You cannot select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
380 lines
13 KiB
Fortran
380 lines
13 KiB
Fortran
! Parallel Sparse BLAS GPU plugin
|
|
! (C) Copyright 2013
|
|
!
|
|
! Salvatore Filippone
|
|
! Alessandro Fanfarillo
|
|
!
|
|
! Redistribution and use in source and binary forms, with or without
|
|
! modification, are permitted provided that the following conditions
|
|
! are met:
|
|
! 1. Redistributions of source code must retain the above copyright
|
|
! notice, this list of conditions and the following disclaimer.
|
|
! 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 PSBLAS 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 PSBLAS 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
|
|
! INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
|
|
! CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
|
|
! ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
|
|
! POSSIBILITY OF SUCH DAMAGE.
|
|
!
|
|
|
|
|
|
module psb_d_cuda_hlg_mat_mod
|
|
|
|
use iso_c_binding
|
|
use psb_d_mat_mod
|
|
use psb_d_hll_mat_mod
|
|
|
|
|
|
integer(psb_ipk_), parameter, private :: is_host = -1
|
|
integer(psb_ipk_), parameter, private :: is_sync = 0
|
|
integer(psb_ipk_), parameter, private :: is_dev = 1
|
|
|
|
type, extends(psb_d_hll_sparse_mat) :: psb_d_cuda_hlg_sparse_mat
|
|
!
|
|
! ITPACK/HLL format, extended.
|
|
! We are adding here the routines to create a copy of the data
|
|
! into the GPU.
|
|
! If HAVE_SPGPU is undefined this is just
|
|
! a copy of HLL, indistinguishable.
|
|
!
|
|
type(c_ptr) :: deviceMat = c_null_ptr
|
|
integer :: devstate = is_host
|
|
|
|
contains
|
|
procedure, nopass :: get_fmt => d_cuda_hlg_get_fmt
|
|
procedure, pass(a) :: sizeof => d_cuda_hlg_sizeof
|
|
procedure, pass(a) :: vect_mv => psb_d_cuda_hlg_vect_mv
|
|
procedure, pass(a) :: csmm => psb_d_cuda_hlg_csmm
|
|
procedure, pass(a) :: csmv => psb_d_cuda_hlg_csmv
|
|
procedure, pass(a) :: in_vect_sv => psb_d_cuda_hlg_inner_vect_sv
|
|
procedure, pass(a) :: scals => psb_d_cuda_hlg_scals
|
|
procedure, pass(a) :: scalv => psb_d_cuda_hlg_scal
|
|
procedure, pass(a) :: reallocate_nz => psb_d_cuda_hlg_reallocate_nz
|
|
procedure, pass(a) :: allocate_mnnz => psb_d_cuda_hlg_allocate_mnnz
|
|
! Note: we do *not* need the TO methods, because the parent type
|
|
! methods will work.
|
|
procedure, pass(a) :: cp_from_coo => psb_d_cuda_cp_hlg_from_coo
|
|
procedure, pass(a) :: cp_from_fmt => psb_d_cuda_cp_hlg_from_fmt
|
|
procedure, pass(a) :: mv_from_coo => psb_d_cuda_mv_hlg_from_coo
|
|
procedure, pass(a) :: mv_from_fmt => psb_d_cuda_mv_hlg_from_fmt
|
|
procedure, pass(a) :: free => d_cuda_hlg_free
|
|
procedure, pass(a) :: mold => psb_d_cuda_hlg_mold
|
|
procedure, pass(a) :: is_host => d_cuda_hlg_is_host
|
|
procedure, pass(a) :: is_dev => d_cuda_hlg_is_dev
|
|
procedure, pass(a) :: is_sync => d_cuda_hlg_is_sync
|
|
procedure, pass(a) :: set_host => d_cuda_hlg_set_host
|
|
procedure, pass(a) :: set_dev => d_cuda_hlg_set_dev
|
|
procedure, pass(a) :: set_sync => d_cuda_hlg_set_sync
|
|
