Merge pull request #3536 from xianyi/develop

Update from develop for release 0.3.20

Author: martin-frbg

CMakeLists.txt

@@ -251,12 +251,14 @@ if (APPLE AND DYNAMIC_ARCH AND BUILD_SHARED_LIBS)
   set (CMAKE_Fortran_USE_RESPONSE_FILE_FOR_OBJECTS 1)
   set (CMAKE_Fortran_CREATE_SHARED_LIBRARY
  "sh -c 'cat ${CMAKE_BINARY_DIR}/CMakeFiles/openblas_shared.dir/objects*.rsp | xargs -n 1024 ar -ru libopenblas.a && exit 0' "
+ "sh -c 'ar -ru libopenblas.a ${CMAKE_BINARY_DIR}/driver/others/CMakeFiles/driver_others.dir/xerbla.c.o && exit 0' "
  "sh -c 'echo \"\" | ${CMAKE_Fortran_COMPILER} -o dummy.o -c -x f95-cpp-input - '"
  "sh -c '${CMAKE_Fortran_COMPILER} -fpic -shared -Wl,-all_load -Wl,-force_load,libopenblas.a -Wl,-noall_load dummy.o -o ${CMAKE_LIBRARY_OUTPUT_DIRECTORY}/libopenblas.${OpenBLAS_MAJOR_VERSION}.${OpenBLAS_MINOR_VERSION}.dylib'"
  "sh -c 'ls -l ${CMAKE_BINARY_DIR}/lib'")
   else ()
   set (CMAKE_C_CREATE_SHARED_LIBRARY
    "sh -c 'cat ${CMAKE_BINARY_DIR}/CMakeFiles/openblas_shared.dir/objects*.rsp | xargs -n 1024 ar -ru libopenblas.a && exit 0' "
+   "sh -c 'ar -ru libopenblas.a ${CMAKE_BINARY_DIR}/driver/others/CMakeFiles/driver_others.dir/xerbla.c.o && exit 0' "
    "sh -c '${CMAKE_C_COMPILER} -fpic -shared -Wl,-all_load -Wl,-force_load,libopenblas.a -Wl,-noall_load -o ${CMAKE_LIBRARY_OUTPUT_DIRECTORY}/libopenblas.${OpenBLAS_MAJOR_VERSION}.${OpenBLAS_MINOR_VERSION}.dylib'")
   endif ()
 endif()

CONTRIBUTORS.md

@@ -201,3 +201,9 @@ In chronological order:
 * Bine Brank <https://github.com/binebrank>
   * [2021-10-27] Add vector-length-agnostic DGEMM kernels for Arm SVE
   * [2021-11-20] Vector-length-agnostic Arm SVE copy routines for DGEMM, DTRMM, DSYMM
+  * [2021-11-12] SVE kernels for SGEMM, STRMM and corresponding SVE copy functions
+  * [2022-01-06] SVE kernels for CGEMM, ZGEMM, CTRMM, ZTRMM and corresponding SVE copy functions
+  * [2022-01-18] SVE kernels and copy functions for TRSM
+
+* Ilya Kurdyukov <https://github.com/ilyakurdyukov>
+  * [2021-02-21] Add basic support for the Elbrus E2000 architecture

Changelog.txt

@@ -1,4 +1,39 @@
 OpenBLAS ChangeLog
+====================================================================
+Version 0.3.20
+ 20-Feb-2022
+
+general:
+ - some code cleanup, with added casts etc.
+ - fixed obtaining the cpu count with OpenMP and OMP_PROC_BIND unset
+ - fixed pivot index calculation by ?LASWP for negative increments other than one
+ - fixed input argument check in LAPACK ? GEQRT2
+ - improved the check for a Fortran compiler in CMAKE builds
+ - disabled building OpenBLAS' optimized versions of LAPACK complex SPMV,SPR,SYMV,SYR with NO_LAPACK=1
+ - fixed building of LAPACK on certain distributed filesystems with parallel gmake
+ - fixed building the shared library on MacOS with classic flang
+
+x86_64:
+ - fixed cross-compilation with CMAKE for CORE2 target
+ - fixed miscompilation of AVX512 code in DYNAMIC_ARCH builds
+ - added support for the "incidental" AVX512 hardware in Alder Lake when enabled in BIOS
+
+E2K:
+ - add new architecture (Russian Elbrus E2000 family)
+
+SPARC:
+ - fix IMIN/IMAX
+
+ARMV8:
+ - added SVE-enabled CGEMM and ZGEMM kernels for ARMV8SVE and A64FX
+ - added support for Neoverse N2 and V1 cpus
+
+MIPS,MIPS64:
+ - fixed autodetection of MSA capability
+
+LOONGARCH64:
+ - added an optimized DGEMM kernel
+
 ====================================================================
 Version 0.3.19
  19-Dec-2021

