Skip to content

GitLab

elpa
elpa
Commit 05ca1675 authored by Andreas Marek's avatar Andreas Marek
Browse files
Options

Disentangle 64bit integer blas from MPI

parent 043ddf39
Hide whitespace changes
Inline Side-by-side
Showing with 588 additions and 435 deletions
.gitlab-ci.yml
View file @ 05ca1675
......@@ -87,7 +87,7 @@ gnu-gnu-nompi-noopenmp-ilp64:
when: on_success
expire_in: 2 month
script:
- ./ci_test_scripts/run_ci_tests.sh -c "CC=\"gcc\" CFLAGS=\"-O3 -mavx\" FC=\"gfortran\" FCFLAGS=\"-O3 -mavx\" SCALAPACK_LDFLAGS=\"$MKL_GFORTRAN_SCALAPACK_LDFLAGS_NOMPI_NOOMP_ILP64 \" SCALAPACK_FCFLAGS=\"$MKL_GFORTRAN_SCALAPACK_FCFLAGS_NOMPI_NOOMP_ILP64 \" --enable-option-checking=fatal --with-mpi=no --disable-openmp --disable-gpu --enable-avx --disable-avx2 --disable-avx512 --enable-64bit-integer-support || { cat config.log; exit 1; }" -j 8 -t $MPI_TASKS -m $MATRIX_SIZE -n $NUMBER_OF_EIGENVECTORS -b $BLOCK_SIZE -s $SKIP_STEP -i $INTERACTIVE_RUN -S $SLURM
- ./ci_test_scripts/run_ci_tests.sh -c "CC=\"gcc\" CFLAGS=\"-O3 -mavx\" FC=\"gfortran\" FCFLAGS=\"-O3 -mavx\" SCALAPACK_LDFLAGS=\"$MKL_GFORTRAN_SCALAPACK_LDFLAGS_NOMPI_NOOMP_ILP64 \" SCALAPACK_FCFLAGS=\"$MKL_GFORTRAN_SCALAPACK_FCFLAGS_NOMPI_NOOMP_ILP64 \" --enable-option-checking=fatal --with-mpi=no --disable-openmp --disable-gpu --enable-avx --disable-avx2 --disable-avx512 --enable-64bit-integer-math-support || { cat config.log; exit 1; }" -j 8 -t $MPI_TASKS -m $MATRIX_SIZE -n $NUMBER_OF_EIGENVECTORS -b $BLOCK_SIZE -s $SKIP_STEP -i $INTERACTIVE_RUN -S $SLURM
# gnu-gnu-ilp64-nompi-openmp
......@@ -98,7 +98,7 @@ gnu-gnu-nompi-openmp-ilp64:
when: on_success
expire_in: 2 month
script:
- ./ci_test_scripts/run_ci_tests.sh -c "CC=\"gcc\" CFLAGS=\"-O3 -mavx\" FC=\"gfortran\" FCFLAGS=\"-O3 -mavx\" SCALAPACK_LDFLAGS=\"$MKL_GFORTRAN_SCALAPACK_LDFLAGS_NOMPI_OMP_ILP64 \" SCALAPACK_FCFLAGS=\"$MKL_GFORTRAN_SCALAPACK_FCFLAGS_NOMPI_OMP_ILP64 \" --enable-option-checking=fatal --with-mpi=no --enable-openmp --disable-gpu --enable-avx --disable-avx2 --disable-avx512 --enable-64bit-integer-support || { cat config.log; exit 1; }" -j 8 -t $MPI_TASKS -m $MATRIX_SIZE -n $NUMBER_OF_EIGENVECTORS -b $BLOCK_SIZE -s $SKIP_STEP -i $INTERACTIVE_RUN -S $SLURM
- ./ci_test_scripts/run_ci_tests.sh -c "CC=\"gcc\" CFLAGS=\"-O3 -mavx\" FC=\"gfortran\" FCFLAGS=\"-O3 -mavx\" SCALAPACK_LDFLAGS=\"$MKL_GFORTRAN_SCALAPACK_LDFLAGS_NOMPI_OMP_ILP64 \" SCALAPACK_FCFLAGS=\"$MKL_GFORTRAN_SCALAPACK_FCFLAGS_NOMPI_OMP_ILP64 \" --enable-option-checking=fatal --with-mpi=no --enable-openmp --disable-gpu --enable-avx --disable-avx2 --disable-avx512 --enable-64bit-integer-math-support || { cat config.log; exit 1; }" -j 8 -t $MPI_TASKS -m $MATRIX_SIZE -n $NUMBER_OF_EIGENVECTORS -b $BLOCK_SIZE -s $SKIP_STEP -i $INTERACTIVE_RUN -S $SLURM
# python tests
......