procedure, pass(a) :: sync => d_cuda_hlg_sync
|
|
procedure, pass(a) :: from_gpu => psb_d_cuda_hlg_from_gpu
|
|
procedure, pass(a) :: to_gpu => psb_d_cuda_hlg_to_gpu
|
|
final :: d_cuda_hlg_finalize
|
|
end type psb_d_cuda_hlg_sparse_mat
|
|
|
|
private :: d_cuda_hlg_get_nzeros, d_cuda_hlg_free, d_cuda_hlg_get_fmt, &
|
|
& d_cuda_hlg_get_size, d_cuda_hlg_sizeof, d_cuda_hlg_get_nz_row
|
|
|
|
|
|
interface
|
|
subroutine psb_d_cuda_hlg_vect_mv(alpha,a,x,beta,y,info,trans)
|
|
import :: psb_d_cuda_hlg_sparse_mat, psb_dpk_, psb_d_base_vect_type, psb_ipk_
|
|
class(psb_d_cuda_hlg_sparse_mat), intent(in) :: a
|
|
real(psb_dpk_), intent(in) :: alpha, beta
|
|
class(psb_d_base_vect_type), intent(inout) :: x
|
|
class(psb_d_base_vect_type), intent(inout) :: y
|
|
integer(psb_ipk_), intent(out) :: info
|
|
character, optional, intent(in) :: trans
|
|
end subroutine psb_d_cuda_hlg_vect_mv
|
|
end interface
|
|
|
|
interface
|
|
subroutine psb_d_cuda_hlg_inner_vect_sv(alpha,a,x,beta,y,info,trans)
|
|
import :: psb_ipk_, psb_d_cuda_hlg_sparse_mat, psb_dpk_, psb_d_base_vect_type
|
|
class(psb_d_cuda_hlg_sparse_mat), intent(in) :: a
|
|
real(psb_dpk_), intent(in) :: alpha, beta
|
|
class(psb_d_base_vect_type), intent(inout) :: x, y
|
|
integer(psb_ipk_), intent(out) :: info
|
|
character, optional, intent(in) :: trans
|
|
end subroutine psb_d_cuda_hlg_inner_vect_sv
|
|
end interface
|
|
|
|
interface
|
|
subroutine psb_d_cuda_hlg_reallocate_nz(nz,a)
|
|
import :: psb_d_cuda_hlg_sparse_mat, psb_ipk_
|
|
integer(psb_ipk_), intent(in) :: nz
|
|
class(psb_d_cuda_hlg_sparse_mat), intent(inout) :: a
|
|
end subroutine psb_d_cuda_hlg_reallocate_nz
|
|
end interface
|
|
|
|
interface
|
|
subroutine psb_d_cuda_hlg_allocate_mnnz(m,n,a,nz)
|
|
import :: psb_d_cuda_hlg_sparse_mat, psb_ipk_
|
|
integer(psb_ipk_), intent(in) :: m,n
|
|
class(psb_d_cuda_hlg_sparse_mat), intent(inout) :: a
|
|
integer(psb_ipk_), intent(in), optional :: nz
|
|
end subroutine psb_d_cuda_hlg_allocate_mnnz
|
|
end interface
|
|
|
|
interface
|
|
subroutine psb_d_cuda_hlg_mold(a,b,info)
|
|
import :: psb_d_cuda_hlg_sparse_mat, psb_d_base_sparse_mat, psb_ipk_
|
|
class(psb_d_cuda_hlg_sparse_mat), intent(in) :: a
|
|
class(psb_d_base_sparse_mat), intent(inout), allocatable :: b
|
|
integer(psb_ipk_), intent(out) :: info
|
|
end subroutine psb_d_cuda_hlg_mold
|
|
end interface
|
|
|
|
interface
|
|
subroutine psb_d_cuda_hlg_from_gpu(a,info)
|
|
import :: psb_d_cuda_hlg_sparse_mat, psb_ipk_
|
|
class(psb_d_cuda_hlg_sparse_mat), intent(inout) :: a
|
|
integer(psb_ipk_), intent(out) :: info
|
|
end subroutine psb_d_cuda_hlg_from_gpu
|
|
end interface
|
|
|
|
interface
|
|
subroutine psb_d_cuda_hlg_to_gpu(a,info, nzrm)
|
|
import :: psb_d_cuda_hlg_sparse_mat, psb_ipk_
|
|
class(psb_d_cuda_hlg_sparse_mat), intent(inout) :: a
|
|
integer(psb_ipk_), intent(out) :: info
|
|
integer(psb_ipk_), intent(in), optional :: nzrm
|
|
end subroutine psb_d_cuda_hlg_to_gpu
|
|
end interface
|
|
|
|
interface
|
|
subroutine psb_d_cuda_cp_hlg_from_coo(a,b,info)
|
|
