From a6847f46555c7946261d3af585c0118f1a5ddd25 Mon Sep 17 00:00:00 2001
From: Mathieu Faverge <mathieu.faverge@inria.fr>
Date: Wed, 20 Jun 2018 20:38:28 +0200
Subject: [PATCH 01/22] Add fortran pkg-config file
---
cmake_modules/GenSPMPkgConfig.cmake | 52 +++++++++++++++++++----------
tools/spmf.pc.in | 23 +++++++++++++
2 files changed, 58 insertions(+), 17 deletions(-)
create mode 100644 tools/spmf.pc.in
diff --git a/cmake_modules/GenSPMPkgConfig.cmake b/cmake_modules/GenSPMPkgConfig.cmake
index 935b432f..33b8e79f 100644
--- a/cmake_modules/GenSPMPkgConfig.cmake
+++ b/cmake_modules/GenSPMPkgConfig.cmake
@@ -80,23 +80,41 @@
###
macro(spm_generate_pkgconfig_file)
- set(SPM_PKGCONFIG_LIBS "-lspm")
- set(SPM_PKGCONFIG_LIBS_PRIVATE "-lm")
- set(SPM_PKGCONFIG_REQUIRED "")
- set(SPM_PKGCONFIG_REQUIRED_PRIVATE "")
-
- #clean_lib_list(SPM)
-
- set(_output_spm_file "${CMAKE_BINARY_DIR}/spm.pc")
- configure_file(
- "${CMAKE_CURRENT_SOURCE_DIR}/tools/spm.pc.in"
- "${_output_spm_file}"
- @ONLY
- )
- install(
- FILES ${_output_spm_file}
- DESTINATION lib/pkgconfig
- )
+ # The link flags specific to this package and any required libraries
+ # that don't support PkgConfig
+ set(SPM_PKGCONFIG_LIBS spm)
+
+ # The link flags for private libraries required by this package but not
+ # exposed to applications
+ set(SPM_PKGCONFIG_LIBS_PRIVATE "")
+
+ # A list of packages required by this package
+ set(SPM_PKGCONFIG_REQUIRED "")
+
+ # A list of private packages required by this package but not exposed to
+ # applications
+ set(SPM_PKGCONFIG_REQUIRED_PRIVATE "")
+
+ # Define required package
+ # -----------------------
+ clean_lib_list(SPM)
+
+ # Create .pc file
+ # ---------------
+ configure_file(
+ "${CMAKE_CURRENT_SOURCE_DIR}/tools/spm.pc.in"
+ "${CMAKE_BINARY_DIR}/lib/pkgconfig/spm.pc" @ONLY)
+
+ configure_file(
+ "${CMAKE_CURRENT_SOURCE_DIR}/tools/spmf.pc.in"
+ "${CMAKE_BINARY_DIR}/lib/pkgconfig/spmf.pc" @ONLY)
+
+ # installation
+ # ------------
+ install(FILES
+ "${CMAKE_BINARY_DIR}/lib/pkgconfig/spm.pc"
+ "${CMAKE_BINARY_DIR}/lib/pkgconfig/spmf.pc"
+ DESTINATION lib/pkgconfig)
endmacro(spm_generate_pkgconfig_file)
diff --git a/tools/spmf.pc.in b/tools/spmf.pc.in
new file mode 100644
index 00000000..c368a6d7
--- /dev/null
+++ b/tools/spmf.pc.in
@@ -0,0 +1,23 @@
+#
+# @file spmf.pc
+#
+# @copyright 2016-2018 Bordeaux INP, CNRS (LaBRI UMR 5800), Inria,
+# Univ. Bordeaux. All rights reserved.
+#
+# @version 1.0.0
+# @author Mathieu Faverge
+# @date 2017-06-24
+#
+prefix=@CMAKE_INSTALL_PREFIX@
+exec_prefix=${prefix}
+libdir=${exec_prefix}/lib
+includedir=${exec_prefix}/include
+
+Name: SPM - Fortran
+Description: SParse Matrix package - Fortran interface
+Version: @SPM_VERSION_MAJOR@.@SPM_VERSION_MINOR@.@SPM_VERSION_MICRO@
+Cflags: -I${includedir}
+Libs: -L${libdir} -lspmf @SPM_PKGCONFIG_LIBS@
+Libs.private: @SPM_PKGCONFIG_LIBS_PRIVATE@
+Requires: spm @SPM_PKGCONFIG_REQUIRED@
+Requires.private: @SPM_PKGCONFIG_REQUIRED_PRIVATE@
--
GitLab
From 64b6a9bbba917ec1b205228eb09a1b0336bb279f Mon Sep 17 00:00:00 2001
From: Mathieu Faverge <mathieu.faverge@inria.fr>
Date: Wed, 20 Jun 2018 20:39:53 +0200
Subject: [PATCH 02/22] Fix issue with norm naming for forward error
---
src/z_spm_genrhs.c | 15 +++++++++------
1 file changed, 9 insertions(+), 6 deletions(-)
diff --git a/src/z_spm_genrhs.c b/src/z_spm_genrhs.c
index e5b4783d..76f3a248 100644
--- a/src/z_spm_genrhs.c
+++ b/src/z_spm_genrhs.c
@@ -451,7 +451,7 @@ z_spmCheckAxb( spm_fixdbl_t eps, int nrhs,
*/
if ( x0 != NULL ) {
double normX0;
- double forw, nr, nx;
+ double forw, nr, nx, nx0;
int fail;
forward = 0.;
@@ -461,14 +461,15 @@ z_spmCheckAxb( spm_fixdbl_t eps, int nrhs,
for( i=0; i<nrhs; i++, zx += ldx, zx0 += ldx0 ) {
- nx = LAPACKE_zlange( LAPACK_COL_MAJOR, 'I', spm->n, 1, zx0, ldx0 );
+ nx0 = LAPACKE_zlange( LAPACK_COL_MAJOR, 'I', spm->n, 1, zx0, ldx0 );
+ nx = LAPACKE_zlange( LAPACK_COL_MAJOR, 'I', spm->n, 1, zx, ldx );
cblas_zaxpy( spm->n, CBLAS_SADDR(mzone),
zx, 1, zx0, 1);
nr = LAPACKE_zlange( LAPACK_COL_MAJOR, 'I', spm->n, 1, zx0, ldx0 );
- forw = (nr / nx) / eps;
+ forw = (nr / nx0) / eps;
normX0 = ( nx > normX0 ) ? nx : normX0;
normR = ( nr > normR ) ? nr : normR;
@@ -477,9 +478,11 @@ z_spmCheckAxb( spm_fixdbl_t eps, int nrhs,
fail = isnan(nx) || isinf(nx) || isnan(forw) || isinf(forw) || (forw > 1.e2);
if ( fail ) {
fprintf( stdout,
- " || x0_%d ||_oo %e\n"
+ " || x_%d ||_oo %e\n"
+ " || x0_%d - x_%d ||_oo %e\n"
" || x0_%d - x_%d ||_oo / (||x0_%d||_oo * eps) %e (%s)\n",
- i, nr,
+ i, nx,
+ i, i, nr,
i, i, i, forw,
fail ? "FAILED" : "SUCCESS" );
}
@@ -488,7 +491,7 @@ z_spmCheckAxb( spm_fixdbl_t eps, int nrhs,
}
fprintf( stdout,
- " max(|| x0_i ||_oo) %e\n"
+ " max(|| x_i ||_oo) %e\n"
" max(|| x0_i - x_i ||_oo) %e\n"
" max(|| x0_i - x_i ||_oo / || x0_i ||_oo) %e (%s)\n",
normX0, normR, forward,
--
GitLab
From 8efaadfb0824997d14065117ac9524f582be3504 Mon Sep 17 00:00:00 2001
From: Mathieu Faverge <mathieu.faverge@inria.fr>
Date: Wed, 20 Jun 2018 20:49:19 +0200
Subject: [PATCH 03/22] Remove the call to clean for now
---
cmake_modules/GenSPMPkgConfig.cmake | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/cmake_modules/GenSPMPkgConfig.cmake b/cmake_modules/GenSPMPkgConfig.cmake
index 33b8e79f..c8780307 100644
--- a/cmake_modules/GenSPMPkgConfig.cmake
+++ b/cmake_modules/GenSPMPkgConfig.cmake
@@ -97,7 +97,7 @@ macro(spm_generate_pkgconfig_file)
# Define required package
# -----------------------
- clean_lib_list(SPM)
+ #clean_lib_list(SPM)
# Create .pc file
# ---------------
--
GitLab
From 38fcf0da1311934421f02f299db8d5e17f8110e6 Mon Sep 17 00:00:00 2001
From: Mathieu Faverge <mathieu.faverge@inria.fr>
Date: Thu, 21 Jun 2018 04:47:20 +0200
Subject: [PATCH 04/22] Remove the GenSpmPkgConfig file for the generic one
---
CMakeLists.txt | 14 ++-
cmake_modules/GenSPMPkgConfig.cmake | 143 ----------------------------
2 files changed, 11 insertions(+), 146 deletions(-)
delete mode 100644 cmake_modules/GenSPMPkgConfig.cmake
diff --git a/CMakeLists.txt b/CMakeLists.txt
index ee61de10..8c904a5f 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -249,9 +249,17 @@ install(FILES
DESTINATION include )
### Build pkg-config and environment file
-include(GenSPMPkgConfig)
-spm_generate_pkgconfig_file()
-spm_generate_env_file()
+include(GenPkgConfig)
+list(APPEND SPM_PKGCONFIG_LIBS_PRIVATE
+ ${LAPACKE_LIBRARIES_DEP}
+ ${CBLAS_LIBRARIES_DEP}
+ )
+generate_pkgconfig_files(
+ "${CMAKE_CURRENT_SOURCE_DIR}/tools/spm.pc.in"
+ "${CMAKE_CURRENT_SOURCE_DIR}/tools/spmf.pc.in"
+ PROJECTNAME SPM )
+
+generate_env_file( PROJECTNAME SPM )
### Add documented files to the global property
add_documented_files(
diff --git a/cmake_modules/GenSPMPkgConfig.cmake b/cmake_modules/GenSPMPkgConfig.cmake
deleted file mode 100644
index c8780307..00000000
--- a/cmake_modules/GenSPMPkgConfig.cmake
+++ /dev/null
@@ -1,143 +0,0 @@
-###
-#
-# @copyright (c) 2009-2014 The University of Tennessee and The University
-# of Tennessee Research Foundation.
-# All rights reserved.
-# @copyright (c) 2012-2017 Inria. All rights reserved.
-# @copyright (c) 2012-2014 Bordeaux INP, CNRS (LaBRI UMR 5800), Inria, Univ. Bordeaux. All rights reserved.
-#
-###
-#
-# @file GenPkgConfig.cmake
-#
-# @project MORSE
-# MORSE is a software package provided by:
-# Inria Bordeaux - Sud-Ouest,
-# Univ. of Tennessee,
-# King Abdullah Univesity of Science and Technology
-# Univ. of California Berkeley,
-# Univ. of Colorado Denver.
