diff --git a/base/serial/impl/psb_c_coo_impl.F90 b/base/serial/impl/psb_c_coo_impl.F90
index 1b162457..53ffcea9 100644
--- a/base/serial/impl/psb_c_coo_impl.F90
+++ b/base/serial/impl/psb_c_coo_impl.F90
@@ -2879,9 +2879,9 @@ subroutine psb_c_coo_csput_a(nz,ia,ja,val,a,imin,imax,jmin,jmax,info)
     if (info /= 0) write(0,*) name,' point 0:',info,isza,nza,nz
     ! Build phase. Must handle reallocations in a sensible way.
     if (isza < (nza+nz)) then
-      write(0,*) ' before reallocate in csput ',psb_errstatus_fatal(),info
+      !write(0,*) ' before reallocate in csput ',psb_errstatus_fatal(),info
       call a%reallocate(max(nza+nz,int(1.5*isza)))
-      write(0,*) '  after reallocate in csput ',psb_errstatus_fatal(),info            
+      !write(0,*) '  after reallocate in csput ',psb_errstatus_fatal(),info            
     endif
     isza = a%get_size()
     if (isza < (nza+nz)) then
diff --git a/base/serial/impl/psb_c_csr_impl.F90 b/base/serial/impl/psb_c_csr_impl.F90
index 5a795583..0e040a7b 100644
--- a/base/serial/impl/psb_c_csr_impl.F90
+++ b/base/serial/impl/psb_c_csr_impl.F90
@@ -152,7 +152,7 @@ contains
           !$omp parallel do private(i,j, acc) schedule(static)
           do i=1,m
             acc  = czero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -164,7 +164,7 @@ contains
           !$omp parallel do private(i,j, acc)
           do i=1,m
             acc  = czero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -176,7 +176,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = czero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -192,7 +192,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = czero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -204,7 +204,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = czero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -216,7 +216,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = czero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -231,7 +231,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = czero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -243,7 +243,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = czero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -255,7 +255,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = czero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -270,7 +270,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = czero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -282,7 +282,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = czero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -294,7 +294,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = czero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
diff --git a/base/serial/impl/psb_d_coo_impl.F90 b/base/serial/impl/psb_d_coo_impl.F90
index 9a20164a..7a79847e 100644
--- a/base/serial/impl/psb_d_coo_impl.F90
+++ b/base/serial/impl/psb_d_coo_impl.F90
@@ -2879,9 +2879,9 @@ subroutine psb_d_coo_csput_a(nz,ia,ja,val,a,imin,imax,jmin,jmax,info)
     if (info /= 0) write(0,*) name,' point 0:',info,isza,nza,nz
     ! Build phase. Must handle reallocations in a sensible way.
     if (isza < (nza+nz)) then
-      write(0,*) ' before reallocate in csput ',psb_errstatus_fatal(),info
+      !write(0,*) ' before reallocate in csput ',psb_errstatus_fatal(),info
       call a%reallocate(max(nza+nz,int(1.5*isza)))
-      write(0,*) '  after reallocate in csput ',psb_errstatus_fatal(),info            
+      !write(0,*) '  after reallocate in csput ',psb_errstatus_fatal(),info            
     endif
     isza = a%get_size()
     if (isza < (nza+nz)) then
diff --git a/base/serial/impl/psb_d_csr_impl.F90 b/base/serial/impl/psb_d_csr_impl.F90
index af842c16..d6b6e9cb 100644
--- a/base/serial/impl/psb_d_csr_impl.F90
+++ b/base/serial/impl/psb_d_csr_impl.F90
@@ -152,7 +152,7 @@ contains
           !$omp parallel do private(i,j, acc) schedule(static)
           do i=1,m
             acc  = dzero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -164,7 +164,7 @@ contains
           !$omp parallel do private(i,j, acc)
           do i=1,m
             acc  = dzero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -176,7 +176,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = dzero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -192,7 +192,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = dzero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -204,7 +204,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = dzero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -216,7 +216,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = dzero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -231,7 +231,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = dzero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -243,7 +243,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = dzero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -255,7 +255,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = dzero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -270,7 +270,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = dzero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -282,7 +282,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = dzero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -294,7 +294,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = dzero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
diff --git a/base/serial/impl/psb_s_coo_impl.F90 b/base/serial/impl/psb_s_coo_impl.F90
index e944ce7c..b4e6a8df 100644
--- a/base/serial/impl/psb_s_coo_impl.F90
+++ b/base/serial/impl/psb_s_coo_impl.F90
@@ -2879,9 +2879,9 @@ subroutine psb_s_coo_csput_a(nz,ia,ja,val,a,imin,imax,jmin,jmax,info)
     if (info /= 0) write(0,*) name,' point 0:',info,isza,nza,nz
     ! Build phase. Must handle reallocations in a sensible way.
