Merge pull request #1702 from xianyi/develop

Merge develop for 0.3.2
Merge branch 'release-0.3.0' into develop
2018-07-30 07:25:01 +02:00 · 2018-07-29 22:37:09 +02:00 · 2018-07-25 19:55:29 +02:00 · 2018-07-24 17:46:33 +02:00 · 2018-07-22 16:34:09 +02:00 · 2018-07-22 09:19:19 +02:00
11 changed files with 234 additions and 22 deletions
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -6,7 +6,7 @@ cmake_minimum_required(VERSION 2.8.5)
 project(OpenBLAS C ASM)
 set(OpenBLAS_MAJOR_VERSION 0)
 set(OpenBLAS_MINOR_VERSION 3)
-set(OpenBLAS_PATCH_VERSION 2.dev)
+set(OpenBLAS_PATCH_VERSION 2)
 set(OpenBLAS_VERSION "${OpenBLAS_MAJOR_VERSION}.${OpenBLAS_MINOR_VERSION}.${OpenBLAS_PATCH_VERSION}")

 # Adhere to GNU filesystem layout conventions
--- a/Makefile.rule
+++ b/Makefile.rule
@@ -3,7 +3,7 @@
 #

 # This library's version
-VERSION = 0.3.2.dev
+VERSION = 0.3.2

 # If you set the suffix, the library name will be libopenblas_$(LIBNAMESUFFIX).a
 # and libopenblas_$(LIBNAMESUFFIX).so. Meanwhile, the soname in shared library
--- a/cmake/prebuild.cmake
+++ b/cmake/prebuild.cmake
@@ -85,7 +85,7 @@ if (NOT NOFORTRAN)
 endif ()

 # Cannot run getarch on target if we are cross-compiling
-if (DEFINED CORE AND CMAKE_CROSSCOMPILING)
+if (DEFINED CORE AND CMAKE_CROSSCOMPILING AND NOT (${HOST_OS} STREQUAL "WINDOWSSTORE"))
  # Write to config as getarch would

  # TODO: Set up defines that getarch sets up based on every other target
--- a/cpuid_power.c
+++ b/cpuid_power.c
@@ -142,6 +142,52 @@ int detect(void){

  return  CPUTYPE_PPC970;
 #endif
+
+#if defined(__FreeBSD__) || defined(__OpenBSD__) || defined(__DragonFly__)
+int id;
+id = __asm __volatile("mfpvr %0" : "=r"(id));
+switch ( id >> 16 ) {
+  case 0x4e: // POWER9
+    return  return CPUTYPE_POWER8;
+    break;
+  case 0x4d:
+  case 0x4b: // POWER8/8E 
+    return CPUTYPE_POWER8;
+    break;
+  case 0x4a:
+  case 0x3f:  // POWER7/7E
+    return CPUTYPE_POWER6; 
+    break;
+  case 0x3e:
+    return CPUTYPE_POWER6;
+    break;
+  case 0x3a:
+    return CPUTYPE_POWER5;
+    break;
+  case 0x35:
+  case 0x38: // POWER4 /4+ 
+    return CPUTYPE_POWER4;
+    break;
+  case 0x40:
+  case 0x41: // POWER3 /3+ 
+    return CPUTYPE_POWER3;
+    break;
+  case 0x39:
+  case 0x3c:
+  case 0x44:
+  case 0x45:
+    return CPUTYPE_PPC970;
+    break;
+  case 0x70: 
+    return CPUTYPE_CELL;
+    break;
+  case 0x8003: 
+    return CPUTYPE_PPCG4;
+    break;
+  default:  
+    return  CPUTYPE_UNKNOWN;
+  }
+#endif
 }

 void get_architecture(void){
--- a/cpuid_x86.c
+++ b/cpuid_x86.c
@@ -1452,6 +1452,8 @@ int get_cpuname(void){
 	switch (model) {
 	case 1:
 	  // AMD Ryzen
+	case 8:
+	  // AMD Ryzen2
 	  if(support_avx())
 #ifndef NO_AVX2
 	    return CPUTYPE_ZEN;
--- a/driver/others/dynamic.c
+++ b/driver/others/dynamic.c
@@ -607,7 +607,7 @@ static gotoblas_t *get_coretype(void){
 	  }
 	}
      } else if (exfamily == 8) {
-	if (model == 1) {
+	if (model == 1 || model == 8) {
 	  if(support_avx())
 	    return &gotoblas_ZEN;
 	  else{
--- a/driver/others/memory.c
+++ b/driver/others/memory.c
@@ -140,7 +140,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 #endif