-#
-# @version 0.9.1
-# @author Cedric Castagnede
-# @author Emmanuel Agullo
-# @author Mathieu Faverge
-# @author Florent Pruvost
-# @date 10-11-2014
-#
-###
-
-###
-#
-# CONVERT_LIBSTYLE_TO_PKGCONFIG: convert a libraries list to follow the pkg-config style
-# used in CLEAN_LIB_LIST
-#
-###
-# macro(CONVERT_LIBSTYLE_TO_PKGCONFIG _liblist)
-# set(${_liblist}_CPY "${${_liblist}}")
-# set(${_liblist} "")
-# foreach(_dep ${${_liblist}_CPY})
-# if (${_dep} MATCHES "^/")
-# get_filename_component(dep_libname ${_dep} NAME)
-# get_filename_component(dep_libdir ${_dep} DIRECTORY)
-# string(REPLACE "lib" "" dep_libname "${dep_libname}")
-# string(REPLACE ".so" "" dep_libname "${dep_libname}")
-# string(REPLACE ".a" "" dep_libname "${dep_libname}")
-# string(REPLACE ".dylib" "" dep_libname "${dep_libname}")
-# string(REPLACE ".dll" "" dep_libname "${dep_libname}")
-# list(APPEND ${_liblist} -L${dep_libdir} -l${dep_libname})
-# elseif(NOT ${_dep} MATCHES "^-")
-# list(APPEND ${_liblist} "-l${_dep}")
-# else()
-# list(APPEND ${_liblist} ${_dep})
-# endif()
-# endforeach()
-# endmacro(CONVERT_LIBSTYLE_TO_PKGCONFIG)
-
-###
-#
-# CLEAN_LIB_LIST: clean libraries lists to follow the pkg-config style
-# used in GENERATE_PKGCONFIG_FILE
-#
-###
-#macro(CLEAN_LIB_LIST _package)
-# list(REMOVE_DUPLICATES ${_package}_PKGCONFIG_LIBS)
-# list(REMOVE_DUPLICATES ${_package}_PKGCONFIG_LIBS_PRIVATE)
-# list(REMOVE_DUPLICATES ${_package}_PKGCONFIG_REQUIRED)
-# list(REMOVE_DUPLICATES ${_package}_PKGCONFIG_REQUIRED_PRIVATE)
-# convert_libstyle_to_pkgconfig(${_package}_PKGCONFIG_LIBS)
-# convert_libstyle_to_pkgconfig(${_package}_PKGCONFIG_LIBS_PRIVATE)
-# string(REPLACE ";" " " ${_package}_PKGCONFIG_LIBS "${${_package}_PKGCONFIG_LIBS}")
-# string(REPLACE ";" " " ${_package}_PKGCONFIG_LIBS_PRIVATE "${${_package}_PKGCONFIG_LIBS_PRIVATE}")
-# string(REPLACE ";" " " ${_package}_PKGCONFIG_REQUIRED "${${_package}_PKGCONFIG_REQUIRED}")
-# string(REPLACE ";" " " ${_package}_PKGCONFIG_REQUIRED_PRIVATE "${${_package}_PKGCONFIG_REQUIRED_PRIVATE}")
-#endmacro(CLEAN_LIB_LIST)
-
-###
-#
-# GENERATE_PKGCONFIG_FILE: generate files spm.pc
-#
-###
-macro(spm_generate_pkgconfig_file)
-
- # The link flags specific to this package and any required libraries
- # that don't support PkgConfig
- set(SPM_PKGCONFIG_LIBS spm)
-
- # The link flags for private libraries required by this package but not
- # exposed to applications
- set(SPM_PKGCONFIG_LIBS_PRIVATE "")
-
- # A list of packages required by this package
- set(SPM_PKGCONFIG_REQUIRED "")
-
- # A list of private packages required by this package but not exposed to
- # applications
- set(SPM_PKGCONFIG_REQUIRED_PRIVATE "")
-
- # Define required package
- # -----------------------
- #clean_lib_list(SPM)
-
- # Create .pc file
- # ---------------
- configure_file(
- "${CMAKE_CURRENT_SOURCE_DIR}/tools/spm.pc.in"
- "${CMAKE_BINARY_DIR}/lib/pkgconfig/spm.pc" @ONLY)
-
- configure_file(
- "${CMAKE_CURRENT_SOURCE_DIR}/tools/spmf.pc.in"
- "${CMAKE_BINARY_DIR}/lib/pkgconfig/spmf.pc" @ONLY)
-
- # installation
- # ------------
- install(FILES
- "${CMAKE_BINARY_DIR}/lib/pkgconfig/spm.pc"
- "${CMAKE_BINARY_DIR}/lib/pkgconfig/spmf.pc"
- DESTINATION lib/pkgconfig)
-
-endmacro(spm_generate_pkgconfig_file)
-
-###
-#
-# generate_env_file: generate files pastix.pc
-#
-###
-macro(spm_generate_env_file)
-
- # Create .sh file
- # ---------------
- configure_file(
- "${CMAKE_CURRENT_SOURCE_DIR}/tools/spm_env.sh.in"
- "${CMAKE_BINARY_DIR}/bin/spm_env.sh" @ONLY)
-
- # installation
- # ------------
- install(FILES "${CMAKE_BINARY_DIR}/bin/spm_env.sh"
- DESTINATION bin)
-
-endmacro(spm_generate_env_file)
-
-##
-## @end file GenPkgConfig.cmake
-##
--
GitLab
From 9196b916683d7749f8a67c7817d1d4e2a3438f53 Mon Sep 17 00:00:00 2001
From: Mathieu Faverge <mathieu.faverge@inria.fr>
Date: Thu, 28 Jun 2018 09:00:52 +0200
Subject: [PATCH 05/22] Update cmake
---
cmake_modules/morse_cmake | 2 +-
tools/spm.pc.in | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/cmake_modules/morse_cmake b/cmake_modules/morse_cmake
index 47b8241b..65647861 160000
--- a/cmake_modules/morse_cmake
+++ b/cmake_modules/morse_cmake
@@ -1 +1 @@
-Subproject commit 47b8241b579347ecc7213ccb7eba0b387f036c98
+Subproject commit 65647861f43e22c83f5be69c26b64b60461c8e86
diff --git a/tools/spm.pc.in b/tools/spm.pc.in
index 56c3d7b3..42bea7f4 100644
--- a/tools/spm.pc.in
+++ b/tools/spm.pc.in
@@ -7,7 +7,7 @@ Name: SPM
Description: SParse Matrix package
Version: @SPM_VERSION_MAJOR@.@SPM_VERSION_MINOR@.@SPM_VERSION_MICRO@
Cflags: -I${includedir}
-Libs: -L${libdir} @SPM_PKGCONFIG_LIBS@
+Libs: -L${libdir} -lspm @SPM_PKGCONFIG_LIBS@
Libs.private: @SPM_PKGCONFIG_LIBS_PRIVATE@
Requires: @SPM_PKGCONFIG_REQUIRED@
Requires.private: @SPM_PKGCONFIG_REQUIRED_PRIVATE@
--
GitLab
From 5b2ea4e23b237e0d6e979b34b57bc9782c3b1cdf Mon Sep 17 00:00:00 2001
From: Mathieu Faverge <mathieu.faverge@inria.fr>
Date: Thu, 28 Jun 2018 11:30:17 +0200
Subject: [PATCH 06/22] Apply some updates from pastix
---
.gitlab-ci.yml | 13 +++++++------
tools/analysis.sh | 1 +
tools/wrappers/wrap_fortran.py | 7 +------
3 files changed, 9 insertions(+), 12 deletions(-)
diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml
index f25b1b2b..acde4028 100644
--- a/.gitlab-ci.yml
+++ b/.gitlab-ci.yml
@@ -4,7 +4,7 @@ stages:
- build
- test
- sonar
-# - doc
+# - deploy
build_spm:
stage: build
@@ -16,7 +16,7 @@ build_spm:
- git submodule update --init --recursive
- mkdir build
- cd build
- - cmake .. -DCMAKE_INSTALL_PREFIX=${PWD}/../install -DBUILD_SHARED_LIBS=ON -DCMAKE_VERBOSE_MAKEFILE=ON -DCMAKE_C_FLAGS="-O0 -g -fPIC --coverage -Wall -fdiagnostics-show-option -fno-inline" -DCMAKE_EXE_LINKER_FLAGS="--coverage" -DPASTIX_INT64=OFF
+ - cmake .. -DCMAKE_INSTALL_PREFIX=${PWD}/../install -DBUILD_SHARED_LIBS=ON -DCMAKE_VERBOSE_MAKEFILE=ON -DCMAKE_C_FLAGS="-O0 -g -fPIC --coverage -Wall -fdiagnostics-show-option -fno-inline" -DCMAKE_EXE_LINKER_FLAGS="--coverage" -DSPM_INT64=OFF -DCMAKE_EXPORT_COMPILE_COMMANDS=ON
- make -j 4 | tee ../pastix-build.log
- make install | tee -a ../pastix-build.log
only:
@@ -34,6 +34,7 @@ test_spm:
- spm.lcov
- spm-gcov.log
script:
+ - git submodule update --init --recursive
- source install/bin/spm_env.sh
- (cd build &&
eval "ctest
@@ -68,9 +69,12 @@ sonar_spm:
- master@solverstack/spm
# pages:
-# stage: doc
+# stage: deploy
# dependencies:
# - build_spm
+# artifacts:
+# paths:
+# - public
# script:
# - git submodule update --init --recursive
# - mkdir -p build
@@ -78,9 +82,6 @@ sonar_spm:
# - cmake .. -DBUILD_DOCUMENTATION=ON -DSPM_INT64=OFF
# - make docs
# - mv docs/out/html ../public/
-# artifacts:
-# paths:
-# - public
# only:
# - master@solverstack/spm
diff --git a/tools/analysis.sh b/tools/analysis.sh
index 0ec0039d..83739173 100755
--- a/tools/analysis.sh
+++ b/tools/analysis.sh
@@ -62,6 +62,7 @@ sonar.c.compiler.reportPath=spm-build.log
sonar.c.coverage.reportPath=spm-coverage.xml
sonar.c.cppcheck.reportPath=spm-cppcheck.xml
sonar.c.rats.reportPath=spm-rats.xml
+sonar.c.jsonCompilationDatabase=build/compile_commands.json
EOF
# run sonar analysis + publish on sonarqube-dev
diff --git a/tools/wrappers/wrap_fortran.py b/tools/wrappers/wrap_fortran.py
index e4250f83..828a8115 100644
--- a/tools/wrappers/wrap_fortran.py
+++ b/tools/wrappers/wrap_fortran.py
@@ -87,7 +87,7 @@ def iso_c_wrapper_type(arg, args_list, args_size):
else:
is_pointer = False
- if (is_pointer and arg[0].strip() == "void"):
+ if (is_pointer and (arg[0].strip() == "void")):
f_type = "type(c_ptr), "
else:
f_type = types_dict[arg[0]] + ", "
@@ -108,11 +108,6 @@ def iso_c_wrapper_type(arg, args_list, args_size):
f_name = arg[2]
- # Force case of myorder
- if f_name == "myorder":
- f_intent = "intent(in), "
- f_target = "pointer"
-
# detect array argument
if (is_pointer and f_name in arrays_names_2D):
f_array = "(:,:)"
--
GitLab
From df72d824d8b4958825a359db3babd7db77394d6a Mon Sep 17 00:00:00 2001
From: Mathieu Faverge <mathieu.faverge@inria.fr>
Date: Thu, 28 Jun 2018 11:53:01 +0200
Subject: [PATCH 07/22] Silent warning
---
src/drivers/skitf.f | 251 ++++++++++++--------------------------------
1 file changed, 69 insertions(+), 182 deletions(-)
diff --git a/src/drivers/skitf.f b/src/drivers/skitf.f
index 5c597345..cf609894 100644
--- a/src/drivers/skitf.f
+++ b/src/drivers/skitf.f
@@ -29,16 +29,16 @@ c for the rows: perm(i) is the destination of row i in the
c permuted matrix -- also the destination of column i in case
c permutation is symmetric (job .le. 2)
c
-c qperm = same thing for the columns. This should be provided only
+c qperm = same thing for the columns. This should be provided only
c if job=3 or job=4, i.e., only in the case of a nonsymmetric
c permutation of rows and columns. Otherwise qperm is a dummy
c
-c job = integer indicating the work to be done:
+c job = integer indicating the work to be done:
c * job = 1,2 permutation is symmetric Ao :== P * A * transp(P)
-c job = 1 permute a, ja, ia into ao, jao, iao
+c job = 1 permute a, ja, ia into ao, jao, iao
c job = 2 permute matrix ignoring real values.
c * job = 3,4 permutation is non-symmetric Ao :== P * A * Q
-c job = 3 permute a, ja, ia into ao, jao, iao
+c job = 3 permute a, ja, ia into ao, jao, iao
c job = 4 permute matrix ignoring real values.
c
c on return:
@@ -168,7 +168,7 @@ c perm = integer array of length n containing the permutation arrays
c for the rows: perm(i) is the destination of row i in the
c permuted matrix -- also the destination of column i.
c
-c rperm = reverse permutation. defined by rperm(iperm(i))=i for all i
+c rperm = reverse permutation. defined by rperm(iperm(i))=i for all i
c
c job = job indicator. if (job .ne.1) values are not copied (i.e.,
c only pattern is copied).
@@ -351,17 +351,17 @@ c ao, jao, iao in csr format
c
c a, ja, ia = input matrix in csr format
c
-c iperm = integer array of length nrow containing the reverse permutation
+c iperm = integer array of length nrow containing the reverse permutation
c array for the rows. row number iperm(j) in permuted matrix PA
c used to be row number j in unpermuted matrix.
c ---> a(i,j) in the permuted matrix was a(iperm(i),j)
c in the inout matrix.
c
-c job = integer indicating the work to be done:
+c job = integer indicating the work to be done:
c job .ne. 1 : get structure only of output matrix,,
c i.e., ignore real values. (in which case arrays a
c and ao are not used nor accessed).
-c job = 1 get complete data structure of output matrix.
+c job = 1 get complete data structure of output matrix.
c (i.e., including arrays ao and iao).
c------------
c on return:
@@ -809,128 +809,15 @@ c
c---------end-of-readmt-------------------------------------------------
c-----------------------------------------------------------------------
end
-c-----------------------------------------------------------------------
- subroutine roscal(nrow,job,nrm,a,ja,ia,diag,b,jb,ib,ierr)
- real*8 a(*), b(*), diag(nrow)
- integer nrow,job,nrm,ja(*),jb(*),ia(nrow+1),ib(nrow+1),ierr
-c-----------------------------------------------------------------------
-c scales the rows of A such that their norms are one on return
-c 3 choices of norms: 1-norm, 2-norm, max-norm.
-c-----------------------------------------------------------------------
-c on entry:
-c ---------
-c nrow = integer. The row dimension of A
-c
-c job = integer. job indicator. Job=0 means get array b only
-c job = 1 means get b, and the integer arrays ib, jb.
-c
-c nrm = integer. norm indicator. nrm = 1, means 1-norm, nrm =2
-c means the 2-nrm, nrm = 0 means max norm
-c
-c a,
-c ja,
-c ia = Matrix A in compressed sparse row format.
-c
-c on return:
-c----------
-c
-c diag = diagonal matrix stored as a vector containing the matrix
-c by which the rows have been scaled, i.e., on return
-c we have B = Diag*A.
-c
-c b,
-c jb,
-c ib = resulting matrix B in compressed sparse row sparse format.
-c
-c ierr = error message. ierr=0 : Normal return
-c ierr=i > 0 : Row number i is a zero row.
-c Notes:
-c-------
-c 1) The column dimension of A is not needed.
-c 2) algorithm in place (B can take the place of A).
-c-----------------------------------------------------------------
- call rnrms (nrow,nrm,a,ja,ia,diag)
-
- ierr = 0
- do 1 j=1, nrow
- if (diag(j) .eq. 0.0d0) then
- ierr = j
- return
- else
- diag(j) = 1.0d0/diag(j)
- endif
- 1 continue
- call diamua(nrow,job,a,ja,ia,diag,b,jb,ib)
- return
-c-------end-of-roscal---------------------------------------------------
-c-----------------------------------------------------------------------
- end
-c-----------------------------------------------------------------------
- subroutine coscal(nrow,job,nrm,a,ja,ia,diag,b,jb,ib,ierr)
-c-----------------------------------------------------------------------
- real*8 a(*),b(*),diag(nrow)
- integer nrow,job,ja(*),jb(*),ia(nrow+1),ib(nrow+1),ierr
-c-----------------------------------------------------------------------
-c scales the columns of A such that their norms are one on return
-c result matrix written on b, or overwritten on A.
-c 3 choices of norms: 1-norm, 2-norm, max-norm. in place.
-c-----------------------------------------------------------------------
-c on entry:
-c ---------
-c nrow = integer. The row dimension of A
-c
-c job = integer. job indicator. Job=0 means get array b only
-c job = 1 means get b, and the integer arrays ib, jb.
-c
-c nrm = integer. norm indicator. nrm = 1, means 1-norm, nrm =2
-c means the 2-nrm, nrm = 0 means max norm
-c
-c a,
-c ja,
-c ia = Matrix A in compressed sparse row format.
-c
-c on return:
-c----------
-c
-c diag = diagonal matrix stored as a vector containing the matrix
-c by which the columns have been scaled, i.e., on return
-c we have B = A * Diag
-c
-c b,
-c jb,
-c ib = resulting matrix B in compressed sparse row sparse format.
-c
-c ierr = error message. ierr=0 : Normal return
-c ierr=i > 0 : Column number i is a zero row.
-c Notes:
-c-------
-c 1) The column dimension of A is not needed.
-c 2) algorithm in place (B can take the place of A).