Makefile.arm64

@@ -78,6 +78,66 @@ endif
 endif
 endif
 
+# Use a72 tunings because Neoverse-V1 is only available
+# in GCC>=9.4
+ifeq ($(CORE), NEOVERSEV1)
+ifeq (1, $(filter 1,$(GCCVERSIONGTEQ7) $(ISCLANG)))
+ifeq ($(GCCVERSIONGTEQ9), 1)
+ifeq (1, $(filter 1,$(GCCMINORVERSIONGTEQ4) $(GCCVERSIONGTEQ10)))
+CCOMMON_OPT += -march=armv8.4-a -mtune=neoverse-v1
+ifneq ($(F_COMPILER), NAG)
+FCOMMON_OPT += -march=armv8.4-a -mtune=neoverse-v1
+endif
+else
+CCOMMON_OPT += -march=armv8.4-a -mtune=native
+ifneq ($(F_COMPILER), NAG)
+FCOMMON_OPT += -march=armv8.4-a -mtune=native
+endif
+endif
+else
+CCOMMON_OPT += -march=armv8.2-a -mtune=cortex-a72
+ifneq ($(F_COMPILER), NAG)
+FCOMMON_OPT += -march=armv8.2-a -mtune=cortex-a72
+endif
+endif
+else
+CCOMMON_OPT += -march=armv8-a -mtune=cortex-a72
+ifneq ($(F_COMPILER), NAG)
+FCOMMON_OPT += -march=armv8-a -mtune=cortex-a72
+endif
+endif
+endif
+
+# Use a72 tunings because Neoverse-N2 is only available
+# in GCC>=9.4
+ifeq ($(CORE), NEOVERSEN2)
+ifeq (1, $(filter 1,$(GCCVERSIONGTEQ7) $(ISCLANG)))
+ifeq ($(GCCVERSIONGTEQ9), 1)
+ifeq (1, $(filter 1,$(GCCMINORVERSIONGTEQ4) $(GCCVERSIONGTEQ10)))
+CCOMMON_OPT += -march=armv8.5-a -mtune=neoverse-n2
+ifneq ($(F_COMPILER), NAG)
+FCOMMON_OPT += -march=armv8.5-a -mtune=neoverse-n2
+endif
+else
+CCOMMON_OPT += -march=armv8.5-a -mtune=native
+ifneq ($(F_COMPILER), NAG)
+FCOMMON_OPT += -march=armv8.5-a -mtune=native
+endif
+endif
+else
+CCOMMON_OPT += -march=armv8.2-a -mtune=cortex-a72
+ifneq ($(F_COMPILER), NAG)
+FCOMMON_OPT += -march=armv8.2-a -mtune=cortex-a72
+endif
+endif
+else
+CCOMMON_OPT += -march=armv8-a -mtune=cortex-a72
+ifneq ($(F_COMPILER), NAG)
+FCOMMON_OPT += -march=armv8-a -mtune=cortex-a72
+endif
+endif
+endif
+
 # Use a53 tunings because a55 is only available in GCC>=8.1
 ifeq ($(CORE), CORTEXA55)
 ifeq (1, $(filter 1,$(GCCVERSIONGTEQ7) $(ISCLANG)))

Makefile.e2k

@@ -0,0 +1 @@
+COPT	= -Wall -O2 # -DGEMMTEST

Makefile.prebuild

@@ -3,6 +3,10 @@
 export BINARY
 export USE_OPENMP
 
+ifdef DYNAMIC_ARCH
+override HOST_CFLAGS += -DDYNAMIC_ARCH
+endif
+
 ifdef TARGET_CORE
 TARGET_MAKE = Makefile_kernel.conf
 TARGET_CONF = config_kernel.h