import :: psb_d_cuda_hlg_sparse_mat, psb_d_coo_sparse_mat, psb_ipk_
|
|
class(psb_d_cuda_hlg_sparse_mat), intent(inout) :: a
|
|
class(psb_d_coo_sparse_mat), intent(in) :: b
|
|
integer(psb_ipk_), intent(out) :: info
|
|
end subroutine psb_d_cuda_cp_hlg_from_coo
|
|
end interface
|
|
|
|
interface
|
|
subroutine psb_d_cuda_cp_hlg_from_fmt(a,b,info)
|
|
import :: psb_d_cuda_hlg_sparse_mat, psb_d_base_sparse_mat, psb_ipk_
|
|
class(psb_d_cuda_hlg_sparse_mat), intent(inout) :: a
|
|
class(psb_d_base_sparse_mat), intent(in) :: b
|
|
integer(psb_ipk_), intent(out) :: info
|
|
end subroutine psb_d_cuda_cp_hlg_from_fmt
|
|
end interface
|
|
|
|
interface
|
|
subroutine psb_d_cuda_mv_hlg_from_coo(a,b,info)
|
|
import :: psb_d_cuda_hlg_sparse_mat, psb_d_coo_sparse_mat, psb_ipk_
|
|
class(psb_d_cuda_hlg_sparse_mat), intent(inout) :: a
|
|
class(psb_d_coo_sparse_mat), intent(inout) :: b
|
|
integer(psb_ipk_), intent(out) :: info
|
|
end subroutine psb_d_cuda_mv_hlg_from_coo
|
|
end interface
|
|
|
|
|
|
interface
|
|
subroutine psb_d_cuda_mv_hlg_from_fmt(a,b,info)
|
|
import :: psb_d_cuda_hlg_sparse_mat, psb_d_base_sparse_mat, psb_ipk_
|
|
class(psb_d_cuda_hlg_sparse_mat), intent(inout) :: a
|
|
class(psb_d_base_sparse_mat), intent(inout) :: b
|
|
integer(psb_ipk_), intent(out) :: info
|
|
end subroutine psb_d_cuda_mv_hlg_from_fmt
|
|
end interface
|
|
|
|
interface
|
|
subroutine psb_d_cuda_hlg_csmv(alpha,a,x,beta,y,info,trans)
|
|
import :: psb_d_cuda_hlg_sparse_mat, psb_dpk_, psb_ipk_
|
|
class(psb_d_cuda_hlg_sparse_mat), intent(in) :: a
|
|
real(psb_dpk_), intent(in) :: alpha, beta, x(:)
|
|
real(psb_dpk_), intent(inout) :: y(:)
|
|
integer(psb_ipk_), intent(out) :: info
|
|
character, optional, intent(in) :: trans
|
|
end subroutine psb_d_cuda_hlg_csmv
|
|
end interface
|
|
interface
|
|
subroutine psb_d_cuda_hlg_csmm(alpha,a,x,beta,y,info,trans)
|
|
import :: psb_d_cuda_hlg_sparse_mat, psb_dpk_, psb_ipk_
|
|
class(psb_d_cuda_hlg_sparse_mat), intent(in) :: a
|
|
real(psb_dpk_), intent(in) :: alpha, beta, x(:,:)
|
|
real(psb_dpk_), intent(inout) :: y(:,:)
|
|
integer(psb_ipk_), intent(out) :: info
|
|
character, optional, intent(in) :: trans
|
|
end subroutine psb_d_cuda_hlg_csmm
|
|
end interface
|
|
|
|
interface
|
|
subroutine psb_d_cuda_hlg_scal(d,a,info, side)
|
|
import :: psb_d_cuda_hlg_sparse_mat, psb_dpk_, psb_ipk_
|
|
class(psb_d_cuda_hlg_sparse_mat), intent(inout) :: a
|
|
real(psb_dpk_), intent(in) :: d(:)
|
|
integer(psb_ipk_), intent(out) :: info
|
|
character, intent(in), optional :: side
|
|
end subroutine psb_d_cuda_hlg_scal
|
|
end interface
|
|
|
|
interface
|
|
subroutine psb_d_cuda_hlg_scals(d,a,info)
|
|
import :: psb_d_cuda_hlg_sparse_mat, psb_dpk_, psb_ipk_
|
|
class(psb_d_cuda_hlg_sparse_mat), intent(inout) :: a
|
|
real(psb_dpk_), intent(in) :: d
|
|
integer(psb_ipk_), intent(out) :: info
|
|
end subroutine psb_d_cuda_hlg_scals
|
|
end interface
|
|
|
|
|
|
contains
|
|
|
|
! == ===================================
|
|
!
|
|
!
|
|
!
|
|
! Getters
|
|
!
|
|
!
|
|
!
|
|
!
|
|
!
