Merge branch 'OpenMathLib:develop' into issue4905

Author: martin-frbg

.cirrus.yml

@@ -94,16 +94,8 @@ task:
   name: AppleM1/LLVM armv7-androidndk xbuild
   compile_script:
   - brew install --cask 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"
   - export ANDROID_NDK_HOME="/opt/homebrew/share/android-ndk"
-  - ls /opt/homebrew
-  - ls -l /System/Volumes/Data/opt/homebrew/Caskroom/android-ndk
-  - find /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/27/AndroidNDK*.app/Contents/NDK/toolchains/llvm/prebuilt/darwin-x86_64/bin/armv7a-linux-androideabi23-clang
+  - 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:

common_arm64.h

@@ -55,6 +55,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 #ifndef ASSEMBLER
 
 
+#ifndef NO_AFFINITY
 static __inline int WhereAmI(void){
   uint64_t ret;
   __asm__ volatile (
@@ -67,6 +68,7 @@ static __inline int WhereAmI(void){
   if ((int)ret <0) ret = 0;
   return (int)ret;
 }
+#endif
 
 static __inline void blas_lock(volatile BLASULONG *address){
 

kernel/power/KERNEL.POWER10

@@ -17,11 +17,15 @@ SBGEMMOTCOPYOBJ =  sbgemm_otcopy$(TSUFFIX).$(SUFFIX)
 STRMMKERNEL	= sgemm_kernel_power10.c
 DTRMMKERNEL	= dgemm_kernel_power10.c
 ifeq ($(OSNAME), AIX)
-CTRMMKERNEL     = ctrmm_kernel_8x4_power8.S
-ZTRMMKERNEL     = ztrmm_kernel_8x2_power8.S
+#CTRMMKERNEL     = ctrmm_kernel_8x4_power8.S
+#ZTRMMKERNEL     = ztrmm_kernel_8x2_power8.S
+CTRMMKERNEL	= cgemm_kernel_power10.c
+ZTRMMKERNEL	= zgemm_kernel_power10.c
 else
-CTRMMKERNEL	= cgemm_kernel_power10.S
-ZTRMMKERNEL	= zgemm_kernel_power10.S
+#CTRMMKERNEL	= cgemm_kernel_power10.S
+#ZTRMMKERNEL	= zgemm_kernel_power10.S
+CTRMMKERNEL	= cgemm_kernel_power10.c
+ZTRMMKERNEL	= zgemm_kernel_power10.c
 endif
 
 SGEMMKERNEL    =  sgemm_kernel_power10.c
@@ -65,9 +69,11 @@ DGEMM_SMALL_K_TN = dgemm_small_kernel_tn_power10.c
 DGEMM_SMALL_K_B0_TN = dgemm_small_kernel_tn_power10.c
 
 ifeq ($(OSNAME), AIX)
-CGEMMKERNEL    = cgemm_kernel_8x4_power8.S
+#CGEMMKERNEL    = cgemm_kernel_8x4_power8.S
+CGEMMKERNEL    = cgemm_kernel_power10.c
 else
-CGEMMKERNEL    = cgemm_kernel_power10.S
+#CGEMMKERNEL    = cgemm_kernel_power10.S
+CGEMMKERNEL    = cgemm_kernel_power10.c
 endif
 #CGEMMKERNEL     = cgemm_kernel_8x4_power8.S
 CGEMMINCOPY    = ../generic/zgemm_ncopy_8.c
@@ -84,9 +90,11 @@ CGEMMINCOPYOBJ =  cgemm_incopy$(TSUFFIX).$(SUFFIX)
 CGEMMITCOPYOBJ =  cgemm_itcopy$(TSUFFIX).$(SUFFIX)
 
 ifeq ($(OSNAME), AIX)
-ZGEMMKERNEL    = zgemm_kernel_8x2_power8.S
+#ZGEMMKERNEL    = zgemm_kernel_8x2_power8.S
+ZGEMMKERNEL    = zgemm_kernel_power10.c
 else
-ZGEMMKERNEL    = zgemm_kernel_power10.S
+#ZGEMMKERNEL    = zgemm_kernel_power10.S
+ZGEMMKERNEL    = zgemm_kernel_power10.c
 endif
 ZGEMMONCOPY    = ../generic/zgemm_ncopy_2.c
 ZGEMMOTCOPY    = ../generic/zgemm_tcopy_2.c