Makefile.rule

@@ -3,7 +3,7 @@
 #
 
 # This library's version
-VERSION = 0.3.19
+VERSION = 0.3.19.dev
 
 # If you set the suffix, the library name will be libopenblas_$(LIBNAMESUFFIX).a
 # and libopenblas_$(LIBNAMESUFFIX).so. Meanwhile, the soname in shared library

Makefile.system

@@ -277,7 +277,7 @@ HAVE_GAS := $(shell $(AS) -v < /dev/null 2>&1 | grep GNU 2>&1 >/dev/null ; echo
 GETARCH_FLAGS += -DHAVE_GAS=$(HAVE_GAS)
 
 # Generating Makefile.conf and config.h
-DUMMY := $(shell $(MAKE) -C $(TOPDIR) -f Makefile.prebuild CC="$(CC)" FC="$(FC)" HOSTCC="$(HOSTCC)" HOST_CFLAGS="$(GETARCH_FLAGS)" CFLAGS="$(CFLAGS)" BINARY=$(BINARY) USE_OPENMP=$(USE_OPENMP) TARGET_CORE=$(TARGET_CORE) ONLY_CBLAS=$(ONLY_CBLAS) TARGET=$(TARGET) all)
+DUMMY := $(shell $(MAKE) -C $(TOPDIR) -f Makefile.prebuild CC="$(CC)" FC="$(FC)" HOSTCC="$(HOSTCC)" HOST_CFLAGS="$(GETARCH_FLAGS)" CFLAGS="$(CFLAGS)" BINARY=$(BINARY) USE_OPENMP=$(USE_OPENMP) DYNAMIC_ARCH=$(DYNAMIC_ARCH) TARGET_CORE=$(TARGET_CORE) ONLY_CBLAS=$(ONLY_CBLAS) TARGET=$(TARGET) all)
 
 ifndef TARGET_CORE
 include $(TOPDIR)/Makefile.conf
@@ -374,6 +374,7 @@ else
 endif
 GCCMINORVERSIONGTEQ1 := $(shell expr `$(CC) $(GCCDUMPVERSION_PARAM) | cut -f2 -d.` \>= 1)
 GCCMINORVERSIONGTEQ2 := $(shell expr `$(CC) $(GCCDUMPVERSION_PARAM) | cut -f2 -d.` \>= 2)
+GCCMINORVERSIONGTEQ4 := $(shell expr `$(CC) $(GCCDUMPVERSION_PARAM) | cut -f2 -d.` \>= 4)
 GCCMINORVERSIONGTEQ7 := $(shell expr `$(CC) $(GCCDUMPVERSION_PARAM) | cut -f2 -d.` \>= 7)
 endif
 
@@ -654,6 +655,8 @@ DYNAMIC_CORE += CORTEXA57
 DYNAMIC_CORE += CORTEXA72
 DYNAMIC_CORE += CORTEXA73
 DYNAMIC_CORE += NEOVERSEN1
+DYNAMIC_CORE += NEOVERSEV1
+DYNAMIC_CORE += NEOVERSEN2
 DYNAMIC_CORE += CORTEXA55
 DYNAMIC_CORE += FALKOR
 DYNAMIC_CORE += THUNDERX

TargetList.txt

@@ -93,6 +93,8 @@ CORTEXA57
 CORTEXA72
 CORTEXA73
 NEOVERSEN1
+NEOVERSEV1
+NEOVERSEN2
 CORTEXA55
 EMAG8180
 FALKOR
@@ -113,3 +115,7 @@ C910V
 
 11.LOONGARCH64:
 LOONGSON3R5
+
+12. Elbrus E2000:
+E2K
+

azure-pipelines.yml

@@ -224,7 +224,7 @@ jobs:
 
 - job: OSX_IOS_ARMV8
   pool:
-     vmImage: 'macOS-10.15'
+     vmImage: 'macOS-11'
   variables:
      CC: /Applications/Xcode_12.4.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/clang
      CFLAGS: -O2 -Wno-macro-redefined -isysroot /Applications/Xcode_12.4.app/Contents/Developer/Platforms/iPhoneOS.platform/Developer/SDKs/iPhoneOS14.4.sdk -arch arm64 -miphoneos-version-min=10.0