A0;amg_dexample_gpu
 &#x00A0;&#x00A0;use&#x00A0;psb_base_mod
@@ -69,7 +69,7 @@ program&#x00A0;amg_dexample_gpu
 &#x00A0;&#x00A0;type(psb_i_vect_gpu)&#x00A0;&#x00A0;&#x00A0;&#x00A0;&#x00A0;&#x00A0;&#x00A0;::&#x00A0;igmold
 &#x00A0;
 </pre>
-<!--l. 467--><p class="nopar" ></div></div>
+<!--l. 471--><p class="nopar" ></div></div>
 <br /> <div class="caption" 
 ><span class="id">Listing 5: </span><span  
 class="content">setup of a GPU-enabled test program part one.</span></div><!--tex4ht:label?: x16-15001r5 -->
@@ -77,7 +77,7 @@ class="content">setup of a GPU-enabled test program part one.</span></div><!--te

                                                                               
   </div><hr class="endfloat" />
-<!--l. 474--><p class="indent" >   <span 
+<!--l. 478--><p class="indent" >   <span 
 class="cmr-12">In this particular example we are choosing to employ a </span><span class="obeylines-h"><span class="verb"><span 
 class="cmtt-12">HLG</span></span></span> <span 
 class="cmr-12">data structure for</span>
@@ -85,7 +85,7 @@ class="cmr-12">data structure for</span>
 class="cmr-12">sparse matrices on GPUs; for more information please refer to the PSBLAS-EXT users&#8217;</span>
 <span 
 class="cmr-12">guide.</span>
-<!--l. 478--><p class="indent" >   <span 
+<!--l. 482--><p class="indent" >   <span 
 class="cmr-12">We then have to initialize the GPU environment, and pass the appropriate MOLD</span>
 <span 
 class="cmr-12">variables to the build methods (see also the PSBLAS and PSBLAS-EXT users&#8217;</span>
@@ -94,7 +94,7 @@ class="cmr-12">guides).</span>
                                                                               

                                                                               
-<!--l. 481--><p class="indent" >   <a 
+<!--l. 485--><p class="indent" >   <a 
 id="x16-15002r6"></a><hr class="float"><div class="float" 
 >
                                                                               
@@ -102,7 +102,7 @@ class="cmr-12">guides).</span>
                                                                               
 <div class="center" 
 >
-<!--l. 497--><p class="noindent" >
+<!--l. 501--><p class="noindent" >
 <div class="minipage"><pre class="verbatim" id="verbatim-11">
 &#x00A0;&#x00A0;call&#x00A0;psb_init(ctxt)
 &#x00A0;&#x00A0;call&#x00A0;psb_info(ctxt,iam,np)
@@ -116,7 +116,7 @@ class="cmr-12">guides).</span>
 &#x00A0;&#x00A0;call&#x00A0;prec%smoothers_build(a,desc_a,info,&#x00A0;amold=agmold,&#x00A0;vmold=vgmold,&#x00A0;imold=igmold)
 &#x00A0;
 </pre>
-<!--l. 512--><p class="nopar" ></div></div>
+<!--l. 516--><p class="nopar" ></div></div>
 <br /> <div class="caption" 
 ><span class="id">Listing 6: </span><span  
 class="content">setup of a GPU-enabled test program part two.</span></div><!--tex4ht:label?: x16-15002r6 -->
@@ -124,7 +124,7 @@ class="content">setup of a GPU-enabled test program part two.</span></div><!--te

                                                                               
   </div><hr class="endfloat" />
-<!--l. 519--><p class="indent" >   <span 
+<!--l. 523--><p class="indent" >   <span 
 class="cmr-12">Finally, we convert the input matrix, the descriptor and the vectors to use a</span>
 <span 
 class="cmr-12">GPU-enabled internal storage format. We then preallocate the preconditioner</span>
@@ -135,7 +135,7 @@ class="cmr-12">GPU environment</span>
                                                                               

                                                                               
-<!--l. 523--><p class="indent" >   <a 
+<!--l. 527--><p class="indent" >   <a 
 id="x16-15003r7"></a><hr class="float"><div class="float" 
 >
                                                                               