-c-----------------------------------------------------------------
- call cnrms (nrow,nrm,a,ja,ia,diag)
- ierr = 0
- do 1 j=1, nrow
- if (diag(j) .eq. 0.0) then
- ierr = j
- return
- else
- diag(j) = 1.0d0/diag(j)
- endif
- 1 continue
- call amudia (nrow,job,a,ja,ia,diag,b,jb,ib)
- return
-c--------end-of-coscal--------------------------------------------------
-c-----------------------------------------------------------------------
- end
- subroutine rnrms (nrow, nrm, a, ja, ia, diag)
+ subroutine rnrms (nrow, nrm, a, ia, diag)
real*8 a(*), diag(nrow), scal
- integer ja(*), ia(nrow+1)
+ integer ia(nrow+1)
c-----------------------------------------------------------------------
c gets the norms of each row of A. (choice of three norms)
c-----------------------------------------------------------------------
c on entry:
c ---------
-c nrow = integer. The row dimension of A
+c nrow = integer. The row dimension of A
c
c nrm = integer. norm indicator. nrm = 1, means 1-norm, nrm =2
c means the 2-nrm, nrm = 0 means max norm
@@ -981,7 +868,7 @@ c gets the norms of each column of A. (choice of three norms)
c-----------------------------------------------------------------------
c on entry:
c ---------
-c nrow = integer. The row dimension of A
+c nrow = integer. The row dimension of A
c
c nrm = integer. norm indicator. nrm = 1, means 1-norm, nrm =2
c means the 2-nrm, nrm = 0 means max norm
@@ -1030,7 +917,7 @@ c performs the matrix by matrix product B = Diag * A (in place)
c-----------------------------------------------------------------------
c on entry:
c ---------
-c nrow = integer. The row dimension of A
+c nrow = integer. The row dimension of A
c
c job = integer. job indicator. Job=0 means get array b only
c job = 1 means get b, and the integer arrays ib, jb.
@@ -1046,7 +933,7 @@ c----------
c
c b,
c jb,
-c ib = resulting matrix B in compressed sparse row sparse format.
+c ib = resulting matrix B in compressed sparse row sparse format.
c
c Notes:
c-------
@@ -1087,7 +974,7 @@ c performs the matrix by matrix product B = A * Diag (in place)
c-----------------------------------------------------------------------
c on entry:
c ---------
-c nrow = integer. The row dimension of A
+c nrow = integer. The row dimension of A
c
c job = integer. job indicator. Job=0 means get array b only
c job = 1 means get b, and the integer arrays ib, jb.
@@ -1103,7 +990,7 @@ c----------
c
c b,
c jb,
-c ib = resulting matrix B in compressed sparse row sparse format.
+c ib = resulting matrix B in compressed sparse row sparse format.
c
c Notes:
c-------
@@ -1144,7 +1031,7 @@ c performs the matrix sum C = A+B.
c-----------------------------------------------------------------------
c on entry:
c ---------
-c nrow = integer. The row dimension of A and B
+c nrow = integer. The row dimension of A and B
c ncol = integer. The column dimension of A and B.
c job = integer. Job indicator. When job = 0, only the structure
c (i.e. the arrays jc, ic) is computed and the
@@ -1156,9 +1043,9 @@ c ia = Matrix A in compressed sparse row format.
c
c b,
c jb,
-c ib = Matrix B in compressed sparse row format.
+c ib = Matrix B in compressed sparse row format.
c
-c nzmax = integer. The length of the arrays c and jc.
+c nzmax = integer. The length of the arrays c and jc.
c amub will stop if the result matrix C has a number
c of elements that exceeds exceeds nzmax. See ierr.
c
@@ -1166,9 +1053,9 @@ c on return:
c----------
c c,
c jc,
-c ic = resulting matrix C in compressed sparse row sparse format.
+c ic = resulting matrix C in compressed sparse row sparse format.
c
-c ierr = integer. serving as error message.
+c ierr = integer. serving as error message.
c ierr = 0 means normal return,
c ierr .gt. 0 means that amub stopped while computing the
c i-th row of C with i=ierr, because the number
@@ -1176,7 +1063,7 @@ c of elements in C exceeds nzmax.
c
c work arrays:
c------------
-c iw = integer work array of length equal to the number of
+c iw = integer work array of length equal to the number of
c columns in A.
c
c-----------------------------------------------------------------------
@@ -1239,8 +1126,8 @@ c this subroutine transposes a matrix stored in a, ja, ia format.
c ---------------
c on entry:
c----------
-c n = dimension of A.
-c job = integer to indicate whether to fill the values (job.eq.1) of the
+c n = dimension of A.
+c job = integer to indicate whether to fill the values (job.eq.1) of the
c matrix ao or only the pattern., i.e.,ia, and ja (job .ne.1)
c
c ipos = starting position in ao, jao of the transposed matrix.
@@ -1249,19 +1136,19 @@ c Note: this may be useful if one needs to append the data structure
c of the transpose to that of A. In this case use for example
c call csrcsc (n,1,n+2,a,ja,ia,a,ja,ia(n+2))
c for any other normal usage, enter ipos=1.
-c a = real array of length nnz (nnz=number of nonzero elements in input
+c a = real array of length nnz (nnz=number of nonzero elements in input
c matrix) containing the nonzero elements.
-c ja = integer array of length nnz containing the column positions
+c ja = integer array of length nnz containing the column positions
c of the corresponding elements in a.
-c ia = integer of size n+1. ia(k) contains the position in a, ja of
+c ia = integer of size n+1. ia(k) contains the position in a, ja of
c the beginning of the k-th row.
c
c on return:
c ----------
c output arguments:
-c ao = real array of size nzz containing the "a" part of the transpose
-c jao = integer array of size nnz containing the column indices.
-c iao = integer array of size n+1 containing the "ia" index array of
+c ao = real array of size nzz containing the "a" part of the transpose
+c jao = integer array of size nnz containing the column indices.
+c iao = integer array of size n+1 containing the "ia" index array of
c the transpose.
c
c-----------------------------------------------------------------------
@@ -1284,9 +1171,9 @@ c this subroutine transposes a matrix stored in a, ja, ia format.
c ---------------
c on entry:
c----------
-c n = number of rows of CSR matrix.
+c n = number of rows of CSR matrix.
c n2 = number of columns of CSC matrix.
-c job = integer to indicate whether to fill the values (job.eq.1) of the
+c job = integer to indicate whether to fill the values (job.eq.1) of the
c matrix ao or only the pattern., i.e.,ia, and ja (job .ne.1)
c
c ipos = starting position in ao, jao of the transposed matrix.
@@ -1295,19 +1182,19 @@ c Note: this may be useful if one needs to append the data structure
c of the transpose to that of A. In this case use for example
c call csrcsc2 (n,n,1,n+2,a,ja,ia,a,ja,ia(n+2))
c for any other normal usage, enter ipos=1.
-c a = real array of length nnz (nnz=number of nonzero elements in input
+c a = real array of length nnz (nnz=number of nonzero elements in input
c matrix) containing the nonzero elements.
-c ja = integer array of length nnz containing the column positions
+c ja = integer array of length nnz containing the column positions
c of the corresponding elements in a.
-c ia = integer of size n+1. ia(k) contains the position in a, ja of
+c ia = integer of size n+1. ia(k) contains the position in a, ja of
c the beginning of the k-th row.
c
c on return:
c ----------
c output arguments:
-c ao = real array of size nzz containing the "a" part of the transpose
-c jao = integer array of size nnz containing the column indices.
-c iao = integer array of size n+1 containing the "ia" index array of
+c ao = real array of size nzz containing the "a" part of the transpose
+c jao = integer array of size nnz containing the column indices.
+c iao = integer array of size n+1 containing the "ia" index array of
c the transpose.
c
c-----------------------------------------------------------------------
@@ -1808,18 +1695,18 @@ c this subroutine performs an in-place permutation of a real vector x
c according to the permutation array perm(*), i.e., on return,
c the vector x satisfies,
c
-c x(perm(j)) :== x(j), j=1,2,.., n
+c x(perm(j)) :== x(j), j=1,2,.., n
c
c-----------------------------------------------------------------------
c on entry:
c---------
c n = length of vector x.
c perm = integer array of length n containing the permutation array.
-c x = input vector
+c x = input vector
c
c on return:
c----------
-c x = vector x permuted according to x(perm(*)) := x(*)
+c x = vector x permuted according to x(perm(*)) := x(*)
c
c----------------------------------------------------------------------c
c Y. Saad, Sep. 21 1989 c
@@ -1828,7 +1715,7 @@ c local variables
real*8 tmp, tmp1
c
init = 1
- tmp = x(init)
+ tmp = x(init)
ii = perm(init)
perm(init)= -perm(init)
k = 0
@@ -1860,8 +1747,8 @@ c
65 init = init+1
if (init .gt. n) goto 101
if (perm(init) .lt. 0) goto 65
- tmp = x(init)
- ii = perm(init)
+ tmp = x(init)
+ ii = perm(init)
perm(init)=-perm(init)
goto 6
c
@@ -1883,18 +1770,18 @@ c this subroutine performs an in-place permutation of an integer vector
c ix according to the permutation array perm(*), i.e., on return,
c the vector x satisfies,
c
-c ix(perm(j)) :== ix(j), j=1,2,.., n
+c ix(perm(j)) :== ix(j), j=1,2,.., n
c
c-----------------------------------------------------------------------
c on entry:
c---------
c n = length of vector x.
c perm = integer array of length n containing the permutation array.
-c ix = input vector
+c ix = input vector
c
c on return:
c----------
-c ix = vector x permuted according to ix(perm(*)) := ix(*)
+c ix = vector x permuted according to ix(perm(*)) := ix(*)
c
c----------------------------------------------------------------------c
c Y. Saad, Sep. 21 1989 c
@@ -1903,7 +1790,7 @@ c local variables
integer tmp, tmp1
c
init = 1
- tmp = ix(init)
+ tmp = ix(init)
ii = perm(init)
perm(init)= -perm(init)
k = 0
@@ -1935,8 +1822,8 @@ c
65 init = init+1
if (init .gt. n) goto 101
if (perm(init) .lt. 0) goto 65
- tmp = ix(init)
- ii = perm(init)
+ tmp = ix(init)
+ ii = perm(init)
perm(init)=-perm(init)
goto 6
c
@@ -1972,8 +1859,8 @@ c permuted matrix.
c ---> a(i,j) in the original matrix becomes a(perm(i),j)
c in the output matrix.
c
-c job = integer indicating the work to be done:
-c job = 1 permute a, ja, ia into ao, jao, iao
+c job = integer indicating the work to be done:
+c job = 1 permute a, ja, ia into ao, jao, iao
c (including the copying of real values ao and
c the array iao).
c job .ne. 1 : ignore real values.
@@ -2042,13 +1929,13 @@ c nrow = row dimension of the matrix
c
c a, ja, ia = input matrix in csr format.
c
-c perm = integer array of length ncol (number of columns of A
+c perm = integer array of length ncol (number of columns of A
c containing the permutation array the columns:
c a(i,j) in the original matrix becomes a(i,perm(j))
c in the output matrix.
c
-c job = integer indicating the work to be done:
-c job = 1 permute a, ja, ia into ao, jao, iao
+c job = integer indicating the work to be done:
+c job = 1 permute a, ja, ia into ao, jao, iao
c (including the copying of real values ao and
c the array iao).
c job .ne. 1 : ignore real values ao and ignore iao.
@@ -2558,6 +2445,7 @@ c
c select domain with largest size
c
maxsiz = 0
+ nextdom = 0
do j=1, idom
if (size(j) .gt. maxsiz) then
maxsiz = size(j)
@@ -2830,11 +2718,10 @@ c print *,'mask',riord(j),j,nfirst
mask(riord(j)) = 0
12 continue
c-----------------------------------------------------------------------
- 13 continue
c
1 nlev = nlev+1
levels(nlev) = istart + 1
- call add_lvst (istart,iend,nlev,riord,ja,ia,mask,maskval)
+ call add_lvst (istart,iend,riord,ja,ia,mask,maskval)
if (istart .lt. iend) goto 1
2 ii = ii+1
if (ii .le. n) then
@@ -2854,7 +2741,7 @@ c
endif
endif
c-----------------------------------------------------------------------
- 3 levels(nlev+1) = iend+1
+ levels(nlev+1) = iend+1
do j=1, iend
mask(riord(j)) = maskval
enddo
@@ -2862,8 +2749,8 @@ c-----------------------------------------------------------------------
return
end
c-----------------------------------------------------------------------
- subroutine add_lvst(istart,iend,nlev,riord,ja,ia,mask,maskval)
- integer nlev, nod, riord(*), ja(*), ia(*), mask(*)
+ subroutine add_lvst(istart,iend,riord,ja,ia,mask,maskval)
+ integer nod, riord(*), ja(*), ia(*), mask(*)
c-------------------------------------------------------------
c adds one level set to the previous sets..
c span all nodes of previous mask
@@ -3120,7 +3007,7 @@ c
nstuck = 0
c-----------------------------------------------------------------------
100 continue
- idom = mindom(n,ndom,levst,link)
+ idom = mindom(n,ndom,link)
c-----------------------------------------------------------------------
c begin level-set loop
c-----------------------------------------------------------------------
@@ -3141,7 +3028,7 @@ c
do 2 kk=ia(i), ia(i+1)-1
j = ja(kk)
if (marker(j) .eq. 0) then
- call add_lk(j,nod,idom,ndom,lkend,levst,link,nodes,marker)
+ call add_lk(j,nod,idom,lkend,levst,link,nodes,marker)
endif
2 continue
c
@@ -3192,11 +3079,11 @@ c
c
c if no neighboring domain select smallest one
c
- if (idom .eq. 0) idom = mindom(n,ndom,levst,link)
+ if (idom .eq. 0) idom = mindom(n,ndom,link)
c
c add ii to sudomain idom at end of linked list
c
- call add_lk(ii,nod,idom,ndom,lkend,levst,link,nodes,marker)
+ call add_lk(ii,nod,idom,lkend,levst,link,nodes,marker)
goto 3
else
goto 20
@@ -3242,9 +3129,9 @@ c
c-----------------------------------------------------------------------
end
c-----------------------------------------------------------------------
- function mindom(n, ndom, levst, link)
+ function mindom(n, ndom, link)
implicit none
- integer mindom, n, ndom, levst(2*ndom),link(n)
+ integer mindom, n, ndom, link(n)
c-----------------------------------------------------------------------
c returns the domain with smallest size
c-----------------------------------------------------------------------
@@ -3252,6 +3139,7 @@ c locals
c
integer i, nsize, isiz
c
+ mindom = -1
isiz = n+1
do 10 i=1, ndom
nsize = - link(i)
@@ -3264,9 +3152,9 @@ c
return
end
c-----------------------------------------------------------------------
- subroutine add_lk(new,nod,idom,ndom,lkend,levst,link,nodes,marker)
+ subroutine add_lk(new,nod,idom,lkend,levst,link,nodes,marker)
implicit none
- integer new,nod,idom,ndom,lkend,levst(*),link(*),nodes(*),
+ integer new,nod,idom,lkend,levst(*),link(*),nodes(*),
* marker(*)
c-----------------------------------------------------------------------
c adds from head --
@@ -3275,7 +3163,6 @@ c adds one entry (new) to linked list and ipdates everything.