ci_test_scripts/generate_gitlab_ci_tests.py
View file @ 05ca1675
......@@ -348,7 +348,7 @@ ilp64_no_omp_tests = [
'SCALAPACK_LDFLAGS=\\"$MKL_GFORTRAN_SCALAPACK_LDFLAGS_NOMPI_NOOMP_ILP64 \\" '
'SCALAPACK_FCFLAGS=\\"$MKL_GFORTRAN_SCALAPACK_FCFLAGS_NOMPI_NOOMP_ILP64 \\" '
'--enable-option-checking=fatal --with-mpi=no --disable-openmp '
'--disable-gpu --enable-avx --disable-avx2 --disable-avx512 --enable-64bit-integer-support || { cat config.log; exit 1; }'
'--disable-gpu --enable-avx --disable-avx2 --disable-avx512 --enable-64bit-integer-math-support || { cat config.log; exit 1; }'
'" -j 8 -t $MPI_TASKS -m $MATRIX_SIZE -n $NUMBER_OF_EIGENVECTORS -b $BLOCK_SIZE '
'-s $SKIP_STEP -i $INTERACTIVE_RUN -S $SLURM',
"\n",
......@@ -366,7 +366,7 @@ ilp64_no_omp_tests = [
'SCALAPACK_LDFLAGS=\\"$MKL_GFORTRAN_SCALAPACK_LDFLAGS_NOMPI_OMP_ILP64 \\" '
'SCALAPACK_FCFLAGS=\\"$MKL_GFORTRAN_SCALAPACK_FCFLAGS_NOMPI_OMP_ILP64 \\" '
'--enable-option-checking=fatal --with-mpi=no --enable-openmp '
'--disable-gpu --enable-avx --disable-avx2 --disable-avx512 --enable-64bit-integer-support || { cat config.log; exit 1; }'
'--disable-gpu --enable-avx --disable-avx2 --disable-avx512 --enable-64bit-integer-math-support || { cat config.log; exit 1; }'
'" -j 8 -t $MPI_TASKS -m $MATRIX_SIZE -n $NUMBER_OF_EIGENVECTORS -b $BLOCK_SIZE '
'-s $SKIP_STEP -i $INTERACTIVE_RUN -S $SLURM',
"\n",
......
configure.ac
View file @ 05ca1675
......@@ -213,21 +213,21 @@ dnl first long int
AC_CHECK_SIZEOF([long int])
size_of_long_int="${ac_cv_sizeof_long_int}"
dnl then 64bit
dnl then 64bit blas
AC_MSG_CHECKING(whether 64bit integers should be used for math libraries (BLAS/LAPACK/SCALAPACK))
AC_ARG_ENABLE([64bit-integer-support],
AS_HELP_STRING([--64bit-integer-support],
AC_ARG_ENABLE([64bit-integer-math-support],
AS_HELP_STRING([--64bit-integer-math-support],
[allows to link against the 64bit integer versions of the math libraries BLAS, LAPACK, and SCALAPACK]),
[
if test x"$enableval" = x"yes"; then
enable_64bit_integer_support=yes
enable_64bit_integer_math_support=yes
else
enable_64bit_integer_support=no
enable_64bit_integer_math_support=no
fi
],
[enable_64bit_integer_support="no"])
AC_MSG_RESULT([$enable_64bit_integer_support])
if test x"${enable_64bit_integer_support}" = x"yes"; then
[enable_64bit_integer_math_support="no"])
AC_MSG_RESULT([$enable_64bit_integer_math_support])
if test x"${enable_64bit_integer_math_support}" = x"yes"; then
if test x"${enable_legacy}" = x"yes"; then
AC_MSG_ERROR([You cannot both enable 64bit integer support and the legacy interface!])
fi
......@@ -242,9 +242,32 @@ if test x"${enable_64bit_integer_support}" = x"yes"; then
AC_MSG_ERROR([The C data-type "long int" is only ${size_of_long_int} bytes; Needed is 8 bytes])
fi
AC_DEFINE([HAVE_64BIT_INTEGER_SUPPORT], [1], [allow to link against the 64bit integer versions of math libraries])
AC_DEFINE([HAVE_64BIT_INTEGER_MATH_SUPPORT], [1], [allow to link against the 64bit integer versions of math libraries])
fi
AM_CONDITIONAL([HAVE_64BIT_INTEGER_SUPPORT],[test x"$enable_64bit_integer_support" = x"yes"])
AM_CONDITIONAL([HAVE_64BIT_INTEGER_MATH_SUPPORT],[test x"$enable_64bit_integer_math_support" = x"yes"])
dnl then 64bit blas
AC_MSG_CHECKING(whether 64bit integers should be used for the MPI library)
AC_ARG_ENABLE([64bit-integer-mpi-support],
AS_HELP_STRING([--64bit-integer-mpi-support],
[allows to link against the 64bit integer versions of the MPI library]),
[
if test x"$enableval" = x"yes"; then
enable_64bit_integer_mpi_support=yes
else
enable_64bit_integer_mpi_support=no
fi
],
[enable_64bit_integer_mpi_support="no"])
AC_MSG_RESULT([$enable_64bit_integer_mpi_support])
if test x"${enable_64bit_integer_mpi_support}" = x"yes"; then
if test x"${enable_legacy}" = x"yes"; then
AC_MSG_ERROR([You cannot both enable 64bit integer support and the legacy interface!])
fi
AC_DEFINE([HAVE_64BIT_INTEGER_MPI_SUPPORT], [1], [allow to link against the 64bit integer versions of the MPI library])
fi
AM_CONDITIONAL([HAVE_64BIT_INTEGER_MPI_SUPPORT],[test x"$enable_64bit_integer_mpi_support" = x"yes"])
AC_MSG_CHECKING(whether C compiler can use _Generic )
......
src/elpa2/elpa2_bandred_template.F90
View file @ 05ca1675
......@@ -171,8 +171,10 @@
#endif
integer(kind=ik) :: lr_end
integer(kind=ik) :: na_cols
integer(kind=BLAS_KIND) :: na_colsBLAS
#if COMPLEXCASE == 1
integer(kind=ik) :: na_rows
integer(kind=BLAS_KIND) :: na_rowsBLAS
#endif
logical, intent(in) :: wantDebug
......@@ -256,11 +258,13 @@
if (useGPU) then
#ifdef WITH_MPI
#if COMPLEXCASE == 1
na_rows = numroc(int(na,kind=MPI_KIND), int(nblk,kind=MPI_KIND), &
int(my_prow,kind=MPI_KIND), 0_MPI_KIND, int(np_rows,kind=MPI_KIND))
na_rowsBLAS = numroc(int(na,kind=BLAS_KIND), int(nblk,kind=BLAS_KIND), &
int(my_prow,kind=BLAS_KIND), 0_BLAS_KIND, int(np_rows,kind=BLAS_KIND))
na_rows = int(na_rowsBLAS,kind=c_int)
#endif
na_cols = numroc(int(na,kind=MPI_KIND), int(nblk,kind=MPI_KIND), &
int(my_pcol,kind=MPI_KIND), 0_MPI_KIND, int(np_cols,kind=MPI_KIND))
na_colsBLAS = numroc(int(na,kind=BLAS_KIND), int(nblk,kind=BLAS_KIND), &
int(my_pcol,kind=BLAS_KIND), 0_BLAS_KIND, int(np_cols,kind=BLAS_KIND))
na_cols = int(na_colsBLAS,kind=c_int)
#else
#if COMPLEXCASE == 1
na_rows = na
......