|
|
! == ===================================
|
|
|
|
|
|
function d_cuda_hlg_sizeof(a) result(res)
|
|
implicit none
|
|
class(psb_d_cuda_hlg_sparse_mat), intent(in) :: a
|
|
integer(psb_epk_) :: res
|
|
|
|
|
|
if (a%is_dev()) call a%sync()
|
|
res = 8
|
|
res = res + psb_sizeof_dp * size(a%val)
|
|
res = res + psb_sizeof_ip * size(a%irn)
|
|
res = res + psb_sizeof_ip * size(a%idiag)
|
|
res = res + psb_sizeof_ip * size(a%hkoffs)
|
|
res = res + psb_sizeof_ip * size(a%ja)
|
|
! Should we account for the shadow data structure
|
|
! on the GPU device side?
|
|
! res = 2*res
|
|
|
|
end function d_cuda_hlg_sizeof
|
|
|
|
function d_cuda_hlg_get_fmt() result(res)
|
|
implicit none
|
|
character(len=5) :: res
|
|
res = 'HLG'
|
|
end function d_cuda_hlg_get_fmt
|
|
|
|
|
|
|
|
! == ===================================
|
|
!
|
|
!
|
|
!
|
|
! Data management
|
|
!
|
|
!
|
|
!
|
|
!
|
|
!
|
|
! == ===================================
|
|
|
|
subroutine d_cuda_hlg_free(a)
|
|
use hlldev_mod
|
|
implicit none
|
|
integer(psb_ipk_) :: info
|
|
class(psb_d_cuda_hlg_sparse_mat), intent(inout) :: a
|
|
|
|
if (c_associated(a%deviceMat)) &
|
|
& call freeHllDevice(a%deviceMat)
|
|
a%deviceMat = c_null_ptr
|
|
call a%psb_d_hll_sparse_mat%free()
|
|
|
|
return
|
|
|
|
end subroutine d_cuda_hlg_free
|
|
|
|
|
|
subroutine d_cuda_hlg_sync(a)
|
|
implicit none
|
|
class(psb_d_cuda_hlg_sparse_mat), target, intent(in) :: a
|
|
class(psb_d_cuda_hlg_sparse_mat), pointer :: tmpa
|
|
integer(psb_ipk_) :: info
|
|
|
|
tmpa => a
|
|
if (tmpa%is_host()) then
|
|
call tmpa%to_gpu(info)
|
|
else if (tmpa%is_dev()) then
|
|
call tmpa%from_gpu(info)
|
|
end if
|
|
call tmpa%set_sync()
|
|
return
|
|
|
|
end subroutine d_cuda_hlg_sync
|
|
|
|
subroutine d_cuda_hlg_set_host(a)
|
|
implicit none
|
|
class(psb_d_cuda_hlg_sparse_mat), intent(inout) :: a
|
|
|
|
a%devstate = is_host
|
|
end subroutine d_cuda_hlg_set_host
|
|
|
|
subroutine d_cuda_hlg_set_dev(a)
|
|
implicit none
|
|
class(psb_d_cuda_hlg_sparse_mat), intent(inout) :: a
|
|
|
|
a%devstate = is_dev
|
|
end subroutine d_cuda_hlg_set_dev
|
|
|
|
subroutine d_cuda_hlg_set_sync(a)
|
|
implicit none
|
|
class(psb_d_cuda_hlg_sparse_mat), intent(inout) :: a
|
|
|
|
a%devstate = is_sync
|
|
end subroutine d_cuda_hlg_set_sync
|
|
|
|
function d_cuda_hlg_is_dev(a) result(res)
|
|
implicit none
|
|
class(psb_d_cuda_hlg_sparse_mat), intent(in) :: a
|
|
logical :: res
|
|
|
|
res = (a%devstate == is_dev)
|
|
end function d_cuda_hlg_is_dev
|
|
|
|
function d_cuda_hlg_is_host(a) result(res)
|
|
implicit none
|
|
class(psb_d_cuda_hlg_sparse_mat), intent(in) :: a
|
|
logical :: res
|
|
|
|
res = (a%devstate == is_host)
|
|
end function d_cuda_hlg_is_host
|
|
|
|
function d_cuda_hlg_is_sync(a) result(res)
|
|
implicit none
|
|
class(psb_d_cuda_hlg_sparse_mat), intent(in) :: a
|
|
logical :: res
|
|
|
|
res = (a%devstate == is_sync)
|
|
end function d_cuda_hlg_is_sync
|
|
|
|
|
|
subroutine d_cuda_hlg_finalize(a)
|
|
use hlldev_mod
|
|
implicit none
|
|
type(psb_d_cuda_hlg_sparse_mat), intent(inout) :: a
|
|
|
|
if (c_associated(a%deviceMat)) &
|
|
& call freeHllDevice(a%deviceMat)
|
|
a%deviceMat = c_null_ptr
|
|
|
|
return
|
|
end subroutine d_cuda_hlg_finalize
|
|
|
|
end module psb_d_cuda_hlg_mat_mod
|