c new = node to be added
c nod = current number of marked nodes
c idom = domain to which new is to be added
-c ndom = total number of domains
c lkend= location of end of structure (link and nodes)
c levst= pointer array for link, nodes
c link = link array
--
GitLab
From 11362f25b152c012c1f748b8b765ee45a952f7a7 Mon Sep 17 00:00:00 2001
From: Mathieu Faverge <mathieu.faverge@inria.fr>
Date: Thu, 28 Jun 2018 11:53:32 +0200
Subject: [PATCH 08/22] Fix issue with complex numbers on double
---
wrappers/fortran90/examples/spm_user.f90 | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/wrappers/fortran90/examples/spm_user.f90 b/wrappers/fortran90/examples/spm_user.f90
index cdf8df9b..d511aa1a 100644
--- a/wrappers/fortran90/examples/spm_user.f90
+++ b/wrappers/fortran90/examples/spm_user.f90
@@ -73,19 +73,19 @@ program spm_user
if (i < dim1) then
rowptr(l) = i + dim1 * (j-1) + dim1 * dim2 * (k-1) + 1
colptr(l) = (i-1) + dim1 * (j-1) + dim1 * dim2 * (k-1) + 1
- values(l) = - 1. - 1. * I
+ values(l) = - 1.
l = l + 1
end if
if (j < dim2) then
rowptr(l) = (i-1) + dim1 * j + dim1 * dim2 * (k-1) + 1
colptr(l) = (i-1) + dim1 * (j-1) + dim1 * dim2 * (k-1) + 1
- values(l) = - 1. - 1. * I
+ values(l) = - 1.
l = l + 1
end if
if (k < dim3) then
rowptr(l) = (i-1) + dim1 * (j-1) + dim1 * dim2 * k + 1
colptr(l) = (i-1) + dim1 * (j-1) + dim1 * dim2 * (k-1) + 1
- values(l) = -1. - 1. * I
+ values(l) = -1.
l = l + 1
end if
end do
--
GitLab
From 77c7232ba39d0066804f340b0ecbd37c9ad53499 Mon Sep 17 00:00:00 2001
From: Mathieu Faverge <mathieu.faverge@inria.fr>
Date: Thu, 28 Jun 2018 11:55:43 +0200
Subject: [PATCH 09/22] Update the wrapper with last changes on pastix
---
tools/gen_wrappers.py | 5 ++---
wrappers/fortran90/CMakeLists.txt | 4 ++--
wrappers/fortran90/src/spm_enums.F90 | 3 +--
3 files changed, 5 insertions(+), 7 deletions(-)
diff --git a/tools/gen_wrappers.py b/tools/gen_wrappers.py
index e03a3c5b..b923b517 100755
--- a/tools/gen_wrappers.py
+++ b/tools/gen_wrappers.py
@@ -450,14 +450,13 @@ enums_python_coeftype='''
return -1
'''
-enums_fortran_footer='''
- integer, parameter :: spm_int_t = SPM_INT_KIND
+enums_fortran_footer=''' integer, parameter :: spm_int_t = SPM_INT_KIND
contains
function spm_getintsize()
integer :: spm_getintsize
- spm_getintsize = SPM_INT_KIND
+ spm_getintsize = kind(SPM_INT_KIND)
return
end function spm_getintsize
'''
diff --git a/wrappers/fortran90/CMakeLists.txt b/wrappers/fortran90/CMakeLists.txt
index dd80adc3..a7a81fea 100644
--- a/wrappers/fortran90/CMakeLists.txt
+++ b/wrappers/fortran90/CMakeLists.txt
@@ -21,11 +21,11 @@ add_library( spmf
if ( SPM_INT64 )
set_source_files_properties(
src/spm_enums.F90
- PROPERTIES COMPILE_DEFINITIONS "SPM_INT_KIND=8")
+ PROPERTIES COMPILE_DEFINITIONS "SPM_INT_KIND=c_int64_t")
else()
set_source_files_properties(
src/spm_enums.F90
- PROPERTIES COMPILE_DEFINITIONS "SPM_INT_KIND=4")
+ PROPERTIES COMPILE_DEFINITIONS "SPM_INT_KIND=c_int32_t")
endif()
target_link_libraries( spmf spm )
diff --git a/wrappers/fortran90/src/spm_enums.F90 b/wrappers/fortran90/src/spm_enums.F90
index 5b7f9c93..c1919678 100644
--- a/wrappers/fortran90/src/spm_enums.F90
+++ b/wrappers/fortran90/src/spm_enums.F90
@@ -129,14 +129,13 @@ module spm_enums
enumerator :: SpmDirBackward = 392
end enum
-
integer, parameter :: spm_int_t = SPM_INT_KIND
contains
function spm_getintsize()
integer :: spm_getintsize
- spm_getintsize = SPM_INT_KIND
+ spm_getintsize = kind(SPM_INT_KIND)
return
end function spm_getintsize
--
GitLab
From 97754cda904f76e11614cc0e2391b83650a953f9 Mon Sep 17 00:00:00 2001
From: Mathieu Faverge <mathieu.faverge@inria.fr>
Date: Thu, 28 Jun 2018 14:23:25 +0200
Subject: [PATCH 10/22] Update cmake module
---
cmake_modules/morse_cmake | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/cmake_modules/morse_cmake b/cmake_modules/morse_cmake
index 65647861..d44efd0e 160000
--- a/cmake_modules/morse_cmake
+++ b/cmake_modules/morse_cmake
@@ -1 +1 @@
-Subproject commit 65647861f43e22c83f5be69c26b64b60461c8e86
+Subproject commit d44efd0e4af75cf4be6b9483198c595b03ec083a
--
GitLab
From 34a9a90e1ef9ce050acf6b3b1ba654c35fcbe7e3 Mon Sep 17 00:00:00 2001
From: Mathieu Faverge <mathieu.faverge@inria.fr>
Date: Fri, 29 Jun 2018 17:46:20 +0200
Subject: [PATCH 11/22] Silent warnigns with icc
---
src/spm_io.c | 8 ++++----
src/z_spm.c | 6 ++++--
src/z_spm_expand.c | 10 ++++++----
src/z_spm_laplacian.c | 2 +-
4 files changed, 15 insertions(+), 11 deletions(-)
diff --git a/src/spm_io.c b/src/spm_io.c
index 11da33cf..9d536cbe 100644
--- a/src/spm_io.c
+++ b/src/spm_io.c
@@ -508,14 +508,14 @@ spmLoad( spmatrix_t *spm,
return SPM_ERR_BADPARAMETER;
}
- spm->mtxtype = mtxtype;
- spm->flttype = flttype;
- spm->fmttype = fmttype;
+ spm->mtxtype = (spm_mtxtype_t)mtxtype;
+ spm->flttype = (spm_coeftype_t)flttype;
+ spm->fmttype = (spm_fmttype_t)fmttype;
spm->gN = gN;
spm->n = n;
spm->nnz = nnz;
spm->dof = dof;
- spm->layout = layout;
+ spm->layout = (spm_layout_t)layout;
spmUpdateComputedFields( spm );
diff --git a/src/z_spm.c b/src/z_spm.c
index e0d45d07..457e7ba9 100644
--- a/src/z_spm.c
+++ b/src/z_spm.c
@@ -139,8 +139,11 @@ z_spmMergeDuplicate( spmatrix_t *spm )
spm_int_t n = spm->n;
spm_int_t baseval = spm->colptr[0];
spm_int_t dof2 = spm->dof * spm->dof;
- spm_int_t idx, i, j, d, size;
+ spm_int_t idx, i, j, size;
spm_int_t merge = 0;
+#if !defined(PRECISION_p)
+ spm_int_t d;
+#endif
if ( spm->fmttype == SpmCSC ) {
idx = 0;
@@ -196,7 +199,6 @@ z_spmMergeDuplicate( spmatrix_t *spm )
}
}
- (void)d;
return merge;
}
diff --git a/src/z_spm_expand.c b/src/z_spm_expand.c
index 78a3785c..ad4280a1 100644
--- a/src/z_spm_expand.c
+++ b/src/z_spm_expand.c
@@ -46,7 +46,8 @@ z_spmCSCExpand(const spmatrix_t *spm)
spm_int_t i, j, k, ii, jj, dofi, dofj, col, row, baseval, lda;
spm_int_t diag, height;
spm_int_t *newcol, *newrow, *oldcol, *oldrow, *dofs;
- spm_complex64_t *newval, *oldval, *oldval2;
+ spm_complex64_t *newval = NULL;
+ spm_complex64_t *oldval2, *oldval = NULL;
assert( spm->fmttype == SpmCSC );
assert( spm->flttype == SpmComplex64 );
@@ -200,7 +201,6 @@ z_spmCSCExpand(const spmatrix_t *spm)
assert(spm->loc2glob == NULL);
- (void)newval;
(void)lda;
return newspm;
}
@@ -234,7 +234,8 @@ z_spmCSRExpand(const spmatrix_t *spm)
spm_int_t i, j, k, ii, jj, dofi, dofj, col, row, baseval, lda;
spm_int_t diag, height;
spm_int_t *newcol, *newrow, *oldcol, *oldrow, *dofs;
- spm_complex64_t *newval, *oldval, *oldval2;
+ spm_complex64_t *newval = NULL;
+ spm_complex64_t *oldval2, *oldval = NULL;
assert( spm->fmttype == SpmCSR );
assert( spm->flttype == SpmComplex64 );
@@ -419,7 +420,8 @@ z_spmIJVExpand(const spmatrix_t *spm)
spmatrix_t *newspm;
spm_int_t i, j, k, ii, jj, dofi, dofj, col, row, baseval;
spm_int_t *newcol, *newrow, *oldcol, *oldrow, *dofs;
- spm_complex64_t *newval, *oldval=NULL;
+ spm_complex64_t *newval = NULL;
+ spm_complex64_t *oldval = NULL;
assert( spm->fmttype == SpmIJV );
assert( spm->flttype == SpmComplex64 );
diff --git a/src/z_spm_laplacian.c b/src/z_spm_laplacian.c
index 68fc4361..135e61cd 100644
--- a/src/z_spm_laplacian.c
+++ b/src/z_spm_laplacian.c
@@ -310,10 +310,10 @@ z_spmLaplacian_27points( spmatrix_t *spm,
/* Diagonal value */
*rowptr = l;
-#if !defined(PRECISION_p)
degree = 1;
d = 1;
+#if !defined(PRECISION_p)
if (k > 1) {
d++;
}
--
GitLab
From a134134ba4ae422f52b3ff6688093d22b0b4de13 Mon Sep 17 00:00:00 2001
From: Mathieu Faverge <mathieu.faverge@inria.fr>
Date: Mon, 2 Jul 2018 18:10:27 +0200
Subject: [PATCH 12/22] Change the spmcheckandcorrect to fix the C/fortran
issue of compatibility with malloc/free and allocate/deallocate
---
include/spm.h | 10 ++++---
src/spm.c | 33 ++++++++++++++--------
src/z_spm_expand.c | 6 ++--
tools/wrappers/wrap_fortran.py | 1 +
tools/wrappers/wrap_python.py | 1 +
wrappers/fortran90/examples/spm_driver.f90 | 6 ++--
wrappers/fortran90/examples/spm_user.f90 | 13 +++++++--
wrappers/fortran90/src/spmf.f90 | 17 ++++++-----
wrappers/python/spm/__spm__.py | 12 ++++----
9 files changed, 63 insertions(+), 36 deletions(-)
diff --git a/include/spm.h b/include/spm.h
index 6fb956e2..90940b10 100644
--- a/include/spm.h
+++ b/include/spm.h
@@ -48,6 +48,8 @@
*
*/
typedef struct spmatrix_s {
+ int8_t alloc; /**< Internal field to detect if the internal buffers has been
+ user allocated or not. */
spm_mtxtype_t mtxtype; /**< Matrix structure: SpmGeneral, SpmSymmetric
or SpmHermitian. */
spm_coeftype_t flttype; /**< values datatype: SpmPattern, SpmFloat, SpmDouble,
@@ -110,10 +112,10 @@ void spmScalVector( spm_coeftype_t flt, double alpha, spm_int_t n, void *x, sp
* @name SPM subroutines to check format
* @{
*/
-int spmSort( spmatrix_t *spm );
-spm_int_t spmMergeDuplicate( spmatrix_t *spm );
-spm_int_t spmSymmetrize( spmatrix_t *spm );
-spmatrix_t *spmCheckAndCorrect( spmatrix_t *spm );
+int spmSort( spmatrix_t *spm );
+spm_int_t spmMergeDuplicate( spmatrix_t *spm );
+spm_int_t spmSymmetrize( spmatrix_t *spm );
+int spmCheckAndCorrect( const spmatrix_t *in, spmatrix_t *out );
/**
* @}
diff --git a/src/spm.c b/src/spm.c
index 49d1daf9..c2279a82 100644
--- a/src/spm.c
+++ b/src/spm.c
@@ -676,24 +676,28 @@ spmSymmetrize( spmatrix_t *spm )
*
*******************************************************************************
*
- * @param[inout] spm
+ * @param[in] in
* The pointer to the sparse matrix structure to check, and correct.
*
+ * @param[inout] out
+ * On entry, an allocated structure to hold the corrected spm.
+ * On exit, holds the pointer to spm corrected.
+ *
*******************************************************************************
*
- * @return if no changes have been made, the initial sparse matrix is returned,
- * otherwise a copy of the sparse matrix, fixed to meet the PaStiX requirements,
- * is returned.
+ * @return 0 if no changes have been made to the spm matrix.
+ * @return 1 if corrections have been applied to the in sparse matrix.