@@ -143,7 +143,7 @@ class="cmr-12">GPU environment</span>
                                                                               
 <div class="center" 
 >
-<!--l. 552--><p class="noindent" >
+<!--l. 557--><p class="noindent" >
 <div class="minipage"><pre class="verbatim" id="verbatim-12">
 &#x00A0;&#x00A0;call&#x00A0;desc_a%cnv(mold=igmold)
 &#x00A0;&#x00A0;call&#x00A0;a%cscnv(info,mold=agmold)
@@ -167,7 +167,7 @@ class="cmr-12">GPU environment</span>
 &#x00A0;&#x00A0;stop
 &#x00A0;
 </pre>
-<!--l. 579--><p class="nopar" ></div></div>
+<!--l. 584--><p class="nopar" ></div></div>
 <br /> <div class="caption" 
 ><span class="id">Listing 7: </span><span  
 class="content">setup of a GPU-enabled test program part three.</span></div><!--tex4ht:label?: x16-15003r7 -->
@@ -175,25 +175,30 @@ class="content">setup of a GPU-enabled test program part three.</span></div><!--

                                                                               
   </div><hr class="endfloat" />
-<!--l. 587--><p class="indent" >   <span 
-class="cmr-12">It is very important to employ solvers that are suited to the GPU, i.e. solvers that</span>
+<!--l. 592--><p class="indent" >   <span 
+class="cmr-12">It is very important to employ smoothers and coarsest solvers that are suited to the</span>
 <span 
-class="cmr-12">do NOT employ triangular system solve kernels. Solvers that satisfy this constraint</span>
+class="cmr-12">GPU, i.e. methods that do NOT employ triangular system solve kernels. Methods that</span>
 <span 
-class="cmr-12">include:</span>
+class="cmr-12">satisfy this constraint include:</span>
     <ul class="itemize1">
     <li class="itemize"><span class="obeylines-h"><span class="verb"><span 
 class="cmtt-12">JACOBI</span></span></span>
     </li>
     <li class="itemize"><span class="obeylines-h"><span class="verb"><span 
+class="cmtt-12">BJAC</span></span></span> <span 
+class="cmr-12">with the following methods on the local blocks:</span>
+          <ul class="itemize2">
+          <li class="itemize"><span class="obeylines-h"><span class="verb"><span 
 class="cmtt-12">INVK</span></span></span>
-     </li>
-     <li class="itemize"><span class="obeylines-h"><span class="verb"><span 
+          </li>
+          <li class="itemize"><span class="obeylines-h"><span class="verb"><span 
 class="cmtt-12">INVT</span></span></span>
-     </li>
-     <li class="itemize"><span class="obeylines-h"><span class="verb"><span 
+          </li>
+          <li class="itemize"><span class="obeylines-h"><span class="verb"><span 
 class="cmtt-12">AINV</span></span></span></li></ul>
-<!--l. 596--><p class="noindent" ><span 
+     </li></ul>
+<!--l. 604--><p class="noindent" ><span 
 class="cmr-12">and their </span><span 
 class="cmmi-12">&#x2113;</span><sub><span 
 class="cmr-8">1</span></sub> <span 
@@ -211,12 +211,12 @@ ps: $(PS)

 .PHONY: clean clean-tmp preview print gzip gunzip tar ci

-clean:
+clean: clean-tmp
 	rm -f $(PDF) $(PS)
-	cd tmp ; rm -r -f *
+	

 clean-tmp:
-	cd tmp ; rm -r -f *
+	(test -d tmp && cd tmp && rm -r -f *)

 #----------------------------------------------------------------------------

@@ -190,22 +190,23 @@ make install
 \subsection{Bug reporting}
 If you find any bugs in our codes, please report them through our
 issues page on \\[2mm]
-\url{https://github.com/psctoolkit/amg4psblas/issues}\\
+\url{https://github.com/psctoolkit/psctoolkit/issues}\\

 To enable us to track the bug, please provide a log from the failing
 application, the test conditions, and ideally a self-contained test
 program reproducing the issue.

 \subsection{Example and test programs\label{sec:ex_and_test}}
-The package contains the \verb|examples| and \verb|tests| directories;
+The package contains a \verb|samples| directory, divided in two
+subdirs  \verb|simple| and \verb|advanced|;
 both of them are further divided into \verb|fileread| and
 \verb|pdegen| subdirectories. Their purpose is as follows:
 \begin{description}
-\item[\tt examples] contains a set of simple example programs with a
+\item[\tt simple] contains a set of simple example programs with a
  predefined choice of preconditioners, selectable via integer
  values. These are intended to get  acquainted with the
  multilevel preconditioners available in AMG4PSBLAS.
-\item[\tt tests] contains a set of more sophisticated examples that
+\item[\tt advanced] contains a set of more sophisticated examples that
  will allow the user, via the input files in the \verb|runs|
  subdirectories, to experiment with the full range of preconditioners
  implemented in the package.
@@ -159,4 +159,4 @@ Some influential environment variables:
 Use these variables to override the choices made by `configure' or to help
 it to find libraries and programs with nonstandard names/locations.