src/elpa2/qr/elpa_pdlarfb_template.F90
View file @ 05ca1675
......@@ -64,7 +64,7 @@ subroutine qr_pdlarfb_1dcomm_&
! local scalars
integer(kind=ik) :: localsize,offset,baseoffset
integer(kind=ik) :: mpirank, mpiprocs
integer(kind=BLAS_KIND) :: mpirankMPI, mpiprocsMPI, mpierr
integer(kind=MPI_KIND) :: mpirankMPI, mpiprocsMPI, mpierr
if (idx .le. 1) return
......
src/elpa_generalized/cannon.c
View file @ 05ca1675
......@@ -60,13 +60,26 @@
#include <math.h>
#include <complex.h>
#ifdef HAVE_64BIT_INTEGER_SUPPORT
#ifdef HAVE_64BIT_INTEGER_MATH_SUPPORT
#define C_INT_TYPE_PTR long int*
#define C_INT_TYPE long int
#define BLAS_KIND c_int64_t
#else
#define C_INT_TYPE_PTR int*
#define C_INT_TYPE int
#define BLAS_KIND c_int
#endif
#ifdef HAVE_64BIT_INTEGER_MPI_SUPPORT
#define C_INT_MPI_TYPE_PTR long int*
#define C_INT_MPI_TYPE long int
#define MPI_KIND c_int64_t
#else
#define C_INT_MPI_TYPE_PTR int*
#define C_INT_MPI_TYPE int
#define MPI_KIND c_int
#endif
// most of the file is not compiled if not using MPI
#ifdef WITH_MPI
......@@ -84,69 +97,37 @@
#undef REALCASE
/*
!f>#ifdef HAVE_64BIT_INTEGER_SUPPORT
!f> interface
!f> subroutine cannons_reduction_d(A, U, local_rowsCast, local_colsCast, a_desc, Res, toStore, row_comm, col_comm) &
!f> bind(C, name="cannons_reduction_c_d")
!f> use, intrinsic :: iso_c_binding
!f> use precision
!f> real(c_double) :: A(local_rowsCast, local_colsCast), U(local_rowsCast, local_colsCast)
!f> real(c_double) :: Res(local_rowsCast, local_colsCast)
!f> !type(c_ptr), value :: A, U, Res
!f> integer(kind=c_int64_t) :: a_desc(9)
!f> integer(kind=BLAS_KIND) :: a_desc(9)
!f> integer(kind=c_int),value :: local_rowsCast, local_colsCast
!f> integer(kind=c_int64_t),value :: row_comm, col_comm, ToStore
!f> integer(kind=MPI_KIND),value :: row_comm, col_comm, ToStore
!f> end subroutine
!f> end interface
!f>#endif
!f>#ifndef HAVE_64BIT_INTEGER_SUPPORT
!f> interface
!f> subroutine cannons_reduction_d(A, U, local_rowsCast, local_colsCast, a_desc, Res, toStore, row_comm, col_comm) &
!f> bind(C, name="cannons_reduction_c_d")
!f> use, intrinsic :: iso_c_binding
!f> real(c_double) :: A(local_rowsCast, local_colsCast), U(local_rowsCast, local_colsCast)
!f> real(c_double) :: Res(local_rowsCast, local_colsCast)
!f> !type(c_ptr), value :: A, U, Res
!f> integer(kind=c_int) :: a_desc(9)
!f> integer(kind=c_int),value :: local_rowsCast, local_colsCast
!f> integer(kind=c_int),value :: row_comm, col_comm, ToStore
!f> end subroutine
!f> end interface
!f>#endif
*/
void cannons_reduction_c_d(double* A, double* U, int local_rowsCast, int local_colsCast, C_INT_TYPE_PTR a_desc,
double *Res, C_INT_TYPE ToStore, C_INT_TYPE row_comm, C_INT_TYPE col_comm);
double *Res, C_INT_MPI_TYPE ToStore, C_INT_MPI_TYPE row_comm, C_INT_MPI_TYPE col_comm);
/*
!f>#ifdef HAVE_64BIT_INTEGER_SUPPORT
!f> interface
!f> subroutine cannons_triang_rectangular_d(U, B, local_rowsCast, local_colsCast, u_desc, b_desc, Res, row_comm, col_comm) &
!f> bind(C, name="cannons_triang_rectangular_c_d")
!f> use, intrinsic :: iso_c_binding
!f> use precision
!f> real(c_double) :: U(local_rowsCast, local_colsCast), B(local_rowsCast, local_colsCast)
!f> real(c_double) :: Res(local_rowsCast, local_colsCast)
!f> integer(kind=c_int64_t) :: u_desc(9), b_desc(9)
!f> integer(kind=BLAS_KIND) :: u_desc(9), b_desc(9)
!f> integer(kind=c_int),value :: local_rowsCast, local_colsCast
!f> integer(kind=c_int64_t),value :: row_comm, col_comm
!f> end subroutine
!f> end interface
!f>#endif
!f>#ifndef HAVE_64BIT_INTEGER_SUPPORT
!f> interface
!f> subroutine cannons_triang_rectangular_d(U, B, local_rowsCast, local_colsCast, u_desc, b_desc, Res, row_comm, col_comm) &
!f> bind(C, name="cannons_triang_rectangular_c_d")
!f> use, intrinsic :: iso_c_binding
!f> real(c_double) :: U(local_rowsCast, local_colsCast), B(local_rowsCast, local_colsCast)
!f> real(c_double) :: Res(local_rowsCast, local_colsCast)
!f> integer(kind=c_int) :: u_desc(9), b_desc(9)
!f> integer(kind=c_int),value :: local_rowsCast, local_colsCast