 #ifndef BUFFERS_PER_THREAD
-#ifdef USE_OPENMP
+#ifdef USE_OPENMP_UNUSED
 #define BUFFERS_PER_THREAD (MAX_CPU_NUMBER * 2 * MAX_PARALLEL_NUMBER)
 #else
 #define BUFFERS_PER_THREAD NUM_BUFFERS
@@ -363,7 +363,7 @@ int blas_get_cpu_number(void){
 #endif

  // blas_goto_num = 0;
-#ifndef USE_OPENMP
+#ifndef USE_OPENMP_UNUSED
  blas_goto_num=openblas_num_threads_env();
  if (blas_goto_num < 0) blas_goto_num = 0;

@@ -494,10 +494,10 @@ static const int allocation_block_size = BUFFER_SIZE + sizeof(struct alloc_t);
 #endif

 /* Holds pointers to allocated memory */
-#if defined(SMP) && !defined(USE_OPENMP)
+#if defined(SMP) && !defined(USE_OPENMP_UNUSED)
 /* This is the number of threads than can be spawned by the server, which is the
   server plus the number of threads in the thread pool */
-#  define MAX_ALLOCATING_THREADS MAX_CPU_NUMBER * 2 * MAX_PARALLEL_NUMBER +1
+#  define MAX_ALLOCATING_THREADS MAX_CPU_NUMBER * 2 * MAX_PARALLEL_NUMBER * 2
 static int next_memory_table_pos = 0;
 #  if defined(HAS_COMPILER_TLS)
 /* Use compiler generated thread-local-storage */
@@ -532,7 +532,7 @@ static BLASULONG  alloc_lock = 0UL;

 /* Returns a pointer to the start of the per-thread memory allocation data */
 static __inline struct alloc_t ** get_memory_table() {
-#if defined(SMP) && !defined(USE_OPENMP)
+#if defined(SMP) && !defined(USE_OPENMP_UNUSED)
 #  if !defined(HAS_COMPILER_TLS)
 #    if defined(OS_WINDOWS)
  int local_memory_table_pos = (int)::TlsGetValue(local_storage_key);
@@ -1057,7 +1057,7 @@ static volatile int memory_initialized = 0;
 /*                2 : Thread                 */

 static void blas_memory_init(){
-#if defined(SMP) && !defined(USE_OPENMP)
+#if defined(SMP) && !defined(USE_OPENMP_UNUSED)
  next_memory_table_pos = 0;
 #  if !defined(HAS_COMPILER_TLS)
 #    if defined(OS_WINDOWS)
@@ -1279,7 +1279,7 @@ void blas_shutdown(void){
      struct alloc_t *alloc_info = local_memory_table[thread][pos];
      if (alloc_info) {
        alloc_info->release_func(alloc_info);
-        alloc_info = (void *)0;
+        local_memory_table[thread][pos] = (void *)0;
      }
    }
  }
--- a/driver/others/openblas_get_config.c
+++ b/driver/others/openblas_get_config.c
@@ -35,6 +35,12 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

 #include <string.h>

+#if defined(_WIN32) && defined(_MSC_VER)
+#if _MSC_VER < 1900
+#define snprintf _snprintf
+#endif
+#endif
+
 static char* openblas_config_str=""
 #ifdef USE64BITINT
  "USE64BITINT "
--- a/kernel/mips64/KERNEL
+++ b/kernel/mips64/KERNEL
@@ -1,3 +1,12 @@
+CAXPYKERNEL = ../mips/zaxpy.c
+ZAXPYKERNEL = ../mips/zaxpy.c
+SROTKERNEL = ../mips/rot.c
+DROTKERNEL = ../mips/rot.c
+CROTKERNEL = ../mips/zrot.c
+ZROTKERNEL = ../mips/zrot.c
+CSWAPKERNEL = ../mips/zswap.c
+ZSWAPKERNEL = ../mips/zswap.c
+
 ifndef SNRM2KERNEL
 SNRM2KERNEL = snrm2.S
 endif
--- a/kernel/mips64/dot.S
+++ b/kernel/mips64/dot.S
@@ -103,35 +103,83 @@
 	.align 3

 .L12:
+#ifdef DSDOT
+cvt.d.s  a1, a1
+cvt.d.s  b1, b1  
+madd.d s1, s1, a1, b1
+#else
 	MADD	s1, s1, a1, b1
+#endif
 	LD	a1,  4 * SIZE(X)
 	LD	b1,  4 * SIZE(Y)