-Report bugs to <https://github.com/sfilippone/amg4psblas/issues>.
+Report bugs to <https://github.com/psctoolkit/psctoolkit/issues>.
@@ -129,9 +129,9 @@ relevant data structures, performed
 through the PSBLAS routines for sparse matrix and vector management, is not reported
 here for the sake of conciseness.
 The complete code can be found in the example program file \verb|amg_dexample_ml.f90|,
-in the directory \verb|examples/fileread| of the AMG4PSBLAS implementation (see
+in the directory \verb|samples/simple/file|\-\verb|read| of the AMG4PSBLAS implementation (see
 Section~\ref{sec:ex_and_test}). A sample test problem along with the relevant
-input data is available in \verb|examples/fileread/runs|.
+input data is available in \verb|samples/simple/fileread/runs|.
 For details on the use of the PSBLAS routines, see the PSBLAS User's
 Guide~\cite{PSBLASGUIDE}.

@@ -139,7 +139,7 @@ The setup and application of the default multilevel preconditioner
 for the real single precision and the complex, single and double
 precision, versions are obtained with straightforward modifications of the previous
 example (see Section~\ref{sec:userinterface} for details). If these versions are installed,
-the corresponding codes are available in \verb|examples/fileread/|.
+the corresponding codes are available in \verb|samples/simple/file|\-\verb|read|.

 \begin{listing}[tbp]
 \begin{center}
@@ -174,7 +174,7 @@ the corresponding codes are available in \verb|examples/fileread/|.
 ! with basic smoothed aggregation, 1 hybrid forward/backward
 ! GS sweep as pre/post-smoother and UMFPACK as coarsest-level
 ! solver
-  call P%init('ML',info)
+  call P%init(ctxt,'ML',info)
 !
 ! build the preconditioner
  call P%hierarchy_build(A,desc_A,info)
@@ -184,8 +184,8 @@ the corresponding codes are available in \verb|examples/fileread/|.
 ! set the solver parameters and the initial guess
 ... ...
 !
-! solve Ax=b with preconditioned CG
-  call psb_krylov('CG',A,P,b,x,tol,desc_A,info)
+! solve Ax=b with preconditioned FCG
+  call psb_krylov('FCG',A,P,b,x,tol,desc_A,info)
 ... ...
 !
 ! deallocate the preconditioner
@@ -229,7 +229,7 @@ stop
 ! with basic smoothed aggregation, 1 hybrid forward/backward
 ! GS sweep as pre/post-smoother and UMFPACK as coarsest-level
 ! solver
-  call P%init('ML',info)
+  call P%init(ctxt,'ML',info)
 !
 ! build the preconditioner
  call P%hierarchy_build(A,desc_A,info)
@@ -239,8 +239,8 @@ stop
 ! set the solver parameters and the initial guess
  ... ...
 !
-! solve Ax=b with preconditioned CG
-  call psb_krylov('CG',A,P,b,x,tol,desc_A,info)
+! solve Ax=b with preconditioned FCG
+  call psb_krylov('FCG',A,P,b,x,tol,desc_A,info)
  ... ...
 !
 ! deallocate the preconditioner
@@ -300,14 +300,14 @@ The corresponding example program is available in the file

