Skip to content

GitLab

elpa
elpa
Commit d123a9d9 authored by Andreas Marek's avatar Andreas Marek
Browse files
Options

Unify real/complex GPU code path

parent 36a6642b
Hide whitespace changes
Inline Side-by-side
Showing with 45 additions and 182 deletions
src/GPU/mod_cuda.F90
View file @ d123a9d9
......@@ -201,14 +201,14 @@ module cuda_functions
implicit none
integer(kind=C_intptr_T), value :: dst
integer(kind=C_intptr_T), value :: dst
integer(kind=c_intptr_t), intent(in), value :: dpitch
integer(kind=C_intptr_T), value :: src
integer(kind=C_intptr_T), value :: src
integer(kind=c_intptr_t), intent(in), value :: spitch
integer(kind=c_intptr_t), intent(in), value :: width
integer(kind=c_intptr_t), intent(in), value :: height
integer(kind=C_INT), intent(in), value :: dir
integer(kind=C_INT) :: istat
integer(kind=C_INT), intent(in), value :: dir
integer(kind=C_INT) :: istat
end function cuda_memcpy2d_c
end interface
......
src/elpa2/elpa2_bandred_template.F90
View file @ d123a9d9
......@@ -171,7 +171,8 @@
#endif
#if COMPLEXCASE == 1
complex(kind=COMPLEX_DATATYPE), allocatable :: tmpCUDA(:,:), vmrCUDA(:,:), umcCUDA(:,:) ! note the different dimension in real case
! complex(kind=COMPLEX_DATATYPE), allocatable :: tmpCUDA(:,:), vmrCUDA(:,:), umcCUDA(:,:) ! note the different dimension in real case
complex(kind=COMPLEX_DATATYPE), allocatable :: tmpCUDA(:), vmrCUDA(:), umcCUDA(:)
complex(kind=COMPLEX_DATATYPE), allocatable :: tmpCPU(:,:), vmrCPU(:,:), umcCPU(:,:)
complex(kind=COMPLEX_DATATYPE), allocatable :: vr(:)
#endif
......@@ -438,12 +439,8 @@
endif
endif
#if REALCASE == 1
allocate(vmrCUDA(vmr_size), stat=istat, errmsg=errorMessage)
#endif
#if COMPLEXCASE == 1
allocate(vmrCUDA(max(l_rows,1),2*n_cols), stat=istat, errmsg=errorMessage)
#endif
if (istat .ne. 0) then
print *,"bandred_&
&MATH_DATATYPE&
......@@ -480,12 +477,8 @@
endif
#if REALCASE == 1
allocate(umcCUDA(umc_size), stat=istat, errmsg=errorMessage)
#endif
#if COMPLEXCASE == 1
allocate(umcCUDA(max(l_cols,1),2*n_cols), stat=istat, errmsg=errorMessage)
#endif
if (istat .ne. 0) then
print *,"bandred_&
&MATH_DATATYPE&
......@@ -540,7 +533,7 @@
vmrCUDA(1 : cur_l_rows * n_cols) = CONST_0_0
#endif
#if COMPLEXCASE == 1
vmrCUDA(1:l_rows,1:n_cols) = CONST_COMPLEX_0_0
vmrCUDA(1: cur_l_rows * n_cols) = CONST_COMPLEX_0_0
#endif
else
#if REALCASE == 1
......@@ -574,31 +567,11 @@
if (my_pcol == cur_pcol) then
successCUDA = cuda_memcpy2d(loc(a(1, lc_start)), &
#if REALCASE == 1
lda * size_of_datatype, &
#endif
#if COMPLEXCASE == 1
int((lda*size_of_datatype),kind=c_intptr_t), &
#endif
#if REALCASE == 1