!f> integer(kind=c_int),value :: row_comm, col_comm
!f> integer(kind=MPI_KIND),value :: row_comm, col_comm
!f> end subroutine
!f> end interface
!f>#endif
*/
void cannons_triang_rectangular_c_d(double* U, double* B, int local_rowsCast, int local_colsCast,
C_INT_TYPE_PTR u_desc, C_INT_TYPE_PTR b_desc, double *Res, C_INT_TYPE row_comm, C_INT_TYPE col_comm);
C_INT_TYPE_PTR u_desc, C_INT_TYPE_PTR b_desc, double *Res, C_INT_MPI_TYPE row_comm, C_INT_MPI_TYPE col_comm);
//***********************************************************************************************************
......@@ -159,68 +140,37 @@ void cannons_triang_rectangular_c_d(double* U, double* B, int local_rowsCast, in
#undef REALCASE
/*
!f>#ifdef HAVE_64BIT_INTEGER_SUPPORT
!f> interface
!f> subroutine cannons_reduction_f(A, U, local_rowsCast, local_colsCast, a_desc, Res, toStore, row_comm, col_comm) &
!f> bind(C, name="cannons_reduction_c_f")
!f> use, intrinsic :: iso_c_binding
!f> use precision
!f> real(c_float) :: A(local_rowsCast, local_colsCast), U(local_rowsCast, local_colsCast)
!f> real(c_float) :: Res(local_rowsCast, local_colsCast)
!f> !type(c_ptr), value :: A, U, Res
!f> integer(kind=c_int64_t) :: a_desc(9)
!f> integer(kind=BLAS_KIND) :: a_desc(9)
!f> integer(kind=c_int),value :: local_rowsCast, local_colsCast
!f> integer(kind=c_int64_t),value :: row_comm, col_comm, ToStore
!f> integer(kind=MPI_KIND),value :: row_comm, col_comm, ToStore
!f> end subroutine
!f> end interface
!f>#endif
!f>#ifndef HAVE_64BIT_INTEGER_SUPPORT
!f> interface
!f> subroutine cannons_reduction_f(A, U, local_rowsCast, local_colsCast, a_desc, Res, toStore, row_comm, col_comm) &
!f> bind(C, name="cannons_reduction_c_f")
!f> use, intrinsic :: iso_c_binding
!f> real(c_float) :: A(local_rowsCast, local_colsCast), U(local_rowsCast, local_colsCast)
!f> real(c_float) :: Res(local_rowsCast, local_colsCast)
!f> !type(c_ptr), value :: A, U, Res
!f> integer(kind=c_int) :: a_desc(9)
!f> integer(kind=c_int),value :: local_rowsCast, local_colsCast
!f> integer(kind=c_int),value :: row_comm, col_comm, ToStore
!f> end subroutine
!f> end interface
!f>#endif
*/
void cannons_reduction_c_f(float* A, float* U, int local_rowsCast, int local_colsCast, C_INT_TYPE_PTR a_desc,
float *Res, C_INT_TYPE ToStore, C_INT_TYPE row_comm, C_INT_TYPE col_comm);
float *Res, C_INT_MPI_TYPE ToStore, C_INT_MPI_TYPE row_comm, C_INT_MPI_TYPE col_comm);
/*
!f>#ifdef HAVE_64BIT_INTEGER_SUPPORT
!f> interface
!f> subroutine cannons_triang_rectangular_f(U, B, local_rowsCast, local_colsCast, u_desc, b_desc, Res, row_comm, col_comm) &
!f> bind(C, name="cannons_triang_rectangular_c_f")
!f> use, intrinsic :: iso_c_binding
!f> real(c_float) :: U(local_rowsCast, local_colsCast), B(local_rowsCast, local_colsCast)
!f> real(c_float) :: Res(local_rowsCast, local_colsCast)
!f> integer(kind=c_int64_t) :: u_desc(9), b_desc(9)
!f> integer(kind=c_int),value :: local_rowsCast, local_colsCast
!f> integer(kind=c_int64_t),value :: row_comm, col_comm
!f> end subroutine
!f> end interface
!f>#endif
!f>#ifndef HAVE_64BIT_INTEGER_SUPPORT
!f> interface
!f> subroutine cannons_triang_rectangular_f(U, B, local_rowsCast, local_colsCast, u_desc, b_desc, Res, row_comm, col_comm) &
!f> bind(C, name="cannons_triang_rectangular_c_f")
!f> use, intrinsic :: iso_c_binding
!f> use precision
!f> real(c_float) :: U(local_rowsCast, local_colsCast), B(local_rowsCast, local_colsCast)
!f> real(c_float) :: Res(local_rowsCast, local_colsCast)
!f> integer(kind=c_int) :: u_desc(9), b_desc(9)
!f> integer(kind=BLAS_KIND) :: u_desc(9), b_desc(9)
!f> integer(kind=c_int),value :: local_rowsCast, local_colsCast
!f> integer(kind=c_int),value :: row_comm, col_comm
!f> integer(kind=MPI_KIND),value :: row_comm, col_comm
!f> end subroutine
!f> end interface
!f>#endif
*/
void cannons_triang_rectangular_c_f(float* U, float* B, int local_rowsCast, int local_colsCast,
C_INT_TYPE_PTR u_desc, C_INT_TYPE_PTR b_desc, float *Res, C_INT_TYPE row_comm, C_INT_TYPE col_comm);