+#ifdef DSDOT
+cvt.d.s  a2, a2
+cvt.d.s  b2, b2
+madd.d s2, s2, a2, b2
+#else
 	MADD	s2, s2, a2, b2
+#endif
 	LD	a2,  5 * SIZE(X)
 	LD	b2,  5 * SIZE(Y)

+#ifdef DSDOT
+cvt.d.s  a3, a3
+cvt.d.s  b3, b3
+madd.d s1, s1, a3, b3
+#else
 	MADD	s1, s1, a3, b3
+#endif
 	LD	a3,  6 * SIZE(X)
 	LD	b3,  6 * SIZE(Y)

+#ifdef DSDOT
+cvt.d.s  a4, a4
+cvt.d.s  b4, b4
+madd.d s2, s2, a4, b4
+#else
 	MADD	s2, s2, a4, b4
+#endif
 	LD	a4,  7 * SIZE(X)
 	LD	b4,  7 * SIZE(Y)

+#ifdef DSDOT
+cvt.d.s  a1, a1
+cvt.d.s  b1, b1
+madd.d s1, s1, a1, b1
+#else
 	MADD	s1, s1, a1, b1
+#endif
 	LD	a1,  8 * SIZE(X)
 	LD	b1,  8 * SIZE(Y)

+#ifdef DSDOT
+cvt.d.s  a2, a2
+cvt.d.s  b2, b2
+madd.d s2, s2, a2, b2
+#else
 	MADD	s2, s2, a2, b2
+#endif
 	LD	a2,  9 * SIZE(X)
 	LD	b2,  9 * SIZE(Y)

+#ifdef DSDOT
+cvt.d.s  a3, a3
+cvt.d.s  b3, b3
+madd.d s1, s1, a3, b3
+#else
 	MADD	s1, s1, a3, b3
+#endif
 	LD	a3, 10 * SIZE(X)
 	LD	b3, 10 * SIZE(Y)

+#ifdef DSDOT
+cvt.d.s  a4, a4
+cvt.d.s  b4, b4
+madd.d s2, s2, a4, b4
+#else
 	MADD	s2, s2, a4, b4
+#endif
 	LD	a4, 11 * SIZE(X)
 	LD	b4, 11 * SIZE(Y)

@@ -143,29 +191,77 @@
 	.align 3

 .L13:
+#ifdef DSDOT
+cvt.d.s  a1, a1
+cvt.d.s  b1, b1
+madd.d s1, s1, a1, b1
+#else
 	MADD	s1, s1, a1, b1
+#endif
 	LD	a1,  4 * SIZE(X)
 	LD	b1,  4 * SIZE(Y)

+#ifdef DSDOT
+cvt.d.s a2, a2
+cvt.d.s b2, b2
+madd.d s2, s2, a2, b2
+#else
 	MADD	s2, s2, a2, b2
+#endif
 	LD	a2,  5 * SIZE(X)
 	LD	b2,  5 * SIZE(Y)

+#ifdef DSDOT
+cvt.d.s a3, a3 
+cvt.d.s b3, b3
+madd.d s1, s1, a3, b3
+#else
 	MADD	s1, s1, a3, b3
+#endif
 	LD	a3,  6 * SIZE(X)
 	LD	b3,  6 * SIZE(Y)

+#ifdef DSDOT
+cvt.d.s a4, a4
+cvt.d.s b4, b4
+madd.d s2, s2, a4, b4
+#else
 	MADD	s2, s2, a4, b4
+#endif
 	LD	a4,  7 * SIZE(X)
 	LD	b4,  7 * SIZE(Y)

+#ifdef DSDOT
+cvt.d.s  a1, a1
+cvt.d.s  b1, b1
+madd.d s1, s1, a1, b1
+#else
 	MADD	s1, s1, a1, b1
+#endif
 	daddiu	X, X, 8 * SIZE
+#ifdef DSDOT
+cvt.d.s  a2, a2
+cvt.d.s  b2, b2
+madd.d s2, s2, a2, b2
+#else
 	MADD	s2, s2, a2, b2
+#endif
 	daddiu	Y, Y, 8 * SIZE