(a_dev + ((lc_start-1) * lda*size_of_datatype)), &
#endif
#if COMPLEXCASE == 1
(a_dev + int( ( (lc_start-1) * lda*size_of_datatype),kind=c_intptr_t )), &
#endif
#if REALCASE == 1
lda*size_of_datatype, lr_end*size_of_datatype, &
#endif
#if COMPLEXCASE == 1
int(lda*size_of_datatype,kind=c_intptr_t), &
int(lr_end*size_of_datatype,kind=c_intptr_t), &
#endif
#if REALCASE == 1
(lc_end - lc_start+1), cudaMemcpyDeviceToHost)
#endif
#if COMPLEXCASE == 1
int((lc_end - lc_start+1),kind=c_intptr_t),int(cudaMemcpyDeviceToHost,kind=c_int))
#endif
if (.not.(successCUDA)) then
......@@ -607,7 +580,6 @@
&: error in cudaMemcpy2d"
stop 1
endif
endif
endif ! useGPU
......@@ -726,7 +698,7 @@
vr(lr+1) = tau
#ifdef WITH_MPI
if (wantDebug) call obj%timer%start("mpi_communication")
call MPI_Bcast(vr, lr+1, &
call MPI_Bcast(vr, lr+1, &
#if REALCASE == 1
MPI_REAL_PRECISION, &
#endif
......@@ -739,12 +711,7 @@
#endif /* WITH_MPI */
if (useGPU) then
#if REALCASE == 1
vmrCUDA(cur_l_rows * (lc - 1) + 1 : cur_l_rows * (lc - 1) + lr) = vr(1:lr)
#endif
#if COMPLEXCASE == 1
vmrCUDA(1:lr,lc) = vr(1:lr)
#endif
else
vmrCPU(1:lr,lc) = vr(1:lr)
endif
......@@ -943,37 +910,17 @@
cur_pcol = pcol(istep*nbw+1, nblk, np_cols)
if (my_pcol == cur_pcol) then
successCUDA = cuda_memcpy2d((a_dev+ &
#if REALCASE == 1
((lc_start-1)*lda*size_of_datatype)), &
#endif
#if COMPLEXCASE == 1
int(((lc_start-1)*lda*size_of_datatype),kind=c_intptr_t)), &
#endif
#if REALCASE == 1
lda*size_of_datatype, loc(a(1, lc_start)), &
#endif
#if COMPLEXCASE == 1
int(lda*size_of_datatype,kind=c_intptr_t), loc(a(1,lc_start)), &
#endif
#if REALCASE == 1
lda*size_of_datatype, lr_end*size_of_datatype, &
#endif
#if COMPLEXCASE == 1
int(lda*size_of_datatype,kind=c_intptr_t), &
int(lr_end*size_of_datatype,kind=c_intptr_t), &
#endif
#if REALCASE == 1
(lc_end - lc_start+1),cudaMemcpyHostToDevice)
#endif
#if COMPLEXCASE == 1
int((lc_end - lc_start+1),kind=c_intptr_t), &
int(cudaMemcpyHostToDevice,kind=c_int))
#endif
if (.not.(successCUDA)) then
print *, "bandred_&
&MATH_DATATYPE&
&: cuda memcpy a_dev failed ", istat
&MATH_DATATYPE&
&: cuda memcpy a_dev failed ", istat
stop 1
endif
endif
......@@ -993,13 +940,9 @@
call PRECISION_HERK('U', 'C', &
#endif
n_cols, l_rows, ONE, &
#if REALCASE == 1
vmrCUDA, cur_l_rows, &
#endif
#if COMPLEXCASE == 1
vmrCUDA, ubound(vmrCUDA,dim=1), &
#endif
ZERO, vav, ubound(vav,dim=1))
vmrCUDA, cur_l_rows, &
ZERO, vav, ubound(vav,dim=1))
else ! useGPU
if (l_rows>0) &
#if REALCASE == 1
......@@ -1050,20 +993,9 @@
&MATH_DATATYPE&
&_&
&PRECISION &
#if REALCASE == 1
(obj, vmrCUDA, cur_l_rows, &
#endif
#if COMPLEXCASE == 1
(obj, vmrCUDA, ubound(vmrCUDA,dim=1), &