C_INT_TYPE_PTR u_desc, C_INT_TYPE_PTR b_desc, float *Res, C_INT_MPI_TYPE row_comm, C_INT_MPI_TYPE col_comm);
//***********************************************************************************************************
......@@ -233,68 +183,36 @@ void cannons_triang_rectangular_c_f(float* U, float* B, int local_rowsCast, int
#undef COMPLEXCASE
/*
!f>#ifdef HAVE_64BIT_INTEGER_SUPPORT
!f> interface
!f> subroutine cannons_reduction_dc(A, U, local_rowsCast, local_colsCast, a_desc, Res, toStore, row_comm, col_comm) &
!f> bind(C, name="cannons_reduction_c_dc")
!f> use, intrinsic :: iso_c_binding
!f> use precision
!f> complex(c_double) :: A(local_rowsCast, local_colsCast), U(local_rowsCast, local_colsCast)
!f> complex(c_double) :: Res(local_rowsCast, local_colsCast)
!f> !type(c_ptr), value :: A, U, Res
!f> integer(kind=c_int64_t) :: a_desc(9)
!f> integer(kind=BLAS_KIND) :: a_desc(9)
!f> integer(kind=c_int),value :: local_rowsCast, local_colsCast
!f> integer(kind=c_int64_t),value :: row_comm, col_comm, ToStore
!f> end subroutine
!f> end interface
!f>#endif
!f>#ifndef HAVE_64BIT_INTEGER_SUPPORT
!f> interface
!f> subroutine cannons_reduction_dc(A, U, local_rowsCast, local_colsCast, a_desc, Res, toStore, row_comm, col_comm) &
!f> bind(C, name="cannons_reduction_c_dc")
!f> use, intrinsic :: iso_c_binding
!f> complex(c_double) :: A(local_rowsCast, local_colsCast), U(local_rowsCast, local_colsCast)
!f> complex(c_double) :: Res(local_rowsCast, local_colsCast)
!f> !type(c_ptr), value :: A, U, Res
!f> integer(kind=c_int) :: a_desc(9)
!f> integer(kind=c_int),value :: local_rowsCast, local_colsCast
!f> integer(kind=c_int),value :: row_comm, col_comm, ToStore
!f> integer(kind=MPI_KIND),value :: row_comm, col_comm, ToStore
!f> end subroutine
!f> end interface
!f>#endif
*/
void cannons_reduction_c_dc(double complex* A, double complex* U, int local_rowsCast, int local_colsCasr, C_INT_TYPE_PTR a_desc,
double complex *Res, C_INT_TYPE ToStore, C_INT_TYPE row_comm, C_INT_TYPE col_comm);
double complex *Res, C_INT_MPI_TYPE ToStore, C_INT_MPI_TYPE row_comm, C_INT_MPI_TYPE col_comm);
/*
!f>#ifdef HAVE_64BIT_INTEGER_SUPPORT
!f> interface
!f> subroutine cannons_triang_rectangular_dc(U, B, local_rowsCast, local_colsCast, u_desc, b_desc, Res, row_comm, col_comm) &
!f> bind(C, name="cannons_triang_rectangular_c_dc")
!f> use, intrinsic :: iso_c_binding
!f> complex(c_double) :: U(local_rowsCast, local_colsCast), B(local_rowsCast, local_colsCast)
!f> complex(c_double) :: Res(local_rowsCast, local_colsCast)
!f> integer(kind=c_int64_t) :: u_desc(9), b_desc(9)
!f> integer(kind=c_int),value :: local_rowsCast, local_colsCast
!f> integer(kind=c_int64_t),value :: row_comm, col_comm
!f> end subroutine
!f> end interface
!f>#endif
!f>#ifndef HAVE_64BIT_INTEGER_SUPPORT
!f> interface
!f> subroutine cannons_triang_rectangular_dc(U, B, local_rowsCast, local_colsCast, u_desc, b_desc, Res, row_comm, col_comm) &
!f> bind(C, name="cannons_triang_rectangular_c_dc")
!f> use, intrinsic :: iso_c_binding
!f> use precision
!f> complex(c_double) :: U(local_rowsCast, local_colsCast), B(local_rowsCast, local_colsCast)
!f> complex(c_double) :: Res(local_rowsCast, local_colsCast)
!f> integer(kind=c_int) :: u_desc(9), b_desc(9)
!f> integer(kind=BLAS_KIND) :: u_desc(9), b_desc(9)
!f> integer(kind=c_int),value :: local_rowsCast, local_colsCast
!f> integer(kind=c_int),value :: row_comm, col_comm
!f> integer(kind=MPI_KIND),value :: row_comm, col_comm
!f> end subroutine
!f> end interface
!f>#endif
*/
void cannons_triang_rectangular_c_dc(double complex* U, double complex* B, int local_rowsCast, int local_colsCast,
C_INT_TYPE_PTR u_desc, C_INT_TYPE_PTR b_desc, double complex *Res, C_INT_TYPE row_comm, C_INT_TYPE col_comm);
C_INT_TYPE_PTR u_desc, C_INT_TYPE_PTR b_desc, double complex *Res, C_INT_MPI_TYPE row_comm, C_INT_MPI_TYPE col_comm);
//***********************************************************************************************************
#define COMPLEXCASE 1
......@@ -306,66 +224,36 @@ void cannons_triang_rectangular_c_dc(double complex* U, double complex* B, int l