*
*******************************************************************************/
-spmatrix_t *
-spmCheckAndCorrect( spmatrix_t *spm )
+int
+spmCheckAndCorrect( const spmatrix_t *in,
+ spmatrix_t *out )
{
spmatrix_t *newspm = NULL;
spm_int_t count;
/* Let's work on a copy */
- newspm = spmCopy( spm );
+ newspm = spmCopy( in );
/* PaStiX works on CSC matrices */
spmConvert( SpmCSC, newspm );
@@ -733,15 +737,18 @@ spmCheckAndCorrect( spmatrix_t *spm )
* Check if we return the new one, or the original one because no changes
* have been made
*/
- if (( spm->fmttype != newspm->fmttype ) ||
- ( spm->nnzexp != newspm->nnzexp ) )
+ if (( in->fmttype != newspm->fmttype ) ||
+ ( in->nnzexp != newspm->nnzexp ) )
{
- return newspm;
+ memcpy( out, newspm, sizeof(spmatrix_t) );
+ free( newspm );
+ return 1;
}
else {
+ memcpy( out, in, sizeof(spmatrix_t) );
spmExit( newspm );
- free(newspm);
- return (spmatrix_t*)spm;
+ free( newspm );
+ return 0;
}
}
@@ -815,6 +822,8 @@ spmCopy( const spmatrix_t *spm )
newspm->values = malloc(valsize);
memcpy( newspm->values, spm->values, valsize );
}
+
+ newspm->alloc = 1;
return newspm;
}
diff --git a/src/z_spm_expand.c b/src/z_spm_expand.c
index ad4280a1..0fa2d278 100644
--- a/src/z_spm_expand.c
+++ b/src/z_spm_expand.c
@@ -46,7 +46,9 @@ z_spmCSCExpand(const spmatrix_t *spm)
spm_int_t i, j, k, ii, jj, dofi, dofj, col, row, baseval, lda;
spm_int_t diag, height;
spm_int_t *newcol, *newrow, *oldcol, *oldrow, *dofs;
+#if !defined(PRECISION_p)
spm_complex64_t *newval = NULL;
+#endif
spm_complex64_t *oldval2, *oldval = NULL;
assert( spm->fmttype == SpmCSC );
@@ -420,9 +422,10 @@ z_spmIJVExpand(const spmatrix_t *spm)
spmatrix_t *newspm;
spm_int_t i, j, k, ii, jj, dofi, dofj, col, row, baseval;
spm_int_t *newcol, *newrow, *oldcol, *oldrow, *dofs;
+#if !defined(PRECISION_p)
spm_complex64_t *newval = NULL;
+#endif
spm_complex64_t *oldval = NULL;
-
assert( spm->fmttype == SpmIJV );
assert( spm->flttype == SpmComplex64 );
@@ -568,7 +571,6 @@ z_spmIJVExpand(const spmatrix_t *spm)
assert(spm->loc2glob == NULL);
- (void)newval;
return newspm;
}
diff --git a/tools/wrappers/wrap_fortran.py b/tools/wrappers/wrap_fortran.py
index 828a8115..69242d33 100644
--- a/tools/wrappers/wrap_fortran.py
+++ b/tools/wrappers/wrap_fortran.py
@@ -30,6 +30,7 @@ iindent=3
# translation_table of types
types_dict = {
"int": ("integer(kind=c_int)"),
+ "int8_t": ("integer(kind=c_int8_t)"),
"spm_coeftype_t": ("integer(c_int)"),
"spm_dir_t": ("integer(c_int)"),
"spm_trans_t": ("integer(c_int)"),
diff --git a/tools/wrappers/wrap_python.py b/tools/wrappers/wrap_python.py
index 60945858..f1371f77 100644
--- a/tools/wrappers/wrap_python.py
+++ b/tools/wrappers/wrap_python.py
@@ -26,6 +26,7 @@ iindent=4
# translation_table of types
types_dict = {
"int": ("c_int"),
+ "int8_t": ("c_int8)"),
"spm_coeftype_t": ("c_int"),
"spm_dir_t": ("c_int"),
"spm_trans_t": ("c_int"),
diff --git a/wrappers/fortran90/examples/spm_driver.f90 b/wrappers/fortran90/examples/spm_driver.f90
index fc7f05e4..3e856d21 100644
--- a/wrappers/fortran90/examples/spm_driver.f90
+++ b/wrappers/fortran90/examples/spm_driver.f90
@@ -17,7 +17,7 @@ program spm_driver
implicit none
type(spmatrix_t), target :: spm
- type(spmatrix_t), pointer :: spm2
+ type(spmatrix_t), target :: spm2
real(kind=c_double) :: normA
real(kind=c_double) :: eps = 1.e-15
integer(c_int) :: info
@@ -30,8 +30,8 @@ program spm_driver
! 1- The matrix
call spmReadDriver( SpmDriverLaplacian, "d:10:10:10:4.", spm, info )
- call spmCheckAndCorrect( spm, spm2 )
- if (.not. c_associated(c_loc(spm), c_loc(spm2))) then
+ call spmCheckAndCorrect( spm, spm2, info )
+ if ( info .ne. 0 ) then
call spmExit( spm )
spm = spm2
end if
diff --git a/wrappers/fortran90/examples/spm_user.f90 b/wrappers/fortran90/examples/spm_user.f90
index d511aa1a..8e5f8789 100644
--- a/wrappers/fortran90/examples/spm_user.f90
+++ b/wrappers/fortran90/examples/spm_user.f90
@@ -22,10 +22,11 @@ program spm_user
type(c_ptr) :: x0_ptr, x_ptr, b_ptr
real(kind=c_double) :: eps = 1.e-15
type(spmatrix_t), target :: spm
- type(spmatrix_t), pointer :: spm2
+ type(spmatrix_t), target :: spm2
integer(kind=spm_int_t) :: dim1, dim2, dim3, n, nnz
integer(kind=spm_int_t) :: i, j, k, l, nrhs
integer(c_int) :: info
+ logical :: spm_modified
!
! Generate a 10x10x10 complex Laplacian in IJV format
@@ -112,8 +113,9 @@ program spm_user
call spmUpdateComputedFields( spm )
- call spmCheckAndCorrect( spm, spm2 )
- if (.not. c_associated(c_loc(spm), c_loc(spm2))) then
+ call spmCheckAndCorrect( spm, spm2, info )
+ spm_modified = .not. ( info .eq. 0 )
+ if ( spm_modified ) then
deallocate(rowptr)
deallocate(colptr)
deallocate(values)
@@ -149,6 +151,11 @@ program spm_user
call spmCheckAxb( eps, nrhs, spm, x0_ptr, spm%n, b_ptr, spm%n, x_ptr, spm%n, info )
call spmExit( spm )
+ if ( .not. spm_modified ) then
+ deallocate(rowptr)
+ deallocate(colptr)
+ deallocate(values)
+ end if
deallocate(x0)
deallocate(x)
deallocate(b)
diff --git a/wrappers/fortran90/src/spmf.f90 b/wrappers/fortran90/src/spmf.f90
index f3c722a0..4a85b145 100644
--- a/wrappers/fortran90/src/spmf.f90
+++ b/wrappers/fortran90/src/spmf.f90
@@ -19,6 +19,7 @@ module spmf
implicit none
type, bind(c) :: spmatrix_t
+ integer(kind=c_int8_t) :: alloc
integer(c_int) :: mtxtype
integer(c_int) :: flttype
integer(c_int) :: fmttype
@@ -234,13 +235,14 @@ module spmf
end interface
interface
- function spmCheckAndCorrect_c(spm) &
+ function spmCheckAndCorrect_c(in, out) &
bind(c, name='spmCheckAndCorrect')
use iso_c_binding
import spmatrix_t
implicit none
- type(c_ptr) :: spmCheckAndCorrect_c
- type(c_ptr), value :: spm
+ integer(kind=c_int) :: spmCheckAndCorrect_c
+ type(c_ptr), value :: in
+ type(c_ptr), value :: out
end function spmCheckAndCorrect_c
end interface
@@ -566,13 +568,14 @@ contains
value = spmSymmetrize_c(c_loc(spm))
end subroutine spmSymmetrize
- subroutine spmCheckAndCorrect(spm, spmo)
+ subroutine spmCheckAndCorrect(in, out, info)
use iso_c_binding
implicit none
- type(spmatrix_t), intent(inout), target :: spm
- type(spmatrix_t), intent(out), pointer :: spmo
+ type(spmatrix_t), intent(in), target :: in
+ type(spmatrix_t), intent(inout), target :: out
+ integer(kind=c_int), intent(out) :: info
- call c_f_pointer(spmCheckAndCorrect_c(c_loc(spm)), spmo)
+ info = spmCheckAndCorrect_c(c_loc(in), c_loc(out))
end subroutine spmCheckAndCorrect
subroutine spmGenRHS(type, nrhs, spm, x, ldx, b, ldb, info)
diff --git a/wrappers/python/spm/__spm__.py b/wrappers/python/spm/__spm__.py
index a4b902fb..507e0d83 100644
--- a/wrappers/python/spm/__spm__.py
+++ b/wrappers/python/spm/__spm__.py
@@ -23,7 +23,8 @@ from . import libspm
from .enum import __spm_int__
class pyspm_spmatrix_t(Structure):
- _fields_ = [("mtxtype", c_int ),
+ _fields_ = [("alloc", c_int8) ),
+ ("mtxtype", c_int ),
("flttype", c_int ),
("fmttype", c_int ),
("gN", __spm_int__ ),
@@ -119,10 +120,11 @@ def pyspm_spmSymmetrize( spm ):
libspm.spmSymmetrize.restype = __spm_int__
return libspm.spmSymmetrize( spm )
-def pyspm_spmCheckAndCorrect( spm ):
- libspm.spmCheckAndCorrect.argtypes = [ POINTER(pyspm_spmatrix_t) ]
- libspm.spmCheckAndCorrect.restype = POINTER(pyspm_spmatrix_t)
- return libspm.spmCheckAndCorrect( spm )
+def pyspm_spmCheckAndCorrect( in, out ):
+ libspm.spmCheckAndCorrect.argtypes = [ POINTER(pyspm_spmatrix_t),
+ POINTER(pyspm_spmatrix_t) ]
+ libspm.spmCheckAndCorrect.restype = c_int
+ return libspm.spmCheckAndCorrect( in, out )
def pyspm_spmGenRHS( type, nrhs, spm, x, ldx, b, ldb ):
libspm.spmGenRHS.argtypes = [ c_int, __spm_int__, POINTER(pyspm_spmatrix_t),
--
GitLab
From db5e70bc7aecf169da090beeae1630540483cbf7 Mon Sep 17 00:00:00 2001
From: Mathieu Faverge <mathieu.faverge@inria.fr>
Date: Mon, 2 Jul 2018 18:10:56 +0200
Subject: [PATCH 13/22] Missing assert
---
src/z_spm_convert_to_ijv.c | 12 ++++++++++++
1 file changed, 12 insertions(+)
diff --git a/src/z_spm_convert_to_ijv.c b/src/z_spm_convert_to_ijv.c
index 6404b208..47328e87 100644
--- a/src/z_spm_convert_to_ijv.c
+++ b/src/z_spm_convert_to_ijv.c
@@ -41,6 +41,12 @@ z_spmConvertCSC2IJV( spmatrix_t *spm )
spm_int_t *col_ijv, *colptr;
spm_int_t i, j, baseval, nnz;
+ /*
+ * Make sure that the spm has been internally allocated to avoid freeing
+ * user's allocated array
+ */
+ assert( spm->alloc );
+
/*
* Check the baseval
*/
@@ -91,6 +97,12 @@ z_spmConvertCSR2IJV( spmatrix_t *spm )
spm_int_t *row_ijv, *rowptr;
spm_int_t i, j, baseval, nnz;
+ /*
+ * Make sure that the spm has been internally allocated to avoid freeing
+ * user's allocated array
+ */
+ assert( spm->alloc );
+
/*
* Check the baseval
*/
--
GitLab
From 85c682753eeab112d72e208f01d4a2dc58ef865b Mon Sep 17 00:00:00 2001
From: Mathieu Faverge <mathieu.faverge@inria.fr>
Date: Mon, 2 Jul 2018 19:36:33 +0200
Subject: [PATCH 14/22] Chane in/out as it is keaywords in python/fortran
---
include/spm.h | 2 +-
src/spm.c | 14 ++---
tools/wrappers/wrap_python.py | 2 +-
wrappers/fortran90/src/spmf.f90 | 14 ++---
wrappers/python/spm/__spm__.py | 6 +-
wrappers/python/spm/spm.py | 105 ++++++++++++++++++++++++++++----
6 files changed, 113 insertions(+), 30 deletions(-)
diff --git a/include/spm.h b/include/spm.h
index 90940b10..8c04dbe0 100644
--- a/include/spm.h
+++ b/include/spm.h
@@ -115,7 +115,7 @@ void spmScalVector( spm_coeftype_t flt, double alpha, spm_int_t n, void *x, sp
int spmSort( spmatrix_t *spm );
spm_int_t spmMergeDuplicate( spmatrix_t *spm );
spm_int_t spmSymmetrize( spmatrix_t *spm );
-int spmCheckAndCorrect( const spmatrix_t *in, spmatrix_t *out );
+int spmCheckAndCorrect( const spmatrix_t *spm_in, spmatrix_t *spm_out );
/**
* @}
diff --git a/src/spm.c b/src/spm.c
index c2279a82..bfb6d47e 100644
--- a/src/spm.c
+++ b/src/spm.c
@@ -690,14 +690,14 @@ spmSymmetrize( spmatrix_t *spm )
*
*******************************************************************************/
int
-spmCheckAndCorrect( const spmatrix_t *in,
- spmatrix_t *out )
+spmCheckAndCorrect( const spmatrix_t *spm_in,
+ spmatrix_t *spm_out )
{
spmatrix_t *newspm = NULL;
spm_int_t count;
/* Let's work on a copy */
- newspm = spmCopy( in );
+ newspm = spmCopy( spm_in );
/* PaStiX works on CSC matrices */
spmConvert( SpmCSC, newspm );
@@ -737,15 +737,15 @@ spmCheckAndCorrect( const spmatrix_t *in,
* Check if we return the new one, or the original one because no changes
* have been made
*/
- if (( in->fmttype != newspm->fmttype ) ||
- ( in->nnzexp != newspm->nnzexp ) )
+ if (( spm_in->fmttype != newspm->fmttype ) ||
+ ( spm_in->nnzexp != newspm->nnzexp ) )
{
- memcpy( out, newspm, sizeof(spmatrix_t) );
+ memcpy( spm_out, newspm, sizeof(spmatrix_t) );
free( newspm );
return 1;
}
else {
- memcpy( out, in, sizeof(spmatrix_t) );