 For all the previous preconditioners, example programs where the sparse matrix and
 the right-hand side are generated by discretizing a PDE with Dirichlet
-boundary conditions are also available in the directory \verb|examples/pdegen|.
+boundary conditions are also available in the directory \verb|samples/simple/pdegen|.
 \vspace{-1em}\begin{listing}[tbh]
 \ifpdf%
 \begin{minted}[breaklines=true,bgcolor=bg,fontsize=\small]{fortran}
 ! build a V-cycle preconditioner with 1 block-Jacobi sweep (with
 ! ILU(0) on the blocks) as pre- and post-smoother, and 8  block-Jacobi
 ! sweeps (with ILU(0) on the blocks) as coarsest-level solver
-call P%init('ML',info)
+call P%init(ctxt,'ML',info)
 call P%set('SMOOTHER_TYPE','BJAC',info)
 call P%set('COARSE_SOLVE','BJAC',info)
 call P%set('COARSE_SWEEPS',8,info)
@@ -323,7 +323,7 @@ call P%smoothers_build(A,desc_A,info)
 ! build a V-cycle preconditioner with 1 block-Jacobi sweep (with
 ! ILU(0) on the blocks) as pre- and post-smoother, and 8  block-Jacobi
 ! sweeps (with ILU(0) on the blocks) as coarsest-level solver
-  call P%init('ML',info)
+  call P%init(ctxt,'ML',info)
  call P%set('SMOOTHER_TYPE','BJAC',info)
  call P%set('COARSE_SOLVE','BJAC',info)
  call P%set('COARSE_SWEEPS',8,info)
@@ -345,7 +345,7 @@ call P%smoothers_build(A,desc_A,info)
 !2 hybrid Gauss-Seidel sweeps as pre- and post-smoother,
 !and parallel flexible Conjugate Gradient coupled with the block-Jacobi 
 !preconditioner having ILU(0) on the blocks as coarsest solver. 
-call P%init('ML',info)
+call P%init(ctxt,'ML',info)
 call P%set('PAR_AGGR_ALG','COUPLED',info)
 call P%set('AGGR_TYPE','MATCHBOXP',info)
 call P%set('AGGR_SIZE',8,info)
@@ -353,6 +353,8 @@ call P%set('ML_CYCLE','WCYCLE',info)
 call P%set('SMOOTHER_TYPE','FBGS',info)
 call P%set('SMOOTHER_SWEEPS',2,info)
 call P%set('COARSE_SOLVE','KRM',info)
+call P%set('COARSE_MAT','DIST',info)
+call P%set('KRM_METHOD','FCG',info)
 call P%hierarchy_build(A,desc_A,info)
 call P%smoothers_build(A,desc_A,info)
 \end{minted}
@@ -365,14 +367,16 @@ call P%smoothers_build(A,desc_A,info)
 ! build a W-cycle preconditioner with 2 hybrid Gauss-Seidel sweeps
 ! as pre- and post-smoother, a distributed coarsest
 ! matrix, and MUMPS as coarsest-level solver
-  call P%init('ML',info)
+  call P%init(ctxt,'ML',info)
  call P%set('PAR_AGGR_ALG','COUPLED',info)
-call P%set('AGGR_TYPE','MATCHBOXP',info)
-call P%set('AGGR_SIZE',8,info)
+  call P%set('AGGR_TYPE','MATCHBOXP',info)
+  call P%set('AGGR_SIZE',8,info)
  call P%set('ML_CYCLE','WCYCLE',info)
  call P%set('SMOOTHER_TYPE','FBGS',info)
  call P%set('SMOOTHER_SWEEPS',2,info)
  call P%set('COARSE_SOLVE','KRM',info)
+  call P%set('COARSE_MAT','DIST',info)
+  call P%set('KRM_METHOD','FCG',info)
  call P%hierarchy_build(A,desc_A,info)
  call P%smoothers_build(A,desc_A,info)
 ... ...
@@ -387,7 +391,7 @@ call P%set('AGGR_SIZE',8,info)
 \ifpdf
 \begin{minted}[breaklines=true,bgcolor=bg,fontsize=\small]{fortran}
 ! build a one-level RAS with overlap 2 and ILU(0) on the local blocks.
-call P%init('AS',info)
+call P%init(ctxt,'AS',info)
 call P%set('SUB_OVR',2,info)
 call P%build(A,desc_A,info)
 ... ...
@@ -401,7 +405,7 @@ call P%build(A,desc_A,info)
 \begin{verbatim}
 ... ...
 ! set RAS with overlap 2 and ILU(0) on the local blocks
-  call P%init('AS',info)
+  call P%init(ctxt,'AS',info)
  call P%set('SUB_OVR',2,info)
  call P%bld(A,desc_A,info)
 ... ...
@@ -535,7 +539,8 @@ Krylov method. At the end of the code, we close the GPU environment
  call prec%allocate_wrk(info)
  t1 = psb_wtime()
  call psb_krylov(s_choice%kmethd,a,prec,b,x,s_choice%eps,&
-       & desc_a,info,itmax=s_choice%itmax,iter=iter,err=err,itrace=s_choice%itrace,&
+       & desc_a,info,itmax=s_choice%itmax,iter=iter,err=err,&
+       & itrace=s_choice%itrace,&
       & istop=s_choice%istopc,irst=s_choice%irst)
  call prec%deallocate_wrk(info)
  call psb_barrier(ctxt)
@@ -584,15 +589,18 @@ Krylov method. At the end of the code, we close the GPU environment
 \caption{setup of a GPU-enabled test program part three.\label{fig:gpu-ex3}}
 \end{listing}

-It is very important to employ solvers that are suited
-to the GPU, i.e.  solvers that do NOT  employ triangular
-system solve kernels.  Solvers that satisfy this constraint include:
+It is very important to employ smoothers and coarsest solvers that are suited
+to the GPU, i.e.  methods that do NOT employ triangular
+system solve kernels. Methods that satisfy this constraint include:
 \begin{itemize}
 \item \verb|JACOBI|
+\item \verb|BJAC| with the following methods on the local blocks:
+\begin{itemize}
 \item \verb|INVK|
 \item \verb|INVT|
 \item \verb|AINV|
 \end{itemize}
+\end{itemize}
 and their $\ell_1$ variants. 