#endif
mpi_comm_rows, &
#if REALCASE == 1
umcCUDA(cur_l_cols * n_cols + 1), cur_l_cols, &
#endif
#if COMPLEXCASE == 1
umcCUDA(1,n_cols+1), ubound(umcCUDA,dim=1), &
#endif
mpi_comm_cols, 1, istep*nbw, n_cols, nblk)
(obj, vmrCUDA, cur_l_rows, mpi_comm_rows, &
umcCUDA(cur_l_cols * n_cols + 1), cur_l_cols, &
mpi_comm_cols, 1, istep*nbw, n_cols, nblk)
else ! useGPU
call elpa_transpose_vectors_&
&MATH_DATATYPE&
......@@ -1124,7 +1056,7 @@
if (n_way > 1) then
#if REALCASE == 1
!$omp do
#endif
#endif
do i=1,min(l_cols_tile, l_cols)
#if REALCASE == 1
umcCPU(i,1:n_cols) = CONST_0_0
......@@ -1210,8 +1142,8 @@
vmrCUDA(cur_l_rows * n_cols + 1 : cur_l_rows * n_cols * 2) = CONST_0_0
#endif
#if COMPLEXCASE == 1
umcCUDA(1:l_cols,1:n_cols) = CONST_COMPLEX_0_0
vmrCUDA(1:l_rows,n_cols+1:2*n_cols) = CONST_COMPLEX_0_0
umcCUDA(1 : l_cols * n_cols) = CONST_COMPLEX_0_0
vmrCUDA(cur_l_rows * n_cols + 1 : cur_l_rows * n_cols * 2) = CONST_COMPLEX_0_0
#endif
else ! useGPU
#if REALCASE == 1
......@@ -1228,12 +1160,7 @@
if (useGPU) then
successCUDA = cuda_memcpy(vmr_dev, &
#if REALCASE == 1
loc(vmrCUDA(1)),&
#endif
#if COMPLEXCASE == 1
loc(vmrCUDA(1,1)), &
#endif
vmr_size*size_of_datatype,cudaMemcpyHostToDevice)
if (.not.(successCUDA)) then
print *,"bandred_&
......@@ -1241,13 +1168,9 @@
&: error in cudaMemcpy vmr_dev 3"
stop 1
endif
successCUDA = cuda_memcpy(umc_dev, &
#if REALCASE == 1
loc(umcCUDA(1)), &
#endif
#if COMPLEXCASE == 1
loc(umcCUDA(1,1)), &
#endif
umc_size*size_of_datatype,cudaMemcpyHostToDevice)
if (.not.(successCUDA)) then
print *,"bandred_&
......@@ -1295,7 +1218,6 @@
size_of_datatype), &
cur_l_rows)
call obj%timer%stop("cublas")
else ! useGPU
call obj%timer%start("blas")
......@@ -1320,12 +1242,7 @@
if (useGPU) then
successCUDA = cuda_memcpy( &
#if REALCASE == 1
loc(vmrCUDA(1)), &
#endif
#if COMPLEXCASE == 1
loc(vmrCUDA(1,1)), &
#endif
vmr_dev,vmr_size*size_of_datatype,cudaMemcpyDeviceToHost)
if (.not.(successCUDA)) then
print *,"bandred_&
......@@ -1335,12 +1252,7 @@
endif
successCUDA = cuda_memcpy( &
#if REALCASE == 1
loc(umcCUDA(1)), &
#endif
#if COMPLEXCASE == 1
loc(umcCUDA(1,1)), &
#endif
umc_dev, umc_size*size_of_datatype,cudaMemcpyDeviceToHost)
if (.not.(successCUDA)) then
print *,"bandred_&
......@@ -1371,20 +1283,9 @@
&MATH_DATATYPE&
&_&
&PRECISION &
#if REALCASE == 1
(obj, vmrCUDA(cur_l_rows * n_cols + 1),cur_l_rows, &
#endif
#if COMPLEXCASE == 1
(obj, vmrCUDA(1,n_cols+1),ubound(vmrCUDA,dim=1), &
#endif
mpi_comm_rows, umcCUDA, &
#if REALCASE == 1
cur_l_cols, &
#endif
#if COMPLEXCASE == 1
ubound(umcCUDA,dim=1), &
#endif
mpi_comm_cols, istep*nbw, n_cols, nblk)