+ memcpy( spm_out, spm_in, sizeof(spmatrix_t) );
spmExit( newspm );
free( newspm );
return 0;
diff --git a/tools/wrappers/wrap_python.py b/tools/wrappers/wrap_python.py
index f1371f77..44a75e53 100644
--- a/tools/wrappers/wrap_python.py
+++ b/tools/wrappers/wrap_python.py
@@ -26,7 +26,7 @@ iindent=4
# translation_table of types
types_dict = {
"int": ("c_int"),
- "int8_t": ("c_int8)"),
+ "int8_t": ("c_int8"),
"spm_coeftype_t": ("c_int"),
"spm_dir_t": ("c_int"),
"spm_trans_t": ("c_int"),
diff --git a/wrappers/fortran90/src/spmf.f90 b/wrappers/fortran90/src/spmf.f90
index 4a85b145..f71a02d0 100644
--- a/wrappers/fortran90/src/spmf.f90
+++ b/wrappers/fortran90/src/spmf.f90
@@ -235,14 +235,14 @@ module spmf
end interface
interface
- function spmCheckAndCorrect_c(in, out) &
+ function spmCheckAndCorrect_c(spm_in, spm_out) &
bind(c, name='spmCheckAndCorrect')
use iso_c_binding
import spmatrix_t
implicit none
integer(kind=c_int) :: spmCheckAndCorrect_c
- type(c_ptr), value :: in
- type(c_ptr), value :: out
+ type(c_ptr), value :: spm_in
+ type(c_ptr), value :: spm_out
end function spmCheckAndCorrect_c
end interface
@@ -568,14 +568,14 @@ contains
value = spmSymmetrize_c(c_loc(spm))
end subroutine spmSymmetrize
- subroutine spmCheckAndCorrect(in, out, info)
+ subroutine spmCheckAndCorrect(spm_in, spm_out, info)
use iso_c_binding
implicit none
- type(spmatrix_t), intent(in), target :: in
- type(spmatrix_t), intent(inout), target :: out
+ type(spmatrix_t), intent(in), target :: spm_in
+ type(spmatrix_t), intent(inout), target :: spm_out
integer(kind=c_int), intent(out) :: info
- info = spmCheckAndCorrect_c(c_loc(in), c_loc(out))
+ info = spmCheckAndCorrect_c(c_loc(spm_in), c_loc(spm_out))
end subroutine spmCheckAndCorrect
subroutine spmGenRHS(type, nrhs, spm, x, ldx, b, ldb, info)
diff --git a/wrappers/python/spm/__spm__.py b/wrappers/python/spm/__spm__.py
index 507e0d83..b39def3c 100644
--- a/wrappers/python/spm/__spm__.py
+++ b/wrappers/python/spm/__spm__.py
@@ -23,7 +23,7 @@ from . import libspm
from .enum import __spm_int__
class pyspm_spmatrix_t(Structure):
- _fields_ = [("alloc", c_int8) ),
+ _fields_ = [("alloc", c_int8 ),
("mtxtype", c_int ),
("flttype", c_int ),
("fmttype", c_int ),
@@ -120,11 +120,11 @@ def pyspm_spmSymmetrize( spm ):
libspm.spmSymmetrize.restype = __spm_int__
return libspm.spmSymmetrize( spm )
-def pyspm_spmCheckAndCorrect( in, out ):
+def pyspm_spmCheckAndCorrect( spm_in, spm_out ):
libspm.spmCheckAndCorrect.argtypes = [ POINTER(pyspm_spmatrix_t),
POINTER(pyspm_spmatrix_t) ]
libspm.spmCheckAndCorrect.restype = c_int
- return libspm.spmCheckAndCorrect( in, out )
+ return libspm.spmCheckAndCorrect( spm_in, spm_out )
def pyspm_spmGenRHS( type, nrhs, spm, x, ldx, b, ldb ):
libspm.spmGenRHS.argtypes = [ c_int, __spm_int__, POINTER(pyspm_spmatrix_t),
diff --git a/wrappers/python/spm/spm.py b/wrappers/python/spm/spm.py
index 016a4cb0..eb8ca174 100644
--- a/wrappers/python/spm/spm.py
+++ b/wrappers/python/spm/spm.py
@@ -38,14 +38,15 @@ class spmatrix():
if mtxtype_ == mtxtype.HerPosDef:
mtxtype_ = mtxtype.Hermitian
- self.spm_c = pyspm_spmatrix_t( mtxtype_,
- coeftype.Double,
- fmttype.CSC,
- 0, 0, 0, 0, 0, 0, 0, 0,
- 1, None,
- layout.ColMajor,
- None, None, None, None )
+ self.spm_c = pyspm_spmatrix_t( 0, mtxtype_,
+ coeftype.Double,
+ fmttype.CSC,
+ 0, 0, 0, 0, 0, 0, 0, 0,
+ 1, None,
+ layout.ColMajor,
+ None, None, None, None )
self.id_ptr = pointer( self.spm_c )
+
if A is not None:
self.fromsps( A, mtxtype_ )
elif driver is not None:
@@ -124,7 +125,55 @@ class spmatrix():
pyspm_spmReadDriver( driver, filename.encode('utf-8'), self.id_ptr )
+ print( "After driver")
self.dtype = coeftype.getnptype( self.spm_c.flttype )
+ print( "Before Check")
+ self.checkAndCorrect()
+ print( "After Check")
+
+ def fromnull( self, A ):
+ """
+ Initialize the SPM wrapper by loading the libraries
+ """
+
+ if not sps.isspmatrix(A):
+ raise TypeError( "A must be of type scipy.sparse matrix" )
+
+ # Assume A is already in Scipy sparse format
+ self.dtype = A.dtype
+ flttype = coeftype.getptype( A.dtype )
+ if flttype == -1:
+ raise TypeError( "Invalid data type. Must be part of (f4, f8, c8 or c16)" )
+
+ A = sps.csc_matrix( A )
+ A.sort_indices()
+
+ # Pointer variables
+ self.py_colptr = np.array( A.indptr[:], dtype=__spm_int__ )
+ self.py_rowptr = np.array( A.indices[:], dtype=__spm_int__ )
+ self.py_values = np.array( A.data[:] )
+
+ if mtxtype_ == mtxtype.SymPosDef:
+ mtxtype_ = mtxtype.Symmetric
+ if mtxtype_ == mtxtype.HerPosDef:
+ mtxtype_ = mtxtype.Hermitian
+
+ self.spm_c.mtxtype = mtxtype_
+ self.spm_c.flttype = flttype
+ self.spm_c.fmttype = fmttype.CSC
+ self.spm_c.n = A.shape[0]
+ self.spm_c.nnz = A.getnnz()
+ self.spm_c.dof = 1
+ self.spm_c.dofs = None
+ self.spm_c.layout = layout.ColMajor
+ self.spm_c.colptr = self.py_colptr.ctypes.data_as(POINTER(__spm_int__))
+ self.spm_c.rowptr = self.py_rowptr.ctypes.data_as(POINTER(__spm_int__))
+ self.spm_c.loc2glob = None
+ self.spm_c.values = self.py_values.ctypes.data_as(c_void_p)
+
+ self.id_ptr = pointer( self.spm_c )
+
+ self.updateComputedField()
self.checkAndCorrect()
def scale( self, alpha ):
@@ -150,13 +199,16 @@ class spmatrix():
def checkAndCorrect( self ):
spm1 = self.id_ptr
- spm2 = pyspm_spmCheckAndCorrect( self.id_ptr )
- if (( spm1.contents.fmttype == spm2.contents.fmttype ) and
- ( spm1.contents.nnzexp == spm2.contents.nnzexp ) ):
+ spm2 = spmatrix()
+ print( spm1 )
+ print( spm2 )
+ rc = pyspm_spmCheckAndCorrect( self.id_ptr, spm2.id_ptr )
+ if ( rc == 0 ):
return
+ self = spm2
self.spm_c = cast( spm2, POINTER(pyspm_spmatrix_t) ).contents
self.id_ptr = pointer( self.spm_c )
-
+ #
self.py_colptr = np.frombuffer( (__spm_int__ * (n+1)).from_address( cast(self.spm_c.colptr, c_void_p).value ), __spm_int__ ).copy()
self.py_rowptr = np.frombuffer( (__spm_int__ * nnz ).from_address( cast(self.spm_c.rowptr, c_void_p).value ), __spm_int__ ).copy()
self.py_values = np.frombuffer( (cflt * nnz ).from_address( self.spm_c.values ), nflt ).copy()
@@ -225,3 +277,34 @@ class spmatrix():
x.ctypes.data_as(c_void_p), ldx,
b.ctypes.data_as(c_void_p), ldb )
return x, b
+
+ def mult( self, B, C, trans=trans.NoTrans, n=-1, alpha=1.0, beta=0. ):
+
+ m = self.spm_c.n
+
+ B = np.asarray(B, self.dtype)
+ C = np.asarray(C, self.dtype)
+
+ if n == -1:
+ if B.ndim == 1:
+ n = 1
+ else:
+ n = B.shape[1]
+
+ self.__checkVector( m, n, B )
+ self.__checkVector( m, n, C )
+
+ ldb = B.shape[0]
+ ldc = B.shape[0]
+
+ if n == 1:
+ pyspm_spmMatVec( trans, alpha, self.id_ptr,
+ B.ctypes.data_as(c_void_p),
+ beta,
+ C.ctypes.data_as(c_void_p) )
+ else:
+ pyspm_spmMatMat( trans, n, alpha, self.id_ptr,
+ B.ctypes.data_as(c_void_p), ldb,
+ beta,
+ C.ctypes.data_as(c_void_p), ldc )
+
--
GitLab
From 4b1912571149ff7eacd18f5c919ee6ec05e01d33 Mon Sep 17 00:00:00 2001
From: Mathieu Faverge <mathieu.faverge@inria.fr>
Date: Mon, 2 Jul 2018 19:36:56 +0200
Subject: [PATCH 15/22] Silent warings
---
tests/spm_dof_matvec_tests.c | 5 -----
tests/spm_matvec_tests.c | 5 -----
tests/spm_tests.h | 8 ++++----
tests/z_spm_tests.c | 5 +++--
4 files changed, 7 insertions(+), 16 deletions(-)
diff --git a/tests/spm_dof_matvec_tests.c b/tests/spm_dof_matvec_tests.c
index 8d5e1a98..1a9c9738 100644
--- a/tests/spm_dof_matvec_tests.c
+++ b/tests/spm_dof_matvec_tests.c
@@ -23,11 +23,6 @@
#include <time.h>
#include "spm_tests.h"
-int z_spm_matvec_check( int trans, const spmatrix_t *spm );
-int c_spm_matvec_check( int trans, const spmatrix_t *spm );
-int d_spm_matvec_check( int trans, const spmatrix_t *spm );
-int s_spm_matvec_check( int trans, const spmatrix_t *spm );
-
#define PRINT_RES(_ret_) \
if(_ret_) { \
printf("FAILED(%d)\n", _ret_); \
diff --git a/tests/spm_matvec_tests.c b/tests/spm_matvec_tests.c
index 1a7c5e40..5cb80698 100644
--- a/tests/spm_matvec_tests.c
+++ b/tests/spm_matvec_tests.c
@@ -22,11 +22,6 @@
#include <time.h>
#include "spm_tests.h"
-int z_spm_matvec_check( int trans, const spmatrix_t *spm );
-int c_spm_matvec_check( int trans, const spmatrix_t *spm );
-int d_spm_matvec_check( int trans, const spmatrix_t *spm );
-int s_spm_matvec_check( int trans, const spmatrix_t *spm );
-
#define PRINT_RES(_ret_) \
if(_ret_) { \
printf("FAILED(%d)\n", _ret_); \
diff --git a/tests/spm_tests.h b/tests/spm_tests.h
index 765de234..7d79984f 100644
--- a/tests/spm_tests.h
+++ b/tests/spm_tests.h
@@ -72,19 +72,19 @@ int core_sgeadd( spm_trans_t trans,
spm_int_t LDB );
void z_spm_print_check( char *filename, const spmatrix_t *spm );
-int z_spm_matvec_check( int trans, const spmatrix_t *spm );
+int z_spm_matvec_check( spm_trans_t trans, const spmatrix_t *spm );
int z_spm_norm_check( const spmatrix_t *spm );
void c_spm_print_check( char *filename, const spmatrix_t *spm );
-int c_spm_matvec_check( int trans, const spmatrix_t *spm );
+int c_spm_matvec_check( spm_trans_t trans, const spmatrix_t *spm );
int c_spm_norm_check( const spmatrix_t *spm );
void d_spm_print_check( char *filename, const spmatrix_t *spm );
-int d_spm_matvec_check( int trans, const spmatrix_t *spm );
+int d_spm_matvec_check( spm_trans_t trans, const spmatrix_t *spm );
int d_spm_norm_check( const spmatrix_t *spm );
void s_spm_print_check( char *filename, const spmatrix_t *spm );
-int s_spm_matvec_check( int trans, const spmatrix_t *spm );
+int s_spm_matvec_check( spm_trans_t trans, const spmatrix_t *spm );
int s_spm_norm_check( const spmatrix_t *spm );
static inline int
diff --git a/tests/z_spm_tests.c b/tests/z_spm_tests.c
index f6487f29..7c1fdc19 100644
--- a/tests/z_spm_tests.c
+++ b/tests/z_spm_tests.c
@@ -109,7 +109,7 @@ z_spm_print_check( char *filename, const spmatrix_t *spm )
* Check the accuracy of the solution
*/
int
-z_spm_matvec_check( int trans, const spmatrix_t *spm )
+z_spm_matvec_check( spm_trans_t trans, const spmatrix_t *spm )
{
unsigned long long int seed = 35469;
spm_complex64_t *A, *x, *y0, *ys, *yd;
@@ -161,7 +161,8 @@ z_spm_matvec_check( int trans, const spmatrix_t *spm )
}
/* Compute the dense matrix-vector product */
- cblas_zgemm( CblasColMajor, trans, CblasNoTrans, spm->gNexp, 1, spm->gNexp,
+ cblas_zgemm( CblasColMajor, (enum CBLAS_TRANSPOSE)trans, CblasNoTrans,
+ spm->gNexp, 1, spm->gNexp,
CBLAS_SADDR(zalpha), A, spm->gNexp,
x, spm->gNexp,
CBLAS_SADDR(zbeta), yd, spm->gNexp );
--
GitLab
From 34b2733d01fde802a89281e494d809a427b4bc30 Mon Sep 17 00:00:00 2001
From: Mathieu Faverge <mathieu.faverge@inria.fr>
Date: Mon, 2 Jul 2018 19:38:06 +0200
Subject: [PATCH 16/22] Remove useless forward check
---
wrappers/python/spm_driver.py | 3 +--
wrappers/python/spm_scipy.py | 8 +++++---
2 files changed, 6 insertions(+), 5 deletions(-)
diff --git a/wrappers/python/spm_driver.py b/wrappers/python/spm_driver.py