#undef COMPLEXCASE
/*
!f>#ifdef HAVE_64BIT_INTEGER_SUPPORT
!f> interface
!f> subroutine cannons_reduction_fc(A, U, local_rowsCast, local_colsCast, a_desc, Res, toStore, row_comm, col_comm) &
!f> bind(C, name="cannons_reduction_c_fc")
!f> use, intrinsic :: iso_c_binding
!f> use precision
!f> complex(c_float) :: A(local_rowsCast, local_colsCast), U(local_rowsCast, local_colsCast)
!f> complex(c_float) :: Res(local_rowsCast, local_colsCast)
!f> !type(c_ptr), value :: A, U, Res
!f> integer(kind=c_int64_t) :: a_desc(9)
!f> integer(kind=BLAS_KIND) :: a_desc(9)
!f> integer(kind=c_int),value :: local_rowsCast, local_colsCast
!f> integer(kind=c_int64_t),value :: row_comm, col_comm, ToStore
!f> integer(kind=MPI_KIND),value :: row_comm, col_comm, ToStore
!f> end subroutine
!f> end interface
!f>#endif
!f>#ifndef HAVE_64BIT_INTEGER_SUPPORT
!f> interface
!f> subroutine cannons_reduction_fc(A, U, local_rowsCast, local_colsCast, a_desc, Res, toStore, row_comm, col_comm) &
!f> bind(C, name="cannons_reduction_c_fc")
!f> use, intrinsic :: iso_c_binding
!f> complex(c_float) :: A(local_rowsCast, local_colsCast), U(local_rowsCast, local_colsCast)
!f> complex(c_float) :: Res(local_rowsCast, local_colsCast)
!f> !type(c_ptr), value :: A, U, Res
!f> integer(kind=c_int) :: a_desc(9)
!f> integer(kind=c_int),value :: local_rowsCast, local_colsCast
!f> integer(kind=c_int),value :: row_comm, col_comm, ToStore
!f> end subroutine
!f> end interface
!f>#endif
*/
void cannons_reduction_c_fc(float complex* A, float complex* U, int local_rowsCast, int local_colsCast, C_INT_TYPE_PTR a_desc,
float complex *Res, C_INT_TYPE ToStore, C_INT_TYPE row_comm, C_INT_TYPE col_comm);
float complex *Res, C_INT_MPI_TYPE ToStore, C_INT_MPI_TYPE row_comm, C_INT_MPI_TYPE col_comm);
/*
!f>#ifdef HAVE_64BIT_INTEGER_SUPPORT
!f> interface
!f> subroutine cannons_triang_rectangular_fc(U, B, local_rowsCast, local_colsCast, u_desc, b_desc, Res, row_comm, col_comm) &
!f> bind(C, name="cannons_triang_rectangular_c_fc")
!f> use, intrinsic :: iso_c_binding
!f> use precision
!f> complex(c_float) :: U(local_rowsCast, local_colsCast), B(local_rowsCast, local_colsCast)
!f> complex(c_float) :: Res(local_rowsCast, local_colsCast)
!f> integer(kind=c_int64_t) :: u_desc(9), b_desc(9)
!f> integer(kind=BLAS_KIND) :: u_desc(9), b_desc(9)
!f> integer(kind=c_int),value :: local_rowsCast, local_colsCast
!f> integer(kind=c_int64_t),value :: row_comm, col_comm
!f> end subroutine
!f> end interface
!f>#endif
!f>#ifndef HAVE_64BIT_INTEGER_SUPPORT
!f> interface
!f> subroutine cannons_triang_rectangular_fc(U, B, local_rowsCast, local_colsCast, u_desc, b_desc, Res, row_comm, col_comm) &
!f> bind(C, name="cannons_triang_rectangular_c_fc")
!f> use, intrinsic :: iso_c_binding
!f> complex(c_float) :: U(local_rowsCast, local_colsCast), B(local_rowsCast, local_colsCast)
!f> complex(c_float) :: Res(local_rowsCast, local_colsCast)
!f> integer(kind=c_int) :: u_desc(9), b_desc(9)
!f> integer(kind=c_int),value :: local_rowsCast, local_colsCast
!f> integer(kind=c_int),value :: row_comm, col_comm
!f> integer(kind=MPI_KIND),value :: row_comm, col_comm
!f> end subroutine
!f> end interface
!f>#endif
*/
void cannons_triang_rectangular_c_fc(float complex* U, float complex* B, int local_rowsCast, int local_colsCast,
C_INT_TYPE_PTR u_desc, C_INT_TYPE_PTR b_desc, float complex *Res, C_INT_TYPE row_comm, C_INT_TYPE col_comm);
C_INT_TYPE_PTR u_desc, C_INT_TYPE_PTR b_desc, float complex *Res, C_INT_MPI_TYPE row_comm, C_INT_MPI_TYPE col_comm);
#endif
src/elpa_generalized/cannon_back_template.c
View file @ 05ca1675
......@@ -54,16 +54,24 @@
// Author: Valeriy Manin (Bergische Universität Wuppertal)
// integreated into the ELPA library Pavel Kus, Andeas Marek (MPCDF)
#ifdef HAVE_64BIT_INTEGER_SUPPORT
#ifdef HAVE_64BIT_INTEGER_MATH_SUPPORT
#define C_INT_TYPE_PTR long int*
#define C_INT_TYPE long int
#define FORTRAN_INT_TYPE c_int64_t
#define BLAS_KIND c_int64_t
#else
#define C_INT_TYPE_PTR int*
#define C_INT_TYPE int