cur_l_cols, mpi_comm_cols, istep*nbw, n_cols, nblk)
else ! useGPU
call elpa_reduce_add_vectors_&
......@@ -1401,12 +1302,7 @@
if (useGPU) then
#ifdef WITH_MPI
#if REALCASE == 1
allocate(tmpCUDA(l_cols * n_cols), stat=istat, errmsg=errorMessage)
#endif
#if COMPLEXCASE == 1
allocate(tmpCUDA(l_cols,n_cols), stat=istat, errmsg=errorMessage)
#endif
if (istat .ne. 0) then
print *,"bandred_&
&MATH_DATATYPE&
......@@ -1425,12 +1321,7 @@
#endif
MPI_SUM, mpi_comm_rows, ierr)
#if REALCASE == 1
umcCUDA(1 : l_cols * n_cols) = tmpCUDA(1 : l_cols * n_cols)
#endif
#if COMPLEXCASE == 1
umcCUDA(1:l_cols,1:n_cols) = tmpCUDA(1:l_cols,1:n_cols)
#endif
if (wantDebug) call obj%timer%stop("mpi_communication")
#else /* WITH_MPI */
......@@ -1488,12 +1379,7 @@
if (useGPU) then
successCUDA = cuda_memcpy(umc_dev, &
#if REALCASE == 1
loc(umcCUDA(1)), &
#endif
#if COMPLEXCASE == 1
loc(umcCUDA(1,1)), &
#endif
umc_size*size_of_datatype, cudaMemcpyHostToDevice)
if (.not.(successCUDA)) then
print *,"bandred_&
......@@ -1501,7 +1387,6 @@
&: error in cudaMemcpy umc_dev 5"
stop 1
endif
successCUDA = cuda_memcpy(tmat_dev,loc(tmat(1,1,istep)),nbw*nbw*size_of_datatype,cudaMemcpyHostToDevice)
if (.not.(successCUDA)) then
print *,"bandred_&
......@@ -1630,12 +1515,7 @@
call obj%timer%stop("cublas")
successCUDA = cuda_memcpy( &
#if REALCASE == 1
loc(umcCUDA(1)), &
#endif
#if COMPLEXCASE == 1
loc(umcCUDA(1,1)), &
#endif
umc_dev, umc_size*size_of_datatype, cudaMemcpyDeviceToHost)
if (.not.(successCUDA)) then
......@@ -1650,24 +1530,12 @@
&MATH_DATATYPE&
&_&
&PRECISION &
#if REALCASE == 1
(obj, umcCUDA, cur_l_cols, mpi_comm_cols, &
vmrCUDA(cur_l_rows * n_cols + 1), cur_l_rows, mpi_comm_rows, &
#endif
#if COMPLEXCASE == 1
(obj, umcCUDA, ubound(umcCUDA,dim=1), mpi_comm_cols, &
vmrCUDA(1,n_cols+1), ubound(vmrCUDA,dim=1), mpi_comm_rows, &
#endif
1, istep*nbw, n_cols, nblk)
successCUDA = cuda_memcpy(vmr_dev, &
#if REALCASE == 1
loc(vmrCUDA(1)), &
#endif
#if COMPLEXCASE == 1
loc(vmrCUDA(1,1)), &
#endif
vmr_size*size_of_datatype, cudaMemcpyHostToDevice)
if (.not.(successCUDA)) then
print *,"bandred_&
......@@ -1677,12 +1545,7 @@
endif
successCUDA = cuda_memcpy(umc_dev, &
#if REALCASE == 1
loc(umcCUDA(1)), &
#endif
#if COMPLEXCASE == 1
loc(umcCUDA(1,1)), &
#endif
umc_size*size_of_datatype, cudaMemcpyHostToDevice)
if (.not.(successCUDA)) then
print *,"bandred_&
......@@ -1819,8 +1682,8 @@
deallocate(vr, stat=istat, errmsg=errorMessage)
if (istat .ne. 0) then
print *,"bandred_&
&MATH_DATATYPE&
&: error when deallocating vr "//errorMessage
&MATH_DATATYPE&
&: error when deallocating vr "//errorMessage
stop 1
endif