 %%% Local Variables:
@@ -176,9 +176,10 @@ Similar considerations apply to the point-Jacobi, Gauss-Seidel and block-Jacobi
 coarsest-level solvers, and shortcuts are available
 in this case too (see Table~\ref{tab:p_coarse_1}). \\

-\textbf{Remark 3.} Many of the coarsest-level solvers cannot be used
-with both replicated and distributed coarsest-matrix layouts; 
-therefore, setting the solver after the layout may change the layout.
+\textbf{Remark 3.} Many of the coarsest-level solvers apply to a
+specific coarsest-matrix layout; 
+therefore, setting the solver after the layout may change the layout
+to either distributed or replicated.
 Similarly, setting the layout after the solver may change the solver.

 More precisely, UMFPACK and SuperLU require the coarsest-level
@@ -399,8 +400,14 @@ the parameter \texttt{ilev}.} \\
 \fortinline|'COARSE_SOLVE'| & \fortinline|character(len=*)|
                          & \fortinline|'MUMPS'| \par \fortinline|'UMF'| \par
                           \fortinline|'SLU'| \par \fortinline|'SLUDIST'| \par
-			   \fortinline|'JACOBI'| \par \fortinline|'GS'| \par \fortinline|'BJAC'| \par \fortinline|'KRM'| 
-			   \fortinline|'L1-JACOBI'| \par \fortinline|'L1-BJAC'| \par \fortinline|'L1-FBGS'|
+                            \fortinline|'ILU'| \par
+                            \fortinline|'JACOBI'| \par
+                            \fortinline|'GS'| \par
+                            \fortinline|'BJAC'| \par
+                            \fortinline|'KRM'|  \par
+			   \fortinline|'L1-JACOBI'| \par
+                            \fortinline|'L1-BJAC'| \par
+                            \fortinline|'L1-FBGS'| 
                         & See~Note.
                         & Solver used at the coarsest level: sequential
                           LU from MUMPS, UMFPACK, or SuperLU
@@ -39,23 +39,18 @@
 !
 ! This sample program solves a linear system obtained by discretizing a
 ! PDE with Dirichlet BCs. The solver is CG, coupled with one of the
-! following multi-level preconditioner, as explained in Section 4.1 of
+! following multi-level preconditioner, as explained in Section 4.2 of
 ! the AMG4PSBLAS User's and Reference Guide:
 !
-! - choice = 1, the default multi-level preconditioner solver, i.e., 
-! V-cycle with decoupled smoothed aggregation, 1 hybrid forward/backward
-! GS sweep as pre/post-smoother and UMFPACK as coarsest-level
-! solver (Sec. 4.1, Listing 1)
+! - choice = 1, a V-cycle with decoupled smoothed aggregation, 4 Jacobi
+! sweeps as pre/post-smoother and 8 Jacobi sweeps as coarsest-level
+! solver with replicated coarsest matrix
 !
-! - choice = 2, a V-cycle preconditioner with 1 block-Jacobi sweep
-! (with ILU(0) on the blocks) as pre- and post-smoother, and 8 block-Jacobi
-! sweeps (with ILU(0) on the blocks) as coarsest-level solver (Sec. 4.1, Listing 2)
-!
-! - choice = 3,  W-cycle preconditioner based on the coupled aggregation relying
-! on matching, with maximum size of aggregates equal to 8 and smoothed prolongators,
-! 2 hybrid forward/backward GS sweeps as pre/post-smoother, a distributed coarsest
-! matrix, and preconditioned Flexible Conjugate Gradient as coarsest-level solver
-! (Sec. 4.1, Listing 3)
+! - choice = 2, a W-cycle based on the coupled aggregation relying on matching,
+! with maximum size of aggregates equal to 8 and smoothed prolongators,
+! 2 sweeps of Block-Jacobi ipre/post-smoother using approximate inverse INVK and
+! 4 sweeps of Block-Jacobi with INVK as coarsest-level solver on distributed
+! coarsest matrix 
 !
 ! The matrix and the rhs are read from files (if an rhs is not available, the
 ! unit rhs is set).
@@ -183,8 +178,9 @@ program amg_dexample_gpu

 case(1)

-    ! initialize a V-cycle preconditioner with 4 Jacobi sweep 
-    ! and 8 Jacobi sweeps as coarsest-level solver
+    ! initialize a V-cycle preconditioner, relying on decoupled smoothed aggregation 
+    ! with 4 Jacobi sweeps as pre/post-smoother
+    ! and 8 Jacobi sweeps as coarsest-level solver on replicated coarsest matrix

    call P%init(ctxt,'ML',info)
    call P%set('SMOOTHER_TYPE','JACOBI',info)
@@ -195,19 +191,22 @@ program amg_dexample_gpu

  case(2)