index fa88e594..99317ab7 100755
--- a/wrappers/python/spm_driver.py
+++ b/wrappers/python/spm_driver.py
@@ -37,6 +37,5 @@ nrhs = 10
x0, b = A.genRHS( spm.rhstype.RndX, nrhs )
# Check that A * x = b
-x = x0.copy()
-A.checkAxb( x0, b, x )
+A.checkAxb( None, b, x0 )
diff --git a/wrappers/python/spm_scipy.py b/wrappers/python/spm_scipy.py
index 6a574b6a..d2167238 100755
--- a/wrappers/python/spm_scipy.py
+++ b/wrappers/python/spm_scipy.py
@@ -26,11 +26,13 @@ n = 9
A = sps.spdiags([np.ones(n)*i for i in [4, -1, -1, -1, -1]],
[0, 1, 3, -1, -3], n, n)
x0 = np.arange(n)
-b = A.dot(x0)
+b = np.zeros(n)
spmA = spm.spmatrix( A )
+# Multiply A by x
+spmA.mult( x0, b, trans=spm.trans.NoTrans, alpha=1., beta=0. )
+
# Check that A * x = b
-x = x0.copy()
-spmA.checkAxb( x0, b, x )
+spmA.checkAxb( None, b, x0 )
--
GitLab
From 9db56173cafa9503aa214311ca739b6a902e1367 Mon Sep 17 00:00:00 2001
From: Mathieu Faverge <mathieu.faverge@inria.fr>
Date: Mon, 2 Jul 2018 22:43:58 +0200
Subject: [PATCH 17/22] Silent icc warings
---
include/cblas.h | 10 +++++-----
tests/get_options.c | 2 +-
tests/spm_convert_tests.c | 5 +++--
tests/spm_dof_expand_tests.c | 8 +++++---
tests/spm_dof_matvec_tests.c | 4 +++-
tests/spm_dof_norm_tests.c | 10 ++++++----
tests/spm_matvec_tests.c | 6 +++++-
tests/spm_norm_tests.c | 6 ++++--
tests/z_spm_tests.c | 2 +-
9 files changed, 33 insertions(+), 20 deletions(-)
diff --git a/include/cblas.h b/include/cblas.h
index 14756f24..fa70cffa 100644
--- a/include/cblas.h
+++ b/include/cblas.h
@@ -16,11 +16,11 @@ extern "C" { /* Assume C declarations for C++ */
#define CBLAS_INDEX int
#endif
-typedef enum {CblasRowMajor=101, CblasColMajor=102} CBLAS_LAYOUT;
-typedef enum {CblasNoTrans=111, CblasTrans=112, CblasConjTrans=113} CBLAS_TRANSPOSE;
-typedef enum {CblasUpper=121, CblasLower=122} CBLAS_UPLO;
-typedef enum {CblasNonUnit=131, CblasUnit=132} CBLAS_DIAG;
-typedef enum {CblasLeft=141, CblasRight=142} CBLAS_SIDE;
+typedef enum CBLAS_LAYOUT {CblasRowMajor=101, CblasColMajor=102} CBLAS_LAYOUT;
+typedef enum CBLAS_TRANSPOSE {CblasNoTrans=111, CblasTrans=112, CblasConjTrans=113} CBLAS_TRANSPOSE;
+typedef enum CBLAS_UPLO {CblasUpper=121, CblasLower=122} CBLAS_UPLO;
+typedef enum CBLAS_DIAG {CblasNonUnit=131, CblasUnit=132} CBLAS_DIAG;
+typedef enum CBLAS_SIDE {CblasLeft=141, CblasRight=142} CBLAS_SIDE;
typedef CBLAS_LAYOUT CBLAS_ORDER; /* this for backward compatibility with CBLAS_ORDER */
diff --git a/tests/get_options.c b/tests/get_options.c
index 8d8a227b..683d8da0 100644
--- a/tests/get_options.c
+++ b/tests/get_options.c
@@ -105,7 +105,7 @@ spmGetOptions( int argc, char **argv,
exit(0);
}
- *driver = -1;
+ *driver = (spm_driver_t)-1;
do
{
#if defined(HAVE_GETOPT_LONG)
diff --git a/tests/spm_convert_tests.c b/tests/spm_convert_tests.c
index 1bc33bb8..d644cd10 100644
--- a/tests/spm_convert_tests.c
+++ b/tests/spm_convert_tests.c
@@ -45,7 +45,7 @@ int spmComp( const spmatrix_t *spm1,
{
spm_int_t *colptr1, *colptr2;
spm_int_t *rowptr1, *rowptr2;
- int *valptr1, *valptr2;
+ int *valptr1, *valptr2;
spm_int_t i;
if ( spm1->fmttype != SpmCSC ) {
@@ -108,7 +108,8 @@ int main (int argc, char **argv)
char *filename;
spmatrix_t spm, *spm2;
spm_driver_t driver;
- int mtxtype, baseval;
+ spm_mtxtype_t mtxtype;
+ int baseval;
int ret = SPM_SUCCESS;
int err = 0;
FILE *f;
diff --git a/tests/spm_dof_expand_tests.c b/tests/spm_dof_expand_tests.c
index 0a343aa9..e27e495c 100644
--- a/tests/spm_dof_expand_tests.c
+++ b/tests/spm_dof_expand_tests.c
@@ -54,8 +54,10 @@ int main (int argc, char **argv)
spmatrix_t original, *spm;
spm_driver_t driver;
char *filename;
- int spmtype, mtxtype, fmttype, baseval;
- int i, rc, dofmax = 3;
+ spm_mtxtype_t spmtype;
+ spm_mtxtype_t mtxtype;
+ spm_fmttype_t fmttype;
+ int baseval, i, rc, dofmax = 3;
/**
* Get options from command line
@@ -95,7 +97,7 @@ int main (int argc, char **argv)
{
spmBase( &original, baseval );
- for( fmttype=0; fmttype<3; fmttype++ )
+ for( fmttype=SpmCSC; fmttype<=SpmIJV; fmttype++ )
{
spmConvert( fmttype, &original );
spm = spmDofExtend( &original, i, dofmax );
diff --git a/tests/spm_dof_matvec_tests.c b/tests/spm_dof_matvec_tests.c
index 1a9c9738..41723259 100644
--- a/tests/spm_dof_matvec_tests.c
+++ b/tests/spm_dof_matvec_tests.c
@@ -41,7 +41,9 @@ int main (int argc, char **argv)
spmatrix_t original, *spm;
spm_driver_t driver;
char *filename;
- int spmtype, mtxtype, fmttype, baseval;
+ spm_mtxtype_t spmtype, mtxtype;
+ spm_fmttype_t fmttype;
+ int baseval;
int ret = SPM_SUCCESS;
int err = 0;
int rc, i, dofmax = 3;
diff --git a/tests/spm_dof_norm_tests.c b/tests/spm_dof_norm_tests.c
index 6d4e74f6..fe229a3a 100644
--- a/tests/spm_dof_norm_tests.c
+++ b/tests/spm_dof_norm_tests.c
@@ -43,9 +43,11 @@ char* mtxnames[] = { "General", "Symmetric", "Hermitian" };
int main (int argc, char **argv)
{
spmatrix_t original, *spm;
- spm_driver_t driver;
- char *filename;
- int spmtype, mtxtype, fmttype, baseval;
+ spm_driver_t driver;
+ char *filename;
+ spm_mtxtype_t spmtype, mtxtype;
+ spm_fmttype_t fmttype;
+ int baseval;
int rc = SPM_SUCCESS;
int err = 0;
int i, dofmax = 4;
@@ -92,7 +94,7 @@ int main (int argc, char **argv)
{
spmBase( &original, baseval );
- for( fmttype=0; fmttype<3; fmttype++ )
+ for( fmttype=SpmCSC; fmttype<=SpmIJV; fmttype++ )
{
spmConvert( fmttype, &original );
spm = spmDofExtend( &original, i, dofmax );
diff --git a/tests/spm_matvec_tests.c b/tests/spm_matvec_tests.c
index 5cb80698..2ccd492b 100644
--- a/tests/spm_matvec_tests.c
+++ b/tests/spm_matvec_tests.c
@@ -41,7 +41,11 @@ int main (int argc, char **argv)
spmatrix_t spm;
spm_driver_t driver;
char *filename;
- int t,spmtype, mtxtype, fmttype, baseval;
+ spm_trans_t t;
+ spm_mtxtype_t spmtype;
+ spm_mtxtype_t mtxtype;
+ spm_fmttype_t fmttype;
+ int baseval;
int rc = SPM_SUCCESS;
int err = 0;
diff --git a/tests/spm_norm_tests.c b/tests/spm_norm_tests.c
index d6824b73..b2ae5a42 100644
--- a/tests/spm_norm_tests.c
+++ b/tests/spm_norm_tests.c
@@ -45,7 +45,9 @@ int main (int argc, char **argv)
spmatrix_t spm;
spm_driver_t driver;
char *filename;
- int spmtype, mtxtype, fmttype, baseval;
+ spm_mtxtype_t spmtype, mtxtype;
+ spm_fmttype_t fmttype;
+ int baseval;
int rc = SPM_SUCCESS;
int err = 0;
@@ -70,7 +72,7 @@ int main (int argc, char **argv)
spmtype = spm.mtxtype;
printf(" -- SPM Norms Test --\n");
- for( fmttype=0; fmttype<3; fmttype++ ) {
+ for( fmttype=SpmCSC; fmttype<=SpmIJV; fmttype++ ) {
spmConvert( fmttype, &spm );
diff --git a/tests/z_spm_tests.c b/tests/z_spm_tests.c
index 7c1fdc19..17f0bb8c 100644
--- a/tests/z_spm_tests.c
+++ b/tests/z_spm_tests.c
@@ -161,7 +161,7 @@ z_spm_matvec_check( spm_trans_t trans, const spmatrix_t *spm )
}
/* Compute the dense matrix-vector product */
- cblas_zgemm( CblasColMajor, (enum CBLAS_TRANSPOSE)trans, CblasNoTrans,
+ cblas_zgemm( CblasColMajor, (CBLAS_TRANSPOSE)trans, CblasNoTrans,
spm->gNexp, 1, spm->gNexp,
CBLAS_SADDR(zalpha), A, spm->gNexp,
x, spm->gNexp,
--
GitLab
From 0145394fb512d3621f19794b2da21bcfc283d77c Mon Sep 17 00:00:00 2001
From: Mathieu Faverge <mathieu.faverge@inria.fr>
Date: Mon, 2 Jul 2018 23:06:29 +0200
Subject: [PATCH 18/22] remove the useless fromnull
---
wrappers/python/spm/spm.py | 45 --------------------------------------
1 file changed, 45 deletions(-)
diff --git a/wrappers/python/spm/spm.py b/wrappers/python/spm/spm.py
index eb8ca174..3803f7d4 100644
--- a/wrappers/python/spm/spm.py
+++ b/wrappers/python/spm/spm.py
@@ -131,51 +131,6 @@ class spmatrix():
self.checkAndCorrect()
print( "After Check")
- def fromnull( self, A ):
- """
- Initialize the SPM wrapper by loading the libraries
- """
-
- if not sps.isspmatrix(A):
- raise TypeError( "A must be of type scipy.sparse matrix" )
-
- # Assume A is already in Scipy sparse format
- self.dtype = A.dtype
- flttype = coeftype.getptype( A.dtype )
- if flttype == -1:
- raise TypeError( "Invalid data type. Must be part of (f4, f8, c8 or c16)" )
-
- A = sps.csc_matrix( A )
- A.sort_indices()
-
- # Pointer variables
- self.py_colptr = np.array( A.indptr[:], dtype=__spm_int__ )
- self.py_rowptr = np.array( A.indices[:], dtype=__spm_int__ )
- self.py_values = np.array( A.data[:] )
-
- if mtxtype_ == mtxtype.SymPosDef:
- mtxtype_ = mtxtype.Symmetric
- if mtxtype_ == mtxtype.HerPosDef:
- mtxtype_ = mtxtype.Hermitian
-
- self.spm_c.mtxtype = mtxtype_
- self.spm_c.flttype = flttype
- self.spm_c.fmttype = fmttype.CSC
- self.spm_c.n = A.shape[0]
- self.spm_c.nnz = A.getnnz()
- self.spm_c.dof = 1
- self.spm_c.dofs = None
- self.spm_c.layout = layout.ColMajor
- self.spm_c.colptr = self.py_colptr.ctypes.data_as(POINTER(__spm_int__))
- self.spm_c.rowptr = self.py_rowptr.ctypes.data_as(POINTER(__spm_int__))
- self.spm_c.loc2glob = None
- self.spm_c.values = self.py_values.ctypes.data_as(c_void_p)
-
- self.id_ptr = pointer( self.spm_c )
-
- self.updateComputedField()
- self.checkAndCorrect()
-
def scale( self, alpha ):
return pyspm_spmScalMatrix( float(alpha), self.id_ptr )
--
GitLab
From c02380654b45ea7d3727911a9d9390be072549de Mon Sep 17 00:00:00 2001
From: Mathieu Faverge <mathieu.faverge@inria.fr>
Date: Mon, 2 Jul 2018 23:10:41 +0200
Subject: [PATCH 19/22] Remove dirty lines added for debug
---
wrappers/python/spm/spm.py | 6 ------
1 file changed, 6 deletions(-)
diff --git a/wrappers/python/spm/spm.py b/wrappers/python/spm/spm.py
index 3803f7d4..1543aff1 100644
--- a/wrappers/python/spm/spm.py
+++ b/wrappers/python/spm/spm.py
@@ -125,11 +125,8 @@ class spmatrix():
pyspm_spmReadDriver( driver, filename.encode('utf-8'), self.id_ptr )
- print( "After driver")
self.dtype = coeftype.getnptype( self.spm_c.flttype )
- print( "Before Check")
self.checkAndCorrect()
- print( "After Check")
def scale( self, alpha ):
return pyspm_spmScalMatrix( float(alpha), self.id_ptr )
@@ -155,15 +152,12 @@ class spmatrix():
def checkAndCorrect( self ):
spm1 = self.id_ptr
spm2 = spmatrix()
- print( spm1 )
- print( spm2 )
rc = pyspm_spmCheckAndCorrect( self.id_ptr, spm2.id_ptr )
if ( rc == 0 ):
return
self = spm2
self.spm_c = cast( spm2, POINTER(pyspm_spmatrix_t) ).contents
self.id_ptr = pointer( self.spm_c )
- #
self.py_colptr = np.frombuffer( (__spm_int__ * (n+1)).from_address( cast(self.spm_c.colptr, c_void_p).value ), __spm_int__ ).copy()
self.py_rowptr = np.frombuffer( (__spm_int__ * nnz ).from_address( cast(self.spm_c.rowptr, c_void_p).value ), __spm_int__ ).copy()
self.py_values = np.frombuffer( (cflt * nnz ).from_address( self.spm_c.values ), nflt ).copy()
--
GitLab
From 3599defee996d846d4d9dd0b346894dde7e8e92c Mon Sep 17 00:00:00 2001
From: Mathieu Faverge <mathieu.faverge@inria.fr>
Date: Tue, 3 Jul 2018 14:51:15 +0200
Subject: [PATCH 20/22] Remove the alloc field
