Compare revisions

Sergey Kosukhin · Yen-Chen Chen · Yen-Chen Chen · Pradipta Samanta · fffe28d0 · fffe28d0
--- a/.gitlab-ci.yml
+++ b/.gitlab-ci.yml
@@ -5,6 +5,7 @@ include:
 stages:
  - lint
  - build_and_test
+  - build_and_test_OpenMP
  - build_OpenACC
  - coverage
  - prepare
@@ -94,6 +95,57 @@ nvhpc:
    - levante, hpc, dkrz
  needs: ["Check Typo", "Check OpenACC Style", "Check Style", "Check CMake Style", "Check License"]

+OpenMP_gcc11:
+  stage: build_and_test_OpenMP
+  extends: 
+    - .default
+  before_script:
+    - module list
+  script:
+    - module load gcc/11.2.0-gcc-11.2.0
+    - mkdir gcc112
+    - cd gcc112
+    - /sw/spack-levante/cmake-3.23.1-q5kzz6/bin/cmake .. -DCMAKE_BUILD_TYPE=RelWithDebInfo -DFS_ENABLE_OMP=ON -DCMAKE_C_COMPILER=gcc -DCMAKE_Fortran_COMPILER=gfortran -DCMAKE_CXX_COMPILER=g++
+    - make VERBOSE=1
+    - ctest --output-on-failure
+  tags:
+    - levante, hpc, dkrz
+  needs: ["gcc11"]
+
+OpenMP_intel22:
+  stage: build_and_test_OpenMP
+  extends: 
+    - .default
+  before_script:
+    - module list
+  script:
+    - module load gcc/11.2.0-gcc-11.2.0 intel-oneapi-compilers/2022.0.1-gcc-11.2.0
+    - mkdir intel22
+    - cd intel22
+    - /sw/spack-levante/cmake-3.23.1-q5kzz6/bin/cmake .. -DCMAKE_BUILD_TYPE=RelWithDebInfo -DFS_ENABLE_OMP=ON -DCMAKE_C_COMPILER=icc -DCMAKE_Fortran_COMPILER=ifort -DCMAKE_CXX_COMPILER=icpc -DCMAKE_BUILD_RPATH=/sw/spack-levante/gcc-11.2.0-bcn7mb/lib64
+    - make VERBOSE=1
+    - ctest --output-on-failure
+  tags:
+    - levante, hpc, dkrz
+  needs: ["intel22"]
+
+OpenMP_nvhpc:
+  stage: build_and_test_OpenMP
+  extends: 
+    - .default
+  before_script:
+    - module list
+  script:
+    - module load gcc/11.2.0-gcc-11.2.0 nvhpc/22.5-gcc-11.2.0
+    - mkdir nvhpc
+    - cd nvhpc
+    - /sw/spack-levante/cmake-3.23.1-q5kzz6/bin/cmake .. -DCMAKE_BUILD_TYPE=RelWithDebInfo -DFS_ENABLE_OMP=ON -DCMAKE_C_COMPILER=nvc -DCMAKE_Fortran_COMPILER=nvfortran -DCMAKE_CXX_COMPILER=nvc++
+    - make VERBOSE=1
+    - ctest --output-on-failure
+  tags:
+    - levante, hpc, dkrz
+  needs: ["nvhpc"]
+
 OpenACC:
  stage: build_OpenACC
  extends: 
@@ -251,7 +303,7 @@ Prepare Changelog:
  artifacts:
    paths:
    - release_notes.md
-  needs: ["nag", "gcc11", "intel22", "nvhpc", "OpenACC"]
+  needs: ["nag", "OpenMP_gcc11", "OpenMP_intel22", "OpenMP_nvhpc", "OpenACC"]

 Release:
  stage: release

--- a/src/CMakeLists.txt
+++ b/src/CMakeLists.txt
@@ -15,14 +15,12 @@ check_include_files("link.h" HAVE_LINK_H)
 check_include_files("unwind.h" HAVE_UNWIND_H)
 check_include_files("sys/resource.h" HAVE_GETRUSAGE)

+include(TestBigEndian)
+test_big_endian(HAVE_BIG_ENDIAN)
+
 configure_file(${CMAKE_CURRENT_SOURCE_DIR}/config.h.in
               ${CMAKE_CURRENT_BINARY_DIR}/config.h)

-if(FS_ENABLE_OMP)
-  find_package(OpenMP REQUIRED)
-  set(CMAKE_Fortran_FLAGS "${CMAKE_Fortran_FLAGS} ${OpenMP_Fortran_FLAGS}")
-endif()
-
 add_library(
  fortran-support
  ${CMAKE_CURRENT_BINARY_DIR}/config.h
@@ -87,6 +85,11 @@ if(FS_ENABLE_SINGLE_PRECISION)
  target_compile_definitions(fortran-support PRIVATE __SINGLE_PRECISION)
 endif()

+if(FS_ENABLE_OMP)
+  find_package(OpenMP COMPONENTS Fortran REQUIRED)
+  target_link_libraries(fortran-support PUBLIC OpenMP::OpenMP_Fortran)
+endif()
+
 include("${PROJECT_SOURCE_DIR}/cmake/check_macro.cmake")
 check_macro_defined(HAS_OPENACC_MACRO _OPENACC QUIET)
 if(FS_ENABLE_OPENACC)

--- a/src/config.h.in
+++ b/src/config.h.in
@@ -16,3 +16,5 @@
 #cmakedefine HAVE_UNWIND_H @HAVE_UNWIND_H@

 #cmakedefine HAVE_GETRUSAGE @HAVE_GETRUSAGE@
+
+#cmakedefine HAVE_BIG_ENDIAN @HAVE_BIG_ENDIAN@
--- a/src/mo_fortran_tools.F90
+++ b/src/mo_fortran_tools.F90
@@ -2224,26 +2224,12 @@ CONTAINS
        CALL util_stride_1d(in_stride(2), elem_byte_size, &
                            C_LOC(ptr_in(1, 1)), C_LOC(ptr_in(1, 2)))
        base_shape(1) = in_stride(2)
+      ELSE
+        base_shape(1) = in_shape(1)
      END IF
      base_shape(2) = in_shape(2)
      CALL insert_dimension_r_dp_3_2_s(ptr_out, ptr_in(1, 1), &
                                       base_shape, new_dim_rank)
-      IF (in_stride(1) > 1 .OR. in_stride(2) > in_shape(1) &
-          .OR. base_shape(1) /= in_shape(1)) THEN
-        out_stride(1) = in_stride(1)
-        out_stride(2) = 1
-        out_shape(1:out_rank - 1) = in_shape
-        DO i = out_rank, new_dim_rank + 1, -1
-          out_shape(i) = out_shape(i - 1)
-          out_stride(i) = out_stride(i - 1)
-        END DO
-        out_stride(new_dim_rank) = 1
-        out_shape(new_dim_rank) = 1
-        out_shape = (out_shape - 1)*out_stride + 1
-        ptr_out => ptr_out(:out_shape(1):out_stride(1), &
-             &             :out_shape(2):out_stride(2), &
-             &             :out_shape(3):out_stride(3))
-      END IF
    ELSE
      out_shape(1:out_rank - 1) = SHAPE(ptr_in)
      DO i = out_rank, new_dim_rank + 1, -1
@@ -2299,26 +2285,12 @@ CONTAINS
        CALL util_stride_1d(in_stride(2), elem_byte_size, &
                            C_LOC(ptr_in(1, 1)), C_LOC(ptr_in(1, 2)))
        base_shape(1) = in_stride(2)
+      ELSE
+        base_shape(1) = in_shape(1)
      END IF
      base_shape(2) = in_shape(2)
      CALL insert_dimension_r_sp_3_2_s(ptr_out, ptr_in(1, 1), &
                                       base_shape, new_dim_rank)
-      IF (in_stride(1) > 1 .OR. in_stride(2) > in_shape(1) &
-          .OR. base_shape(1) /= in_shape(1)) THEN
-        out_stride(1) = in_stride(1)
-        out_stride(2) = 1
-        out_shape(1:out_rank - 1) = in_shape
-        DO i = out_rank, new_dim_rank + 1, -1
-          out_shape(i) = out_shape(i - 1)
-          out_stride(i) = out_stride(i - 1)
-        END DO
-        out_stride(new_dim_rank) = 1
-        out_shape(new_dim_rank) = 1
-        out_shape = (out_shape - 1)*out_stride + 1
-        ptr_out => ptr_out(:out_shape(1):out_stride(1), &
-             &             :out_shape(2):out_stride(2), &
-             &             :out_shape(3):out_stride(3))
-      END IF
    ELSE
      out_shape(1:out_rank - 1) = SHAPE(ptr_in)
      DO i = out_rank, new_dim_rank + 1, -1
@@ -2374,26 +2346,12 @@ CONTAINS
        CALL util_stride_1d(in_stride(2), elem_byte_size, &
                            C_LOC(ptr_in(1, 1)), C_LOC(ptr_in(1, 2)))
        base_shape(1) = in_stride(2)
+      ELSE
+        base_shape(1) = in_shape(1)
      END IF
      base_shape(2) = in_shape(2)
      CALL insert_dimension_i4_3_2_s(ptr_out, ptr_in(1, 1), &
                                     base_shape, new_dim_rank)
-      IF (in_stride(1) > 1 .OR. in_stride(2) > in_shape(1) &
-          .OR. base_shape(1) /= in_shape(1)) THEN
-        out_stride(1) = in_stride(1)
-        out_stride(2) = 1
-        out_shape(1:out_rank - 1) = in_shape
-        DO i = out_rank, new_dim_rank + 1, -1
-          out_shape(i) = out_shape(i - 1)
-          out_stride(i) = out_stride(i - 1)
-        END DO
-        out_stride(new_dim_rank) = 1
-        out_shape(new_dim_rank) = 1
-        out_shape = (out_shape - 1)*out_stride + 1
-        ptr_out => ptr_out(:out_shape(1):out_stride(1), &
-             &             :out_shape(2):out_stride(2), &
-             &             :out_shape(3):out_stride(3))
-      END IF
    ELSE
      out_shape(1:out_rank - 1) = SHAPE(ptr_in)
      DO i = out_rank, new_dim_rank + 1, -1
@@ -2447,26 +2405,12 @@ CONTAINS
        CALL util_stride_1d(in_stride(2), elem_byte_size, &
                            C_LOC(ptr_in(1, 1)), C_LOC(ptr_in(1, 2)))
        base_shape(1) = in_stride(2)
+      ELSE
+        base_shape(1) = in_shape(1)
      END IF
      base_shape(2) = in_shape(2)
      CALL insert_dimension_l_3_2_s(ptr_out, ptr_in(1, 1), &
                                    base_shape, new_dim_rank)
-      IF (in_stride(1) > 1 .OR. in_stride(2) > in_shape(1) &
-          .OR. base_shape(1) /= in_shape(1)) THEN
-        out_stride(1) = in_stride(1)
-        out_stride(2) = 1
-        out_shape(1:out_rank - 1) = in_shape
-        DO i = out_rank, new_dim_rank + 1, -1
-          out_shape(i) = out_shape(i - 1)
-          out_stride(i) = out_stride(i - 1)
-        END DO
-        out_stride(new_dim_rank) = 1
-        out_shape(new_dim_rank) = 1
-        out_shape = (out_shape - 1)*out_stride + 1
-        ptr_out => ptr_out(:out_shape(1):out_stride(1), &
-             &             :out_shape(2):out_stride(2), &
-             &             :out_shape(3):out_stride(3))
-      END IF
    ELSE
      out_shape(1:out_rank - 1) = SHAPE(ptr_in)
      DO i = out_rank, new_dim_rank + 1, -1

--- a/src/util_hash.c
+++ b/src/util_hash.c
@@ -42,63 +42,21 @@
 * -------------------------------------------------------------------------------
 */

-#define _POSIX_C_SOURCE 200112L
+#include "config.h"

-#include <stdint.h> /* defines uint32_t etc */
-
-#include <sys/param.h> /* attempt to define endianness */
-
-#if (defined(__APPLE__) && defined(__MACH__))
-
-#if (defined(__DARWIN_BYTE_ORDER) \
-     && (__DARWIN_BYTE_ORDER == __DARWIN_LITTLE_ENDIAN))
-#define HASH_LITTLE_ENDIAN 1
-#define HASH_BIG_ENDIAN 0
-#else
-#define HASH_LITTLE_ENDIAN 0
-#define HASH_BIG_ENDIAN 1
-#endif
+#include "util_hash.h"

-#elif defined(_AIX)
+#include <stddef.h>
+#include <stdint.h>

-#if (defined(BYTE_ORDER) && (BYTE_ORDER == LITTLE_ENDIAN))
-#define HASH_LITTLE_ENDIAN 1
-#define HASH_BIG_ENDIAN 0
-#else
+#ifdef HAVE_BIG_ENDIAN
 #define HASH_LITTLE_ENDIAN 0
 #define HASH_BIG_ENDIAN 1
-#endif
-
-#elif defined(__linux)
-
-#include <endian.h>
-
-#if (defined(__BYTE_ORDER) && (__BYTE_ORDER == __LITTLE_ENDIAN))
-#define HASH_LITTLE_ENDIAN 1
-#define HASH_BIG_ENDIAN 0
 #else
-#define HASH_LITTLE_ENDIAN 0
-#define HASH_BIG_ENDIAN 1
-#endif
-
-#elif defined(_SX)
-
-#define HASH_LITTLE_ENDIAN 0
-#define HASH_BIG_ENDIAN 1
-
-#elif defined(__NEC__)
-
 #define HASH_LITTLE_ENDIAN 1
 #define HASH_BIG_ENDIAN 0
-
-#else
-
-#error "Couldn't determine endianness."
-
 #endif

-#include "util_hash.h"
-
 #define rot(x, k) (((x) << (k)) | ((x) >> (32 - (k))))

 /*

--- a/src/util_hash.h
+++ b/src/util_hash.h
@@ -12,6 +12,7 @@
 #ifndef UTIL_HASH_H
 #define UTIL_HASH_H

+#include <stddef.h>
 #include <stdint.h>

 #ifdef __cplusplus
No results found