#define FORTRAN_INT_TYPE c_int
#define BLAS_KIND c_int
#endif
#ifdef HAVE_64BIT_INTEGER_MPI_SUPPORT
#define C_INT_MPI_TYPE_PTR long int*
#define C_INT_MPI_TYPE long int
#define MPI_KIND c_int64_t
#else
#define C_INT_MPI_TYPE_PTR int*
#define C_INT_MPI_TYPE int
#define MPI_KIND c_int
#endif
// it seems, that we need those two levels of indirection to correctly expand macros
#define cannons_triang_rectangular_impl_expand2(SUFFIX) cannons_triang_rectangular_##SUFFIX
......@@ -88,6 +96,7 @@ void cannons_triang_rectangular_impl(math_type* U, math_type* B, C_INT_TYPE np_r
C_INT_TYPE na, nb, nblk, width, na_rows, na_cols, nb_cols, cols_in_buffer_U_my_initial, cols_in_buffer_U, rows_in_buffer_U, Size_receive_U_now, rows_in_buffer_U_now, cols_in_buffer_U_now, rows_in_buffer_U_my_initial;
C_INT_MPI_TYPE Size_receive_U_nowMPI, Size_receive_UMPI, Size_receive_BMPI;
C_INT_TYPE i, j, Size_send_U, Size_receive_U, Size_send_B, Size_receive_B, intNumber, Buf_rows, Buf_cols_U, Buf_cols_B, curr_rows, num_of_iters, cols_in_buffer, rows_in_block, curr_col_loc, cols_in_block, num_of_blocks_in_U_buffer, col_of_origin_U, b_rows_mult, b_cols_mult;
math_type *Buf_to_send_U, *Buf_to_receive_U, *Buf_to_send_B, *Buf_to_receive_B, *Buf_U, *PosBuff;
......@@ -241,9 +250,10 @@ void cannons_triang_rectangular_impl(math_type* U, math_type* B, C_INT_TYPE np_r
{
if(where_to_send_U != my_pcol) // if I need to send and receive on this step
{
MPI_Sendrecv(Buf_to_send_U, Size_send_U, MPI_MATH_DATATYPE_PRECISION_C, where_to_send_U, 0, Buf_U, Size_U_stored, MPI_MATH_DATATYPE_PRECISION_C, from_where_to_receive_U, 0, row_comm, &status);
MPI_Get_count(&status, MPI_MATH_DATATYPE_PRECISION_C, &Size_receive_U_now);
Size_receive_U = Size_receive_U + Size_receive_U_now - 2; // we need only number of elements, so exclude information about cols_in_buffer_U and rows_in_buffer_U
MPI_Sendrecv(Buf_to_send_U, (C_INT_MPI_TYPE) Size_send_U, MPI_MATH_DATATYPE_PRECISION_C, (C_INT_MPI_TYPE) where_to_send_U, 0, Buf_U, (C_INT_MPI_TYPE) Size_U_stored, MPI_MATH_DATATYPE_PRECISION_C, (C_INT_MPI_TYPE) from_where_to_receive_U, 0, row_comm, &status);
MPI_Get_count(&status, MPI_MATH_DATATYPE_PRECISION_C, &Size_receive_U_nowMPI);
Size_receive_U_now = (C_INT_TYPE) Size_receive_U_nowMPI;
Size_receive_U = Size_receive_U + Size_receive_U_now - 2; // we need only number of elements, so exclude information about cols_in_buffer_U and rows_in_buffer_
cols_in_buffer_U_now = Buf_U[Size_receive_U_now - 2];
cols_in_buffer_U = cols_in_buffer_U + cols_in_buffer_U_now;
......@@ -309,8 +319,10 @@ void cannons_triang_rectangular_impl(math_type* U, math_type* B, C_INT_TYPE np_r
{
if(my_prow > 0)
{
MPI_Sendrecv(Buf_to_send_U, Size_send_U, MPI_MATH_DATATYPE_PRECISION_C, where_to_send_U, 0, Buf_to_receive_U, Size_U_stored, MPI_MATH_DATATYPE_PRECISION_C, from_where_to_receive_U, 0, row_comm, &status);
MPI_Get_count(&status, MPI_MATH_DATATYPE_PRECISION_C, &Size_receive_U);
MPI_Sendrecv(Buf_to_send_U, (C_INT_MPI_TYPE) Size_send_U, MPI_MATH_DATATYPE_PRECISION_C, (C_INT_MPI_TYPE) where_to_send_U, 0, Buf_to_receive_U, (C_INT_MPI_TYPE) Size_U_stored, MPI_MATH_DATATYPE_PRECISION_C, (C_INT_MPI_TYPE) from_where_to_receive_U, 0, row_comm, &status);
MPI_Get_count(&status, MPI_MATH_DATATYPE_PRECISION_C, &Size_receive_UMPI);
Size_receive_U = (C_INT_TYPE) Size_receive_UMPI;
cols_in_buffer_U = (C_INT_TYPE)Buf_to_receive_U[Size_receive_U-2];
rows_in_buffer_U = (C_INT_TYPE)Buf_to_receive_U[Size_receive_U-1];
}
......@@ -341,9 +353,11 @@ void cannons_triang_rectangular_impl(math_type* U, math_type* B, C_INT_TYPE np_r
{
// form array to send
C_LACPY("A", &na_rows, &nb_cols, B, &na_rows, Buf_to_send_B, &na_rows);
MPI_Sendrecv(Buf_to_send_B, nb_cols*na_rows, MPI_MATH_DATATYPE_PRECISION_C, where_to_send_B, 0, Buf_to_receive_B, nb_cols*Buf_rows, MPI_MATH_DATATYPE_PRECISION_C, from_where_to_receive_B, 0, col_comm, &status);
MPI_Get_count(&status, MPI_MATH_DATATYPE_PRECISION_C, &Size_receive_B); // find out how many elements I have received