---
include/spm.h | 2 --
src/spm.c | 1 -
src/z_spm_convert_to_ijv.c | 12 ------------
wrappers/fortran90/src/spmf.f90 | 1 -
wrappers/python/spm/__spm__.py | 3 +--
wrappers/python/spm/spm.py | 2 +-
6 files changed, 2 insertions(+), 19 deletions(-)
diff --git a/include/spm.h b/include/spm.h
index 8c04dbe0..5f11776a 100644
--- a/include/spm.h
+++ b/include/spm.h
@@ -48,8 +48,6 @@
*
*/
typedef struct spmatrix_s {
- int8_t alloc; /**< Internal field to detect if the internal buffers has been
- user allocated or not. */
spm_mtxtype_t mtxtype; /**< Matrix structure: SpmGeneral, SpmSymmetric
or SpmHermitian. */
spm_coeftype_t flttype; /**< values datatype: SpmPattern, SpmFloat, SpmDouble,
diff --git a/src/spm.c b/src/spm.c
index bfb6d47e..64d74e91 100644
--- a/src/spm.c
+++ b/src/spm.c
@@ -823,7 +823,6 @@ spmCopy( const spmatrix_t *spm )
memcpy( newspm->values, spm->values, valsize );
}
- newspm->alloc = 1;
return newspm;
}
diff --git a/src/z_spm_convert_to_ijv.c b/src/z_spm_convert_to_ijv.c
index 47328e87..6404b208 100644
--- a/src/z_spm_convert_to_ijv.c
+++ b/src/z_spm_convert_to_ijv.c
@@ -41,12 +41,6 @@ z_spmConvertCSC2IJV( spmatrix_t *spm )
spm_int_t *col_ijv, *colptr;
spm_int_t i, j, baseval, nnz;
- /*
- * Make sure that the spm has been internally allocated to avoid freeing
- * user's allocated array
- */
- assert( spm->alloc );
-
/*
* Check the baseval
*/
@@ -97,12 +91,6 @@ z_spmConvertCSR2IJV( spmatrix_t *spm )
spm_int_t *row_ijv, *rowptr;
spm_int_t i, j, baseval, nnz;
- /*
- * Make sure that the spm has been internally allocated to avoid freeing
- * user's allocated array
- */
- assert( spm->alloc );
-
/*
* Check the baseval
*/
diff --git a/wrappers/fortran90/src/spmf.f90 b/wrappers/fortran90/src/spmf.f90
index f71a02d0..639aaacf 100644
--- a/wrappers/fortran90/src/spmf.f90
+++ b/wrappers/fortran90/src/spmf.f90
@@ -19,7 +19,6 @@ module spmf
implicit none
type, bind(c) :: spmatrix_t
- integer(kind=c_int8_t) :: alloc
integer(c_int) :: mtxtype
integer(c_int) :: flttype
integer(c_int) :: fmttype
diff --git a/wrappers/python/spm/__spm__.py b/wrappers/python/spm/__spm__.py
index b39def3c..55116cd1 100644
--- a/wrappers/python/spm/__spm__.py
+++ b/wrappers/python/spm/__spm__.py
@@ -23,8 +23,7 @@ from . import libspm
from .enum import __spm_int__
class pyspm_spmatrix_t(Structure):
- _fields_ = [("alloc", c_int8 ),
- ("mtxtype", c_int ),
+ _fields_ = [("mtxtype", c_int ),
("flttype", c_int ),
("fmttype", c_int ),
("gN", __spm_int__ ),
diff --git a/wrappers/python/spm/spm.py b/wrappers/python/spm/spm.py
index 1543aff1..79aab068 100644
--- a/wrappers/python/spm/spm.py
+++ b/wrappers/python/spm/spm.py
@@ -38,7 +38,7 @@ class spmatrix():
if mtxtype_ == mtxtype.HerPosDef:
mtxtype_ = mtxtype.Hermitian
- self.spm_c = pyspm_spmatrix_t( 0, mtxtype_,
+ self.spm_c = pyspm_spmatrix_t( mtxtype_,
coeftype.Double,
fmttype.CSC,
0, 0, 0, 0, 0, 0, 0, 0,
--
GitLab
From 6290c5185f9c1368ffb5ce466fa0d80b87e5a86a Mon Sep 17 00:00:00 2001
From: Mathieu Faverge <mathieu.faverge@inria.fr>
Date: Tue, 3 Jul 2018 17:19:30 +0200
Subject: [PATCH 21/22] PRovide an spmAlloc function to allocate arrays in C
---
include/spm.h | 5 ++--
src/spm.c | 52 +++++++++++++++++++++++++++++++++
wrappers/fortran90/src/spmf.f90 | 18 ++++++++++++
wrappers/python/spm/__spm__.py | 4 +++
4 files changed, 77 insertions(+), 2 deletions(-)
diff --git a/include/spm.h b/include/spm.h
index 5f11776a..9d4ac2f9 100644
--- a/include/spm.h
+++ b/include/spm.h
@@ -81,8 +81,9 @@ typedef struct spmatrix_s {
* @name SPM basic subroutines
* @{
*/
-void spmInit( spmatrix_t *spm );
-void spmExit( spmatrix_t *spm );
+void spmInit ( spmatrix_t *spm );
+void spmAlloc( spmatrix_t *spm );
+void spmExit ( spmatrix_t *spm );
spmatrix_t *spmCopy( const spmatrix_t *spm );
void spmBase( spmatrix_t *spm, int baseval );
diff --git a/src/spm.c b/src/spm.c
index 64d74e91..5af62625 100644
--- a/src/spm.c
+++ b/src/spm.c
@@ -190,6 +190,58 @@ spmUpdateComputedFields( spmatrix_t *spm )
spm->gnnzexp = spm->nnzexp;
}
+/**
+ *******************************************************************************
+ *
+ * @brief Allocate the arrays of an spm structure with PaStiX internal
+ * allocations.
+ *
+ * This function must be called after initialization of the non-computed fields,
+ * and the call to spmUpdateComputedFields(). It allocates the colptr, rowptr,
+ * dof, and values arrays with PaStiX C malloc function. This is highly
+ * recommended to use this function when using PaStiX from Fortran.
+ *
+ *******************************************************************************
+ *
+ * @param[inout] spm
+ * The sparse matrix fr which the internal arrays needs to be allocated.
+ *
+ *******************************************************************************/
+void
+spmAlloc( spmatrix_t *spm )
+{
+ switch(spm->fmttype){
+ case SpmCSC:
+ colsize = spm->n + 1;
+ rowsize = spm->nnz;
+ valsize = spm->nnzexp;
+ dofsize = spm->n + 1;
+ break;
+ case SpmCSR:
+ colsize = spm->nnz;
+ rowsize = spm->n + 1;
+ valsize = spm->nnzexp;
+ dofsize = spm->n + 1;
+ break;
+ case SpmIJV:
+ default:
+ colsize = spm->nnz;
+ rowsize = spm->nnz;
+ valsize = spm->nnzexp;
+ dofsize = spm->n + 1;
+ }
+
+ spm->colptr = (spm_int_t*)malloc( colsize * sizeof(spm_int_t) );
+ spm->rowptr = (spm_int_t*)malloc( rowsize * sizeof(spm_int_t) );
+ if ( spm->dof < 1 ) {
+ spm->dofs = (spm_int_t*)malloc( dofsize * sizeof(spm_int_t) );
+ }
+ if(spm->values != NULL) {
+ valsize = valsize * spm_size_of( spm->flttype );
+ spm->values = malloc(valsize);
+ }
+}
+
/**
*******************************************************************************
*
diff --git a/wrappers/fortran90/src/spmf.f90 b/wrappers/fortran90/src/spmf.f90
index 639aaacf..ac0ce6d6 100644
--- a/wrappers/fortran90/src/spmf.f90
+++ b/wrappers/fortran90/src/spmf.f90
@@ -49,6 +49,16 @@ module spmf
end subroutine spmInit_c
end interface
+ interface
+ subroutine spmAlloc_c(spm) &
+ bind(c, name='spmAlloc')
+ use iso_c_binding
+ import spmatrix_t
+ implicit none
+ type(c_ptr), value :: spm
+ end subroutine spmAlloc_c
+ end interface
+
interface
subroutine spmExit_c(spm) &
bind(c, name='spmExit')
@@ -417,6 +427,14 @@ contains
call spmInit_c(c_loc(spm))
end subroutine spmInit
+ subroutine spmAlloc(spm)
+ use iso_c_binding
+ implicit none
+ type(spmatrix_t), intent(inout), target :: spm
+
+ call spmAlloc_c(c_loc(spm))
+ end subroutine spmAlloc
+
subroutine spmExit(spm)
use iso_c_binding
implicit none
diff --git a/wrappers/python/spm/__spm__.py b/wrappers/python/spm/__spm__.py
index 55116cd1..97148fc1 100644
--- a/wrappers/python/spm/__spm__.py
+++ b/wrappers/python/spm/__spm__.py
@@ -46,6 +46,10 @@ def pyspm_spmInit( spm ):
libspm.spmInit.argtypes = [ POINTER(pyspm_spmatrix_t) ]
libspm.spmInit( spm )
+def pyspm_spmAlloc( spm ):
+ libspm.spmAlloc.argtypes = [ POINTER(pyspm_spmatrix_t) ]
+ libspm.spmAlloc( spm )
+
def pyspm_spmExit( spm ):
libspm.spmExit.argtypes = [ POINTER(pyspm_spmatrix_t) ]
libspm.spmExit( spm )
--
GitLab
From 006fe1a920f92749d88ec702ab752e97e51c784f Mon Sep 17 00:00:00 2001
From: Mathieu Faverge <mathieu.faverge@inria.fr>
Date: Tue, 3 Jul 2018 17:37:39 +0200
Subject: [PATCH 22/22] Use spmalloc in spm_user
---
src/spm.c | 8 ++--
wrappers/fortran90/examples/spm_user.f90 | 59 +++++++++---------------
2 files changed, 25 insertions(+), 42 deletions(-)
diff --git a/src/spm.c b/src/spm.c
index 5af62625..c38e807d 100644
--- a/src/spm.c
+++ b/src/spm.c
@@ -210,6 +210,8 @@ spmUpdateComputedFields( spmatrix_t *spm )
void
spmAlloc( spmatrix_t *spm )
{
+ spm_int_t colsize, rowsize, valsize, dofsize;
+
switch(spm->fmttype){
case SpmCSC:
colsize = spm->n + 1;
@@ -236,10 +238,8 @@ spmAlloc( spmatrix_t *spm )
if ( spm->dof < 1 ) {
spm->dofs = (spm_int_t*)malloc( dofsize * sizeof(spm_int_t) );
}
- if(spm->values != NULL) {
- valsize = valsize * spm_size_of( spm->flttype );
- spm->values = malloc(valsize);
- }
+ valsize = valsize * spm_size_of( spm->flttype );
+ spm->values = malloc(valsize);
}
/**
diff --git a/wrappers/fortran90/examples/spm_user.f90 b/wrappers/fortran90/examples/spm_user.f90
index 8e5f8789..833bd409 100644
--- a/wrappers/fortran90/examples/spm_user.f90
+++ b/wrappers/fortran90/examples/spm_user.f90
@@ -15,9 +15,9 @@ program spm_user
use spmf
implicit none
- integer(kind=spm_int_t), dimension(:), allocatable, target :: rowptr
- integer(kind=spm_int_t), dimension(:), allocatable, target :: colptr
- real(kind=c_double), dimension(:), allocatable, target :: values
+ integer(kind=spm_int_t), dimension(:), pointer :: rowptr
+ integer(kind=spm_int_t), dimension(:), pointer :: colptr
+ real(kind=c_double), dimension(:), pointer :: values
real(kind=c_double), dimension(:,:), allocatable, target :: x0, x, b
type(c_ptr) :: x0_ptr, x_ptr, b_ptr
real(kind=c_double) :: eps = 1.e-15
@@ -26,7 +26,6 @@ program spm_user
integer(kind=spm_int_t) :: dim1, dim2, dim3, n, nnz
integer(kind=spm_int_t) :: i, j, k, l, nrhs
integer(c_int) :: info
- logical :: spm_modified
!
! Generate a 10x10x10 complex Laplacian in IJV format
@@ -37,9 +36,23 @@ program spm_user
n = dim1 * dim2 * dim3
nnz = (2*(dim1)-1) * dim2 * dim3 + (dim2-1)*dim1*dim3 + dim2*dim1*(dim3-1)
- allocate(rowptr(nnz))
- allocate(colptr(nnz))
- allocate(values(nnz))
+ !
+ ! Create the spm out of the internal data
+ !
+ call spmInit( spm )
+ spm%mtxtype = SpmSymmetric
+ spm%flttype = SpmDouble
+ spm%fmttype = SpmIJV
+ spm%n = n
+ spm%nnz = nnz
+ spm%dof = 1
+
+ call spmUpdateComputedFields( spm )
+ call spmAlloc( spm )
+
+ call c_f_pointer( spm%rowptr, rowptr, [nnz] )
+ call c_f_pointer( spm%colptr, colptr, [nnz] )
+ call c_f_pointer( spm%values, values, [nnz] )
l = 1
do i=1,dim1
@@ -97,33 +110,8 @@ program spm_user
write(6,*) 'l ', l, " nnz ", nnz
end if
- !
- ! Create the spm out of the internal data
- !
- call spmInit( spm )
- spm%mtxtype = SpmSymmetric
- spm%flttype = SpmDouble
- spm%fmttype = SpmIJV
- spm%n = n
- spm%nnz = nnz
- spm%dof = 1
- spm%rowptr = c_loc(rowptr)
- spm%colptr = c_loc(colptr)
- spm%values = c_loc(values)
-
- call spmUpdateComputedFields( spm )
-
call spmCheckAndCorrect( spm, spm2, info )
- spm_modified = .not. ( info .eq. 0 )
- if ( spm_modified ) then
- deallocate(rowptr)
- deallocate(colptr)
- deallocate(values)
-
- spm%rowptr = c_null_ptr
- spm%colptr = c_null_ptr
- spm%values = c_null_ptr
-
+ if ( info .ne. 0 ) then
call spmExit( spm )
spm = spm2
end if
@@ -151,11 +139,6 @@ program spm_user
call spmCheckAxb( eps, nrhs, spm, x0_ptr, spm%n, b_ptr, spm%n, x_ptr, spm%n, info )
call spmExit( spm )
- if ( .not. spm_modified ) then
- deallocate(rowptr)
- deallocate(colptr)
- deallocate(values)
- end if
deallocate(x0)
deallocate(x)
deallocate(b)
--
GitLab