-   ! initialize a V-cycle preconditioner based on the coupled aggregation relying on matching,
+   ! initialize a W-cycle preconditioner based on the coupled aggregation relying on matching,
   ! with maximum size of aggregates equal to 8 and smoothed prolongators,
-   ! Block-Jacobi smoother using approximate inverse INVK and
-   ! and 4 sweeps of INVK on he coarsest level 
+   ! 2 sweeps of Block-Jacobi pre/post-smoother using approximate inverse INVK and
+   ! 4 sweeps of Block-Jacobi with INVK on the coarsest level distributed matrix

    call P%init(ctxt,'ML',info)
    call P%set('PAR_AGGR_ALG','COUPLED',info)
    call P%set('AGGR_TYPE','MATCHBOXP',info)
    call P%set('AGGR_SIZE',8,info)
    call P%set('ML_CYCLE','WCYCLE',info)
+    call P%set('SMOOTHER_TYPE','BJAC',info)
    call P%set('SMOOTHER_SWEEPS',2,info)
    call P%set('SUB_SOLVE','INVK',info)  
-    call P%set('COARSE_SOLVE','INVK',info)
+    call P%set('COARSE_SOLVE','BJAC',info)
+    call P%set('COARSE_SUBSOLVE','INVK',info)
+    call P%set('COARSE_SWEEPS',4,info)
    call P%set('COARSE_MAT','DIST',info)
    kmethod = 'CG'