MPI_Sendrecv(Buf_to_send_B, (C_INT_MPI_TYPE) (nb_cols*na_rows), MPI_MATH_DATATYPE_PRECISION_C, (C_INT_MPI_TYPE) where_to_send_B, 0, Buf_to_receive_B, (C_INT_MPI_TYPE) (nb_cols*Buf_rows), MPI_MATH_DATATYPE_PRECISION_C, (C_INT_MPI_TYPE) from_where_to_receive_B, 0, col_comm, &status);
MPI_Get_count(&status, MPI_MATH_DATATYPE_PRECISION_C, &Size_receive_BMPI); // find out how many elements I have received
Size_receive_B = (C_INT_TYPE) Size_receive_BMPI;
Size_receive_B = Size_receive_B/nb_cols; // how many rows I have received
}
else
{
......@@ -378,13 +392,11 @@ void cannons_triang_rectangular_impl(math_type* U, math_type* B, C_INT_TYPE np_r
Size_send_B = Size_receive_B;
///// shift for U ////////////////////////////////////////////////////////////
MPI_Isend(Buf_to_send_U, Size_send_U, MPI_MATH_DATATYPE_PRECISION_C, where_to_send_U, 0, row_comm, &request_U_Send);
MPI_Irecv(Buf_to_receive_U, ratio*Size_U_stored, MPI_MATH_DATATYPE_PRECISION_C, from_where_to_receive_U, 0, row_comm, &request_U_Recv);
MPI_Isend(Buf_to_send_U, (C_INT_MPI_TYPE) Size_send_U, MPI_MATH_DATATYPE_PRECISION_C, (C_INT_MPI_TYPE) where_to_send_U, 0, row_comm, &request_U_Send);
MPI_Irecv(Buf_to_receive_U, (C_INT_MPI_TYPE) (ratio*Size_U_stored), MPI_MATH_DATATYPE_PRECISION_C, (C_INT_MPI_TYPE) from_where_to_receive_U, 0, row_comm, &request_U_Recv);
///// shift for B /////////////////////////////////////////////
MPI_Isend(Buf_to_send_B, Size_send_B*nb_cols, MPI_MATH_DATATYPE_PRECISION_C, where_to_send_B, 0, col_comm, &request_B_Send);
MPI_Irecv(Buf_to_receive_B, Buf_rows*nb_cols, MPI_MATH_DATATYPE_PRECISION_C, from_where_to_receive_B, 0, col_comm, &request_B_Recv);
MPI_Isend(Buf_to_send_B, (C_INT_MPI_TYPE) (Size_send_B*nb_cols), MPI_MATH_DATATYPE_PRECISION_C, (C_INT_MPI_TYPE) where_to_send_B, 0, col_comm, &request_B_Send);
MPI_Irecv(Buf_to_receive_B, (C_INT_MPI_TYPE) (Buf_rows*nb_cols), MPI_MATH_DATATYPE_PRECISION_C, (C_INT_MPI_TYPE) from_where_to_receive_B, 0, col_comm, &request_B_Recv);
///// multiplication ////////////////////////////////////////////////////////////////////////////////////////////
cols_in_buffer_U = (C_INT_TYPE)Buf_to_send_U[Size_receive_U-2];
rows_in_buffer_U = (C_INT_TYPE)Buf_to_send_U[Size_receive_U-1];
......@@ -426,12 +438,15 @@ void cannons_triang_rectangular_impl(math_type* U, math_type* B, C_INT_TYPE np_r
MPI_Wait(&request_U_Send, &status);
MPI_Wait(&request_U_Recv, &status);
MPI_Get_count(&status, MPI_MATH_DATATYPE_PRECISION_C, &Size_receive_U); // find out how many elements I have received
MPI_Get_count(&status, MPI_MATH_DATATYPE_PRECISION_C, &Size_receive_UMPI); // find out how many elements I have received
Size_receive_U = (C_INT_TYPE) Size_receive_UMPI;
MPI_Wait(&request_B_Send, &status);
MPI_Wait(&request_B_Recv, &status);
MPI_Get_count(&status, MPI_MATH_DATATYPE_PRECISION_C, &Size_receive_B); // find out how many elements I have received
Size_receive_B = Size_receive_B/nb_cols; // how many rows I have received
MPI_Get_count(&status, MPI_MATH_DATATYPE_PRECISION_C, &Size_receive_BMPI); // find out how many elements I have received
Size_receive_B = (C_INT_TYPE) Size_receive_BMPI;
Size_receive_B = (C_INT_TYPE) Size_receive_B / nb_cols; // how many rows I have received
}
// last iteration
......@@ -483,7 +498,7 @@ void cannons_triang_rectangular_impl(math_type* U, math_type* B, C_INT_TYPE np_r
void cannons_triang_rectangular_c_impl(math_type* U, math_type* B, int local_rowsCast, int local_colsCast,
C_INT_TYPE_PTR u_desc, C_INT_TYPE_PTR b_desc, math_type *Res, C_INT_TYPE row_comm, C_INT_TYPE col_comm)
C_INT_TYPE_PTR u_desc, C_INT_TYPE_PTR b_desc, math_type *Res, C_INT_MPI_TYPE row_comm, C_INT_MPI_TYPE col_comm)
{
C_INT_TYPE local_rows, local_cols;
......@@ -494,11 +509,17 @@ void cannons_triang_rectangular_c_impl(math_type* U, math_type* B, int local_row
MPI_Comm c_col_comm = MPI_Comm_f2c(col_comm);
C_INT_TYPE my_prow, my_pcol, np_rows, np_cols;
MPI_Comm_rank(c_row_comm, &my_prow);
MPI_Comm_size(c_row_comm, &np_rows);
MPI_Comm_rank(c_col_comm, &my_pcol);