     if (isza < (nza+nz)) then
-      write(0,*) ' before reallocate in csput ',psb_errstatus_fatal(),info
+      !write(0,*) ' before reallocate in csput ',psb_errstatus_fatal(),info
       call a%reallocate(max(nza+nz,int(1.5*isza)))
-      write(0,*) '  after reallocate in csput ',psb_errstatus_fatal(),info            
+      !write(0,*) '  after reallocate in csput ',psb_errstatus_fatal(),info            
     endif
     isza = a%get_size()
     if (isza < (nza+nz)) then
diff --git a/base/serial/impl/psb_s_csr_impl.F90 b/base/serial/impl/psb_s_csr_impl.F90
index 01657071..49d168bb 100644
--- a/base/serial/impl/psb_s_csr_impl.F90
+++ b/base/serial/impl/psb_s_csr_impl.F90
@@ -152,7 +152,7 @@ contains
           !$omp parallel do private(i,j, acc) schedule(static)
           do i=1,m
             acc  = szero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -164,7 +164,7 @@ contains
           !$omp parallel do private(i,j, acc)
           do i=1,m
             acc  = szero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -176,7 +176,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = szero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -192,7 +192,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = szero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -204,7 +204,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = szero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -216,7 +216,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = szero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -231,7 +231,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = szero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -243,7 +243,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = szero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -255,7 +255,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = szero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -270,7 +270,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = szero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -282,7 +282,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = szero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -294,7 +294,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = szero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
diff --git a/base/serial/impl/psb_z_coo_impl.F90 b/base/serial/impl/psb_z_coo_impl.F90
index f88dc92b..949d0384 100644
--- a/base/serial/impl/psb_z_coo_impl.F90
+++ b/base/serial/impl/psb_z_coo_impl.F90
@@ -2879,9 +2879,9 @@ subroutine psb_z_coo_csput_a(nz,ia,ja,val,a,imin,imax,jmin,jmax,info)
     if (info /= 0) write(0,*) name,' point 0:',info,isza,nza,nz
     ! Build phase. Must handle reallocations in a sensible way.
     if (isza < (nza+nz)) then
-      write(0,*) ' before reallocate in csput ',psb_errstatus_fatal(),info
+      !write(0,*) ' before reallocate in csput ',psb_errstatus_fatal(),info
       call a%reallocate(max(nza+nz,int(1.5*isza)))
-      write(0,*) '  after reallocate in csput ',psb_errstatus_fatal(),info            
+      !write(0,*) '  after reallocate in csput ',psb_errstatus_fatal(),info            
     endif
     isza = a%get_size()
     if (isza < (nza+nz)) then
diff --git a/base/serial/impl/psb_z_csr_impl.F90 b/base/serial/impl/psb_z_csr_impl.F90
index 947c8329..1fc387d8 100644
--- a/base/serial/impl/psb_z_csr_impl.F90
+++ b/base/serial/impl/psb_z_csr_impl.F90
@@ -152,7 +152,7 @@ contains
           !$omp parallel do private(i,j, acc) schedule(static)
           do i=1,m
             acc  = zzero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -164,7 +164,7 @@ contains
           !$omp parallel do private(i,j, acc)
           do i=1,m
             acc  = zzero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -176,7 +176,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = zzero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -192,7 +192,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = zzero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -204,7 +204,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = zzero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -216,7 +216,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = zzero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -231,7 +231,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = zzero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -243,7 +243,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = zzero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -255,7 +255,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = zzero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -270,7 +270,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = zzero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -282,7 +282,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = zzero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
@@ -294,7 +294,7 @@ contains
           !$omp parallel do private(i,j,acc)
           do i=1,m
             acc  = zzero
-            !$omp  simd
+            !$omp  simd reduction(+:acc)
             do j=irp(i), irp(i+1)-1
               acc  = acc + val(j) * x(ja(j))
             enddo
diff --git a/cuda/Makefile b/cuda/Makefile
index 1a595bc7..b47e58ce 100755
--- a/cuda/Makefile
+++ b/cuda/Makefile
@@ -51,6 +51,7 @@ lib: objs ilib cudalib spgpulib
 	$(AR) $(LIBNAME) $(OBJS)
 	/bin/cp -p $(LIBNAME) $(LIBDIR)
 
+$(COBJS): spgpuinc
 
 objs: spgpuinc $(OBJS) iobjs cudaobjs spgpuobjs
 	/bin/cp -p *$(.mod) $(MODDIR)