endif
......@@ -1829,8 +1692,8 @@
deallocate(umcCPU, stat=istat, errmsg=errorMessage)
if (istat .ne. 0) then
print *,"bandred_&
&MATH_DATATYPE&
&: error when deallocating umcCPU "//errorMessage
&MATH_DATATYPE&
&: error when deallocating umcCPU "//errorMessage
stop 1
endif
endif
......@@ -1839,8 +1702,8 @@
deallocate(vmrCPU, stat=istat, errmsg=errorMessage)
if (istat .ne. 0) then
print *,"bandred_&
&MATH_DATATYPE&
&: error when deallocating vmrCPU "//errorMessage
&MATH_DATATYPE&
&: error when deallocating vmrCPU "//errorMessage
stop 1
endif
endif
......@@ -1852,8 +1715,8 @@
successCUDA = cuda_free(vav_dev)
if (.not.(successCUDA)) then
print *,"bandred_&
&MATH_DATATYPE&
&: error in cudaFree vav_dev 4"
&MATH_DATATYPE&
&: error in cudaFree vav_dev 4"
stop 1
endif
endif ! useGPU
......@@ -1862,8 +1725,8 @@
deallocate(vr, stat=istat, errmsg=errorMessage)
if (istat .ne. 0) then
print *,"bandred_&
&MATH_DATATYPE&
&: error when deallocating vr "//errorMessage
&MATH_DATATYPE&
&: error when deallocating vr "//errorMessage
stop 1
endif
endif
......@@ -1872,8 +1735,8 @@
deallocate(umcCPU, stat=istat, errmsg=errorMessage)
if (istat .ne. 0) then
print *,"bandred_&
&MATH_DATATYPE&
&: error when deallocating umcCPU "//errorMessage
&MATH_DATATYPE&
&: error when deallocating umcCPU "//errorMessage
stop 1
endif
endif
......@@ -1882,8 +1745,8 @@
deallocate(vmrCPU, stat=istat, errmsg=errorMessage)
if (istat .ne. 0) then
print *,"bandred_&
&MATH_DATATYPE&
&: error when deallocating vmrCPU "//errorMessage
&MATH_DATATYPE&
&: error when deallocating vmrCPU "//errorMessage
stop 1
endif
endif
......@@ -1903,16 +1766,16 @@
successCUDA = cuda_free(vmr_dev)
if (.not.(successCUDA)) then
print *,"bandred_&
&MATH_DATATYPE&
&: error in cudaFree vmr_dev 6"
&MATH_DATATYPE&
&: error in cudaFree vmr_dev 6"
stop 1
endif
successCUDA = cuda_free(umc_dev)
if (.not.(successCUDA)) then
print *,"bandred_&
&MATH_DATATYPE&
&: error in cudaFree umc_dev 8"
&MATH_DATATYPE&
&: error in cudaFree umc_dev 8"
stop
endif
......@@ -1920,8 +1783,8 @@
deallocate(umcCUDA, stat=istat, errmsg=errorMessage)
if (istat .ne. 0) then
print *,"bandred_&
&MATH_DATATYPE&
&: error when deallocating umcCUDA "//errorMessage
&MATH_DATATYPE&
&: error when deallocating umcCUDA "//errorMessage
stop 1
endif
endif
......@@ -1929,8 +1792,8 @@
deallocate(vmrCUDA, stat=istat, errmsg=errorMessage)
if (istat .ne. 0) then
print *,"bandred_&
&MATH_DATATYPE&
&: error when deallocating vmrCUDA "//errorMessage
&MATH_DATATYPE&
&: error when deallocating vmrCUDA "//errorMessage
stop 1
endif
endif
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment