Merge pull request #5066 from OpenMathLib/develop

Merge changes from develop in preparation of the 0.3.29 release

Author: martin-frbg

.cirrus.yml

@@ -89,20 +89,13 @@ task:
       type: text/plain
 
 macos_instance:
-  image: ghcr.io/cirruslabs/macos-monterey-xcode:latest
+  image: ghcr.io/cirruslabs/macos-sonoma-xcode:latest
 task:
   name: AppleM1/LLVM armv7-androidndk xbuild
   compile_script:
-  - brew install android-ndk
-  - export #PATH=/opt/homebrew/opt/llvm/bin:$PATH
-  - export #LDFLAGS="-L/opt/homebrew/opt/llvm/lib"
-  - export #CPPFLAGS="-I/opt/homebrew/opt/llvm/include" 
-  - ls /System/Volumes/Data/opt/homebrew
-  - ls -l /System/Volumes/Data/opt/homebrew/Caskroom/android-ndk
-  - find /System/Volumes/Data/opt/homebrew -name "armv7a-linux-androideabi*-ranlib"
-  - #export CC=/Applications/Xcode-13.4.1.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/clang
-  - #export CFLAGS="-O2 -unwindlib=none -Wno-macro-redefined -isysroot /Applications/Xcode-13.4.1.app/Contents/Developer/Platforms/iPhoneOS.platform/Developer/SDKs/iPhoneOS16.0.sdk -arch arm64 -miphoneos-version-min=10.0"
-  - export CC=/System/Volumes/Data/opt/homebrew/Caskroom/android-ndk/26d/AndroidNDK*.app/Contents/NDK/toolchains/llvm/prebuilt/darwin-x86_64/bin/armv7a-linux-androideabi23-clang
+  - brew install --cask android-ndk
+  - export ANDROID_NDK_HOME="/opt/homebrew/share/android-ndk"
+  - export CC=/opt/homebrew/share/android-ndk/toolchains/llvm/prebuilt/darwin-x86_64/bin/armv7a-linux-androideabi23-clang
   - make TARGET=ARMV7 ARM_SOFTFP_ABI=1 NUM_THREADS=32 HOSTCC=clang NOFORTRAN=1 RANLIB="ls -l"
   always:
     config_artifacts:
@@ -132,9 +125,9 @@ task:
   - make USE_OPENMP=1
 
 FreeBSD_task:
-  name: FreeBSD-gcc12
+  name: FreeBSD-gcc
   freebsd_instance:
-    image_family: freebsd-13-3
+    image_family: freebsd-14-1
   install_script:
   - pkg update -f && pkg upgrade -y && pkg install -y gmake gcc
   compile_script:
@@ -143,9 +136,9 @@ FreeBSD_task:
 
 
 FreeBSD_task:
-  name: freebsd-gcc12-ilp64
+  name: freebsd-gcc-ilp64
   freebsd_instance:
-    image_family: freebsd-13-3
+    image_family: freebsd-14-1
   install_script:
   - pkg update -f && pkg upgrade -y && pkg install -y gmake gcc
   compile_script:
@@ -155,7 +148,7 @@ FreeBSD_task:
 FreeBSD_task:
   name: FreeBSD-clang-openmp
   freebsd_instance:
-    image_family: freebsd-13-3
+    image_family: freebsd-14-1
   install_script:
   - pkg update -f && pkg upgrade -y && pkg install -y gmake gcc 
   - ln -s /usr/local/lib/gcc13/libgfortran.so.5.0.0 /usr/lib/libgfortran.so

.github/workflows/docs.yml

@@ -23,7 +23,7 @@ jobs:
           python-version: "3.10"
 
       - name: Install MkDocs and doc theme packages
-        run: pip install mkdocs mkdocs-material mkdocs-git-revision-date-localized-plugin
+        run: pip install mkdocs mkdocs-material mkdocs-git-revision-date-localized-plugin mkdocs-mermaid2-plugin
 
       - name: Build docs site
         run: mkdocs build

.github/workflows/dynamic_arch.yml

@@ -158,7 +158,7 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        msystem: [UCRT64, MINGW32, CLANG64, CLANG32]
+        msystem: [UCRT64, MINGW32, CLANG64]
         idx: [int32, int64]
         build-type: [Release]
         include:
@@ -174,14 +174,6 @@ jobs:
             idx: int32
             target-prefix: mingw-w64-clang-x86_64
             fc-pkg: fc
-            # Compiling with Flang 16 seems to cause test errors on machines
-            # with AVX512 instructions. Revisit after MSYS2 distributes Flang 17.
-            no-avx512-flags: -DNO_AVX512=1
-          - msystem: CLANG32
-            idx: int32
-            target-prefix: mingw-w64-clang-i686
-            fc-pkg: cc
-            c-lapack-flags: -DC_LAPACK=ON
           - msystem: UCRT64
             idx: int64
             idx64-flags: -DBINARY=64 -DINTERFACE64=1
@@ -192,9 +184,6 @@ jobs:
             idx64-flags: -DBINARY=64 -DINTERFACE64=1
             target-prefix: mingw-w64-clang-x86_64
             fc-pkg: fc
-            # Compiling with Flang 16 seems to cause test errors on machines
-            # with AVX512 instructions. Revisit after MSYS2 distributes Flang 17.
-            no-avx512-flags: -DNO_AVX512=1
           - msystem: UCRT64
             idx: int32
             target-prefix: mingw-w64-ucrt-x86_64
@@ -203,8 +192,6 @@ jobs:
         exclude:
           - msystem: MINGW32
             idx: int64
-          - msystem: CLANG32
-            idx: int64
 
     defaults:
       run:
@@ -280,8 +267,6 @@ jobs:
                 -DNUM_THREADS=64 \
                 -DTARGET=CORE2 \
                 ${{ matrix.idx64-flags }} \
-                ${{ matrix.c-lapack-flags }} \
-                ${{ matrix.no-avx512-flags }} \
                 -DCMAKE_C_COMPILER_LAUNCHER=ccache \
                 -DCMAKE_Fortran_COMPILER_LAUNCHER=ccache \
                 ..

.github/workflows/harmonyos.yml

@@ -0,0 +1,37 @@
+name: harmonyos
+
+on: [push, pull_request]
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.head_ref || github.run_id }}
+  cancel-in-progress: true
+
+permissions:
+  contents: read # to fetch code (actions/checkout)
+
+jobs:
+  build:
+    if: "github.repository == 'OpenMathLib/OpenBLAS'"
+    runs-on: ubuntu-latest
+    env:
+      OHOS_NDK_CMAKE: $GITHUB_WORKSPACE/ohos-sdk/linux/native/build-tools/cmake/bin/cmake
+      COMMON_CMAKE_OPTIONS: |
+        -DCMAKE_TOOLCHAIN_FILE=$GITHUB_WORKSPACE/ohos-sdk/linux/native/build/cmake/ohos.toolchain.cmake \
+        -DCMAKE_INSTALL_PREFIX=install \
+        -DCMAKE_BUILD_TYPE=Release \
+    steps:
+    - uses: actions/checkout@v4
+    - name: ndk-install
+      run: |
+        wget https://repo.huaweicloud.com/harmonyos/os/4.1.1-Release/ohos-sdk-windows_linux-public.tar.gz
+        tar -xf ohos-sdk-windows_linux-public.tar.gz
+        cd ohos-sdk/linux
+        unzip -q native-linux-x64-4.1.7.8-Release.zip
+        cd -
+    - name: build-armv8
+      run: |
+       mkdir build && cd build
+       ${{ env.OHOS_NDK_CMAKE }} ${{ env.COMMON_CMAKE_OPTIONS }} -DOHOS_ARCH="arm64-v8a" \
+       -DTARGET=ARMV8 -DNOFORTRAN=1 ..
+       ${{ env.OHOS_NDK_CMAKE }} --build . -j $(nproc)
+       

.github/workflows/loongarch64.yml

@@ -9,22 +9,31 @@ concurrency:
 jobs:
   TEST:
     if: "github.repository == 'OpenMathLib/OpenBLAS'"
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-24.04
     strategy:
       fail-fast: false
       matrix:
         include:
           - target: LOONGSONGENERIC
-            triple:  loongarch64-unknown-linux-gnu
+            triple:  loongarch64-linux-gnu
             opts: NO_SHARED=1 DYNAMIC_ARCH=1 TARGET=LOONGSONGENERIC
           - target: LOONGSON3R5
-            triple: loongarch64-unknown-linux-gnu
+            triple: loongarch64-linux-gnu
             opts: NO_SHARED=1 DYNAMIC_ARCH=1 TARGET=LOONGSON3R5
           - target: LOONGSON2K1000
-            triple: loongarch64-unknown-linux-gnu
+            triple: loongarch64-linux-gnu
             opts: NO_SHARED=1 DYNAMIC_ARCH=1 TARGET=LOONGSON2K1000
+          - target: LA64_GENERIC
+            triple:  loongarch64-linux-gnu
+            opts: NO_SHARED=1 DYNAMIC_ARCH=1 TARGET=LA64_GENERIC
+          - target: LA464
+            triple: loongarch64-linux-gnu
+            opts: NO_SHARED=1 DYNAMIC_ARCH=1 TARGET=LA464
+          - target: LA264
+            triple: loongarch64-linux-gnu
+            opts: NO_SHARED=1 DYNAMIC_ARCH=1 TARGET=LA264
           - target: DYNAMIC_ARCH
-            triple: loongarch64-unknown-linux-gnu
+            triple: loongarch64-linux-gnu
             opts: NO_SHARED=1 DYNAMIC_ARCH=1 TARGET=GENERIC
 
     steps:
@@ -33,33 +42,9 @@ jobs:
 
       - name: Install APT deps
         run: |
-          sudo apt-get update
-          sudo apt-get install autoconf automake autotools-dev ninja-build make ccache
-
-      - name: Download and install loongarch64-toolchain
-        run: |
-          wget https://github.com/sunhaiyong1978/CLFS-for-LoongArch/releases/download/8.1/CLFS-loongarch64-8.1-x86_64-cross-tools-gcc-glibc.tar.xz
-          #wget https://github.com/loongson/build-tools/releases/download/2023.08.08/CLFS-loongarch64-8.1-x86_64-cross-tools-gcc-glibc.tar.xz
-          tar -xf CLFS-loongarch64-8.1-x86_64-cross-tools-gcc-glibc.tar.xz -C /opt
-
-      - name: Checkout qemu
-        uses: actions/checkout@v3
-        with:
-          repository: qemu/qemu
-          path: qemu
-          ref: master
-
-      - name: Install qemu
-        run: |
-          cd qemu
-          ./configure --prefix=$GITHUB_WORKSPACE/qemu-install --target-list=loongarch64-linux-user --disable-system --static
-          make -j$(nproc)
-          make install
-
-      - name: Set env
-        run: |
-          echo "LD_LIBRARY_PATH=/opt/cross-tools/target/usr/lib64:/opt/cross-tools/loongarch64-unknown-linux-gnu/lib64:$LD_LIBRARY_PATH" >> $GITHUB_ENV
-          echo "PATH=$GITHUB_WORKSPACE:/opt/cross-tools/bin:$PATH" >> $GITHUB_ENV
+          sudo apt-get update && \
+          sudo apt-get install autoconf automake autotools-dev ninja-build make ccache qemu-user-static \
+                               gcc-14-loongarch64-linux-gnu g++-14-loongarch64-linux-gnu gfortran-14-loongarch64-linux-gnu
 
       - name: Compilation cache
         uses: actions/cache@v3
@@ -80,54 +65,55 @@ jobs:
       - name: Disable utest dsdot:dsdot_n_1
         run: |
           echo -n > utest/test_dsdot.c
-          echo "Due to the qemu versions 7.2 causing utest cases to fail,"
+          echo "Due to the current version of qemu causing utest cases to fail,"
           echo "the utest dsdot:dsdot_n_1 have been temporarily disabled."
 
       - name: Build OpenBLAS
-        run: make CC='ccache ${{ matrix.triple }}-gcc -static' FC='ccache ${{ matrix.triple }}-gfortran -static' ${{ matrix.opts }} HOSTCC='ccache gcc' -j$(nproc)
+        run: |
+          make CC='ccache ${{ matrix.triple }}-gcc-14 -static' FC='ccache ${{ matrix.triple }}-gfortran-14 -static' \
+          RANLIB='ccache ${{ matrix.triple }}-gcc-ranlib-14' ${{ matrix.opts }} HOSTCC='ccache gcc' -j$(nproc)
 
       - name: Test
         run: |
-          export PATH=$GITHUB_WORKSPACE/qemu-install/bin/:$PATH
-          qemu-loongarch64 ./utest/openblas_utest
-          qemu-loongarch64 ./utest/openblas_utest_ext
-          OPENBLAS_NUM_THREADS=2 qemu-loongarch64 ./ctest/xscblat1
-          OPENBLAS_NUM_THREADS=2 qemu-loongarch64 ./ctest/xdcblat1
-          OPENBLAS_NUM_THREADS=2 qemu-loongarch64 ./ctest/xccblat1
-          OPENBLAS_NUM_THREADS=2 qemu-loongarch64 ./ctest/xzcblat1
-          OPENBLAS_NUM_THREADS=2 qemu-loongarch64 ./ctest/xscblat2 < ./ctest/sin2
-          OPENBLAS_NUM_THREADS=2 qemu-loongarch64 ./ctest/xdcblat2 < ./ctest/din2
-          OPENBLAS_NUM_THREADS=2 qemu-loongarch64 ./ctest/xccblat2 < ./ctest/cin2
-          OPENBLAS_NUM_THREADS=2 qemu-loongarch64 ./ctest/xzcblat2 < ./ctest/zin2
-          OPENBLAS_NUM_THREADS=2 qemu-loongarch64 ./ctest/xscblat3 < ./ctest/sin3
-          OPENBLAS_NUM_THREADS=2 qemu-loongarch64 ./ctest/xdcblat3 < ./ctest/din3
-          OPENBLAS_NUM_THREADS=2 qemu-loongarch64 ./ctest/xccblat3 < ./ctest/cin3
-          OPENBLAS_NUM_THREADS=2 qemu-loongarch64 ./ctest/xzcblat3 < ./ctest/zin3
-          OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-loongarch64 ./test/sblat1
-          OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-loongarch64 ./test/dblat1
-          OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-loongarch64 ./test/cblat1
-          OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-loongarch64 ./test/zblat1
-          OPENBLAS_NUM_THREADS=2 qemu-loongarch64 ./test/sblat1
-          OPENBLAS_NUM_THREADS=2 qemu-loongarch64 ./test/dblat1
-          OPENBLAS_NUM_THREADS=2 qemu-loongarch64 ./test/cblat1
-          OPENBLAS_NUM_THREADS=2 qemu-loongarch64 ./test/zblat1
+          qemu-loongarch64-static ./utest/openblas_utest
+          qemu-loongarch64-static ./utest/openblas_utest_ext
+          OPENBLAS_NUM_THREADS=2 qemu-loongarch64-static ./ctest/xscblat1
+          OPENBLAS_NUM_THREADS=2 qemu-loongarch64-static ./ctest/xdcblat1
+          OPENBLAS_NUM_THREADS=2 qemu-loongarch64-static ./ctest/xccblat1
+          OPENBLAS_NUM_THREADS=2 qemu-loongarch64-static ./ctest/xzcblat1
+          OPENBLAS_NUM_THREADS=2 qemu-loongarch64-static ./ctest/xscblat2 < ./ctest/sin2
+          OPENBLAS_NUM_THREADS=2 qemu-loongarch64-static ./ctest/xdcblat2 < ./ctest/din2
+          OPENBLAS_NUM_THREADS=2 qemu-loongarch64-static ./ctest/xccblat2 < ./ctest/cin2
+          OPENBLAS_NUM_THREADS=2 qemu-loongarch64-static ./ctest/xzcblat2 < ./ctest/zin2
+          OPENBLAS_NUM_THREADS=2 qemu-loongarch64-static ./ctest/xscblat3 < ./ctest/sin3
+          OPENBLAS_NUM_THREADS=2 qemu-loongarch64-static ./ctest/xdcblat3 < ./ctest/din3
+          OPENBLAS_NUM_THREADS=2 qemu-loongarch64-static ./ctest/xccblat3 < ./ctest/cin3
+          OPENBLAS_NUM_THREADS=2 qemu-loongarch64-static ./ctest/xzcblat3 < ./ctest/zin3
+          OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-loongarch64-static ./test/sblat1
+          OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-loongarch64-static ./test/dblat1
+          OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-loongarch64-static ./test/cblat1
+          OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-loongarch64-static ./test/zblat1
+          OPENBLAS_NUM_THREADS=2 qemu-loongarch64-static ./test/sblat1
+          OPENBLAS_NUM_THREADS=2 qemu-loongarch64-static ./test/dblat1
+          OPENBLAS_NUM_THREADS=2 qemu-loongarch64-static ./test/cblat1
+          OPENBLAS_NUM_THREADS=2 qemu-loongarch64-static ./test/zblat1
           rm -f ./test/?BLAT2.SUMM
-          OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-loongarch64 ./test/sblat2 < ./test/sblat2.dat
-          OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-loongarch64 ./test/dblat2 < ./test/dblat2.dat
-          OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-loongarch64 ./test/cblat2 < ./test/cblat2.dat
-          OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-loongarch64 ./test/zblat2 < ./test/zblat2.dat
+          OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-loongarch64-static ./test/sblat2 < ./test/sblat2.dat
+          OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-loongarch64-static ./test/dblat2 < ./test/dblat2.dat
+          OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-loongarch64-static ./test/cblat2 < ./test/cblat2.dat
+          OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-loongarch64-static ./test/zblat2 < ./test/zblat2.dat
           rm -f ./test/?BLAT2.SUMM
-          OPENBLAS_NUM_THREADS=2 qemu-loongarch64 ./test/sblat2 < ./test/sblat2.dat
-          OPENBLAS_NUM_THREADS=2 qemu-loongarch64 ./test/dblat2 < ./test/dblat2.dat
-          OPENBLAS_NUM_THREADS=2 qemu-loongarch64 ./test/cblat2 < ./test/cblat2.dat
-          OPENBLAS_NUM_THREADS=2 qemu-loongarch64 ./test/zblat2 < ./test/zblat2.dat
+          OPENBLAS_NUM_THREADS=2 qemu-loongarch64-static ./test/sblat2 < ./test/sblat2.dat
+          OPENBLAS_NUM_THREADS=2 qemu-loongarch64-static ./test/dblat2 < ./test/dblat2.dat
+          OPENBLAS_NUM_THREADS=2 qemu-loongarch64-static ./test/cblat2 < ./test/cblat2.dat
+          OPENBLAS_NUM_THREADS=2 qemu-loongarch64-static ./test/zblat2 < ./test/zblat2.dat
           rm -f ./test/?BLAT3.SUMM
-          OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-loongarch64 ./test/sblat3 < ./test/sblat3.dat
-          OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-loongarch64 ./test/dblat3 < ./test/dblat3.dat
-          OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-loongarch64 ./test/cblat3 < ./test/cblat3.dat
-          OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-loongarch64 ./test/zblat3 < ./test/zblat3.dat
+          OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-loongarch64-static ./test/sblat3 < ./test/sblat3.dat
+          OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-loongarch64-static ./test/dblat3 < ./test/dblat3.dat
+          OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-loongarch64-static ./test/cblat3 < ./test/cblat3.dat
+          OPENBLAS_NUM_THREADS=1 OMP_NUM_THREADS=1 qemu-loongarch64-static ./test/zblat3 < ./test/zblat3.dat
           rm -f ./test/?BLAT3.SUMM
-          OPENBLAS_NUM_THREADS=2 qemu-loongarch64 ./test/sblat3 < ./test/sblat3.dat
-          OPENBLAS_NUM_THREADS=2 qemu-loongarch64 ./test/dblat3 < ./test/dblat3.dat
-          OPENBLAS_NUM_THREADS=2 qemu-loongarch64 ./test/cblat3 < ./test/cblat3.dat
-          OPENBLAS_NUM_THREADS=2 qemu-loongarch64 ./test/zblat3 < ./test/zblat3.dat
+          OPENBLAS_NUM_THREADS=2 qemu-loongarch64-static ./test/sblat3 < ./test/sblat3.dat
+          OPENBLAS_NUM_THREADS=2 qemu-loongarch64-static ./test/dblat3 < ./test/dblat3.dat
+          OPENBLAS_NUM_THREADS=2 qemu-loongarch64-static ./test/cblat3 < ./test/cblat3.dat
+          OPENBLAS_NUM_THREADS=2 qemu-loongarch64-static ./test/zblat3 < ./test/zblat3.dat

.github/workflows/loongarch64_clang.yml

@@ -20,6 +20,12 @@ jobs:
             opts: NO_SHARED=1 DYNAMIC_ARCH=1 TARGET=LOONGSON3R5
           - target: LOONGSON2K1000
             opts: NO_SHARED=1 DYNAMIC_ARCH=1 TARGET=LOONGSON2K1000
+          - target: LA64_GENERIC
+            opts: NO_SHARED=1 DYNAMIC_ARCH=1 TARGET=LA64_GENERIC
+          - target: LA464
+            opts: NO_SHARED=1 DYNAMIC_ARCH=1 TARGET=LA464
+          - target: LA264
+            opts: NO_SHARED=1 DYNAMIC_ARCH=1 TARGET=LA264
           - target: DYNAMIC_ARCH
             opts: NO_SHARED=1 DYNAMIC_ARCH=1 TARGET=GENERIC
 

.github/workflows/nightly-Homebrew-build.yml

@@ -69,7 +69,7 @@ jobs:
           mv *.bottle.tar.gz bottles
 
       - name: Upload bottle
-        uses: actions/upload-artifact@v1
+        uses: actions/upload-artifact@v4
         with:
           name: openblas--HEAD.catalina.bottle.tar.gz
           path: bottles

CMakeLists.txt

@@ -2,13 +2,14 @@
 ## Author: Hank Anderson <hank@statease.com>
 ##
 
-cmake_minimum_required(VERSION 2.8.5)
+cmake_minimum_required(VERSION 3.16.0)
 
+set (CMAKE_ASM_SOURCE_FILE_EXTENSIONS "S")
 project(OpenBLAS C ASM)
 
 set(OpenBLAS_MAJOR_VERSION 0)
 set(OpenBLAS_MINOR_VERSION 3)
-set(OpenBLAS_PATCH_VERSION 28)
+set(OpenBLAS_PATCH_VERSION 28.dev)
 
 set(OpenBLAS_VERSION "${OpenBLAS_MAJOR_VERSION}.${OpenBLAS_MINOR_VERSION}.${OpenBLAS_PATCH_VERSION}")
 
@@ -102,6 +103,10 @@ endif()
 
 message(WARNING "CMake support is experimental. It does not yet support all build options and may not produce the same Makefiles that OpenBLAS ships with.")
 
+if (USE_OPENMP)
+  find_package(OpenMP REQUIRED)
+endif ()
+
 include("${PROJECT_SOURCE_DIR}/cmake/utils.cmake")
 include("${PROJECT_SOURCE_DIR}/cmake/system.cmake")
 
@@ -258,8 +263,13 @@ if (${CMAKE_SYSTEM_NAME} MATCHES "AIX|Android|Linux|FreeBSD|OpenBSD|NetBSD|Drago
   endif()
 endif()
 
-if (APPLE AND BUILD_SHARED_LIBS)
-set(CMAKE_MACOSX_RPATH ON)
+if (USE_OPENMP)
+  if(BUILD_STATIC_LIBS)
+    target_link_libraries(${OpenBLAS_LIBNAME}_static OpenMP::OpenMP_C)
+  endif()
+  if(BUILD_SHARED_LIBS)
+    target_link_libraries(${OpenBLAS_LIBNAME}_shared OpenMP::OpenMP_C)
+  endif()
 endif()
 
 # Seems that this hack doesn't required since macOS 11 Big Sur