@@ -676,8 +676,9 @@ contains
      call read_data(prec%aggr_prol,inp_unit)    ! aggregation type
      call read_data(prec%par_aggr_alg,inp_unit)    ! parallel aggregation alg
      call read_data(prec%aggr_ord,inp_unit)    ! ordering for aggregation
-      call read_data(prec%aggr_filter,inp_unit) ! filtering
      call read_data(prec%mncrratio,inp_unit)  ! minimum aggregation ratio
+      call read_data(prec%aggr_filter,inp_unit) ! filtering
+      call read_data(prec%athres,inp_unit)      ! smoothed aggr thresh
      call read_data(prec%thrvsz,inp_unit)      ! size of aggr thresh vector
      if (prec%thrvsz > 0) then
        call psb_realloc(prec%thrvsz,prec%athresv,info)
@@ -685,7 +686,6 @@ contains
      else
        read(inp_unit,*)                        ! dummy read to skip a record
      end if
-      call read_data(prec%athres,inp_unit)      ! smoothed aggr thresh
      ! coasest-level solver
      call read_data(prec%csolve,inp_unit)      ! coarsest-lev solver
      call read_data(prec%csbsolve,inp_unit)    ! coarsest-lev subsolver
@@ -676,8 +676,9 @@ contains
      call read_data(prec%aggr_prol,inp_unit)    ! aggregation type
      call read_data(prec%par_aggr_alg,inp_unit)    ! parallel aggregation alg
      call read_data(prec%aggr_ord,inp_unit)    ! ordering for aggregation
-      call read_data(prec%aggr_filter,inp_unit) ! filtering
      call read_data(prec%mncrratio,inp_unit)  ! minimum aggregation ratio
+      call read_data(prec%aggr_filter,inp_unit) ! filtering
+      call read_data(prec%athres,inp_unit)      ! smoothed aggr thresh
      call read_data(prec%thrvsz,inp_unit)      ! size of aggr thresh vector
      if (prec%thrvsz > 0) then
        call psb_realloc(prec%thrvsz,prec%athresv,info)
@@ -685,7 +686,6 @@ contains
      else
        read(inp_unit,*)                        ! dummy read to skip a record
      end if
-      call read_data(prec%athres,inp_unit)      ! smoothed aggr thresh
      ! coasest-level solver
      call read_data(prec%csolve,inp_unit)      ! coarsest-lev solver
      call read_data(prec%csbsolve,inp_unit)    ! coarsest-lev subsolver
@@ -676,8 +676,9 @@ contains
      call read_data(prec%aggr_prol,inp_unit)    ! aggregation type
      call read_data(prec%par_aggr_alg,inp_unit)    ! parallel aggregation alg
      call read_data(prec%aggr_ord,inp_unit)    ! ordering for aggregation
-      call read_data(prec%aggr_filter,inp_unit) ! filtering
      call read_data(prec%mncrratio,inp_unit)  ! minimum aggregation ratio
+      call read_data(prec%aggr_filter,inp_unit) ! filtering
+      call read_data(prec%athres,inp_unit)      ! smoothed aggr thresh
      call read_data(prec%thrvsz,inp_unit)      ! size of aggr thresh vector
      if (prec%thrvsz > 0) then
        call psb_realloc(prec%thrvsz,prec%athresv,info)
@@ -685,7 +686,6 @@ contains
      else
        read(inp_unit,*)                        ! dummy read to skip a record
      end if
-      call read_data(prec%athres,inp_unit)      ! smoothed aggr thresh
      ! coasest-level solver
      call read_data(prec%csolve,inp_unit)      ! coarsest-lev solver
      call read_data(prec%csbsolve,inp_unit)    ! coarsest-lev subsolver
@@ -676,8 +676,9 @@ contains
      call read_data(prec%aggr_prol,inp_unit)    ! aggregation type
      call read_data(prec%par_aggr_alg,inp_unit)    ! parallel aggregation alg
      call read_data(prec%aggr_ord,inp_unit)    ! ordering for aggregation
-      call read_data(prec%aggr_filter,inp_unit) ! filtering
      call read_data(prec%mncrratio,inp_unit)  ! minimum aggregation ratio
+      call read_data(prec%aggr_filter,inp_unit) ! filtering
+      call read_data(prec%athres,inp_unit)      ! smoothed aggr thresh
      call read_data(prec%thrvsz,inp_unit)      ! size of aggr thresh vector
      if (prec%thrvsz > 0) then
        call psb_realloc(prec%thrvsz,prec%athresv,info)
@@ -685,7 +686,6 @@ contains
      else
        read(inp_unit,*)                        ! dummy read to skip a record
      end if
-      call read_data(prec%athres,inp_unit)      ! smoothed aggr thresh
      ! coasest-level solver
      call read_data(prec%csolve,inp_unit)      ! coarsest-lev solver
      call read_data(prec%csbsolve,inp_unit)    ! coarsest-lev subsolver
@@ -1,14 +1,14 @@
 !   
 !   
-!                             MLD2P4  version 2.2
-!    MultiLevel Domain Decomposition Parallel Preconditioners Package
-!               based on PSBLAS (Parallel Sparse BLAS version 3.5)
+!                             AMG4PSBLAS version 1.0
+!    Algebraic Multigrid Package
+!               based on PSBLAS (Parallel Sparse BLAS version 3.7)
 !    
-!    (C) Copyright 2008-2018 
+!    (C) Copyright 2021 
 !  
 !        Salvatore Filippone  
 !        Pasqua D'Ambra   
-!        Daniela di Serafino   
+!        Fabio Durastante        
 !   
 !    Redistribution and use in source and binary forms, with or without
 !    modification, are permitted provided that the following conditions
@@ -18,14 +18,14 @@
 !      2. Redistributions in binary form must reproduce the above copyright
 !         notice, this list of conditions, and the following disclaimer in the
 !         documentation and/or other materials provided with the distribution.
-!      3. The name of the MLD2P4 group or the names of its contributors may
+!      3. The name of the AMG4PSBLAS group or the names of its contributors may
 !         not be used to endorse or promote products derived from this
 !         software without specific written permission.
 !   
 !    THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
 !    ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
 !    TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
-!    PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE MLD2P4 GROUP OR ITS CONTRIBUTORS
+!    PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE AMG4PSBLAS GROUP OR ITS CONTRIBUTORS
 !    BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
 !    CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
 !    SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
--- a/Show More
+++ b/Show More
Author	SHA1	Message	Date
Salvatore Filippone	ac42d7b1dd	Sync configure with configure_n	2022-04-14 20:33:02 +02:00
Salvatore Filippone	697f325df6	Fix use of SuperLU_Dist, configure checks and ifdefs	2022-04-13 16:36:32 +02:00
Salvatore Filippone	58d00b16c6	Add message to configure	2022-04-07 10:36:28 +02:00
Salvatore Filippone	425743939c	Fix for new SuperLU_Dist version, change configure	2022-04-06 11:02:32 +02:00
Salvatore Filippone	7e48a0a742	Fix defines for SLUD v7	2022-04-06 09:02:20 +02:00
Salvatore Filippone	4f9254ebb0	Add log entry for configure check on MPICXX libs	2022-03-18 18:31:42 +01:00
Salvatore Filippone	90657b706f	Fix Makefile: mld -> amg	2022-03-15 11:00:52 +01:00
Salvatore Filippone	23a39a6c54	Fix configry redirect check to dev/null	2022-03-15 11:00:22 +01:00
Salvatore Filippone	873f190961	Configry checks for OpenMPI cxx libs	2022-03-15 10:25:58 +01:00
Salvatore Filippone	87cdd76f8d	Fix spurious error notification with prec%descr	2022-02-07 10:41:45 +01:00
Salvatore Filippone	45fabb5214	Move reading aggregation ratio above filtering option.	2021-10-25 14:46:43 +02:00
Salvatore Filippone	a9182021bb	Sample programs adapted for position of ATHRES in control files.	2021-10-25 14:13:25 +02:00
Salvatore Filippone	a8f4009cb1	Take out spurious csize and maxnlev from parmatch aggregator object.	2021-10-22 09:39:18 -04:00
Salvatore Filippone	794080e386	Fix target coarse size handling.	2021-10-22 08:14:26 -04:00
Salvatore Filippone	818f7a78a0	Do not call %default on setting coarse_solve	2021-10-18 14:47:28 +02:00
Salvatore Filippone	939d7c9a89	Do not invoke default() after setting KRM for coarse solver.	2021-10-16 08:37:25 +02:00
Salvatore Filippone	92f7cde375	Make test program to dump preconditioner controlled from input file.	2021-09-27 10:38:49 -04:00
Salvatore Filippone	af178daa84	Modify dump method to print base level matrix.	2021-09-27 14:50:46 +02:00
Salvatore Filippone	49777a379b	Cosmetic change in sample source code	2021-09-27 14:22:34 +02:00
Salvatore Filippone	5768238f66	Typographical fixes.	2021-09-22 05:02:12 -04:00
Salvatore Filippone	4c4b2b282e	Merge branch 'development' of github.com:sfilippone/amg4psblas into development	2021-09-22 04:59:54 -04:00
Salvatore Filippone	9d11a99ed4	Fix settings in samples/PDEGEN	2021-09-22 04:58:26 -04:00
Salvatore Filippone	9bc8b540b3	Fix settings in samples/PDEGEN	2021-08-27 11:42:01 +02:00
Salvatore Filippone	af75364c54	Fix matchbox internal interface names.	2021-07-16 09:16:54 +02:00
Salvatore Filippone	1270498170	Fix examples	2021-07-16 09:16:44 +02:00
Salvatore Filippone	b387308455	Merge branch 'maint-1.0' into development	2021-07-15 12:09:54 +02:00
Salvatore Filippone	aba9b29717	Fix samples/simple internal and external docs.	2021-07-15 11:59:34 +02:00
Salvatore Filippone	94ca610bff	Do not print matching statistics	2021-06-28 18:47:35 +02:00
Salvatore Filippone	2542c0fda4	Do not print matching statistics	2021-06-28 18:42:56 +02:00
Salvatore Filippone	8482067b52	Deactivate MINNRG	2021-06-28 18:42:34 +02:00
Salvatore Filippone	7319dab30f	Deactivate MINNRG	2021-06-21 21:44:09 +02:00
Salvatore Filippone	4bbba3ebd7	Fix interface inconsistencies	2021-06-21 21:38:14 +02:00
Salvatore Filippone	988021ff24	Fix uninitialized warning	2021-06-15 03:33:13 -04:00
Salvatore Filippone	4e177ce926	Merge branch 'development' of github.com:sfilippone/amg4psblas into development	2021-06-14 12:27:38 -04:00
Salvatore Filippone	1fa94d0372	Fix AS%FREE()	2021-06-14 12:26:04 -04:00
Cirdans-Home	0fcbdd74cd	Fixed typo	2021-06-08 17:50:21 +02:00
Cirdans-Home	ba854379e4	Fixed typo	2021-06-08 17:48:19 +02:00
pasquadambra	a6cbd64e65	update	2021-06-08 17:18:02 +02:00
Cirdans-Home	5c589dbf30	Fixed TeX typos and href for issues	2021-06-08 16:15:32 +02:00
pasquadambra	10e9c53e54	updating examples/gpu and doc	2021-06-08 15:15:44 +02:00
Salvatore Filippone	0332920a63	Merge branch 'development' into maint-1.0	2021-05-13 11:39:02 +02:00
Salvatore Filippone	9b9dfbd198	Fix copyright	2021-05-13 11:38:53 +02:00
Salvatore Filippone	5909e541b0	Merge branch 'development' into maint-1.0	2021-05-13 11:36:43 +02:00
Salvatore Filippone	941ca6568a	Fix docs for new samples	2021-05-13 11:36:28 +02:00
Salvatore Filippone	39a9c4e4ed	Fix copyright	2021-05-12 21:34:59 +02:00
Salvatore Filippone	41b4373494	Merge branch 'development' into maint-1.0	2021-05-12 21:30:11 +02:00
Salvatore Filippone	4bf009a1ab	Fix docs for samples	2021-05-12 21:29:36 +02:00