+#ifdef DSDOT
+cvt.d.s  a3, a3
+cvt.d.s  b3, b3
+madd.d s1, s1, a3, b3
+#else
 	MADD	s1, s1, a3, b3
+#endif
+#ifdef DSDOT
+cvt.d.s  a4, a4
+cvt.d.s  b4, b4
+madd.d s2, s2, a4, b4
+#else
 	MADD	s2, s2, a4, b4
+#endif
 	.align 3

 .L15:
@@ -179,8 +275,13 @@
 	LD	a1,  0 * SIZE(X)
 	LD	b1,  0 * SIZE(Y)

+#ifdef DSDOT
+cvt.d.s  a1, a1
+cvt.d.s  b1, b1
+madd.d s1, s1, a1, b1
+#else
 	MADD	s1, s1, a1, b1
-
+#endif
 	daddiu	I, I, -1

 	daddiu	X, X, SIZE
@@ -225,50 +326,85 @@
 	LD	b1,  0 * SIZE(Y)
 	dadd	Y, Y, INCY

+#ifdef DSDOT
+cvt.d.s  a1, a1
+cvt.d.s  b1, b1
+madd.d s1, s1, a1, b1
+#else
 	MADD	s1, s1, a1, b1
-
+#endif
 	LD	a1,  0 * SIZE(X)
 	dadd	X, X, INCX
 	LD	b1,  0 * SIZE(Y)
 	dadd	Y, Y, INCY

+#ifdef DSDOT
+cvt.d.s  a1, a1
+cvt.d.s  b1, b1
+madd.d s2, s2, a1, b1
+#else
 	MADD	s2, s2, a1, b1
-
+#endif
 	LD	a1,  0 * SIZE(X)
 	dadd	X, X, INCX
 	LD	b1,  0 * SIZE(Y)
 	dadd	Y, Y, INCY

+#ifdef DSDOT
+cvt.d.s  a1, a1
+cvt.d.s  b1, b1
+madd.d s1, s1, a1, b1
+#else
 	MADD	s1, s1, a1, b1
-
+#endif
 	LD	a1,  0 * SIZE(X)
 	dadd	X, X, INCX
 	LD	b1,  0 * SIZE(Y)
 	dadd	Y, Y, INCY

+#ifdef DSDOT
+cvt.d.s  a1, a1
+cvt.d.s  b1, b1
+madd.d s2, s2, a1, b1
+#else
 	MADD	s2, s2, a1, b1
-
+#endif
 	LD	a1,  0 * SIZE(X)
 	dadd	X, X, INCX
 	LD	b1,  0 * SIZE(Y)
 	dadd	Y, Y, INCY

+#ifdef DSDOT
+cvt.d.s  a1, a1
+cvt.d.s  b1, b1
+madd.d s1, s1, a1, b1
+#else
 	MADD	s1, s1, a1, b1
-
+#endif
 	LD	a1,  0 * SIZE(X)
 	dadd	X, X, INCX
 	LD	b1,  0 * SIZE(Y)
 	dadd	Y, Y, INCY

+#ifdef DSDOT
+cvt.d.s  a1, a1
+cvt.d.s  b1, b1
+madd.d s2, s2, a1, b1
+#else
 	MADD	s2, s2, a1, b1
-
+#endif
 	LD	a1,  0 * SIZE(X)
 	dadd	X, X, INCX
 	LD	b1,  0 * SIZE(Y)
 	dadd	Y, Y, INCY

+#ifdef DSDOT
+cvt.d.s  a1, a1
+cvt.d.s  b1, b1
+madd.d s1, s1, a1, b1
+#else
 	MADD	s1, s1, a1, b1
-
+#endif
 	LD	a1,  0 * SIZE(X)
 	dadd	X, X, INCX
 	LD	b1,  0 * SIZE(Y)
@@ -277,7 +413,13 @@
 	daddiu	I, I, -1

 	bgtz	I, .L23
+#ifdef DSDOT
+cvt.d.s  a1, a1
+cvt.d.s  b1, b1
+madd.d s2, s2, a1, b1
+#else
 	MADD	s2, s2, a1, b1
+#endif
 	.align 3

 .L25:
@@ -296,13 +438,20 @@
 	daddiu	I, I, -1

 	bgtz	I, .L26
+#ifdef DSDOT
+cvt.d.s  a1, a1
+cvt.d.s  b1, b1
+madd.d s1, s1, a1, b1
+#else
 	MADD	s1, s1, a1, b1
+#endif
 	.align 3

 .L999:
-	ADD	s1, s1, s2
 #ifdef DSDOT
-	cvt.d.s s1, s1
+	add.d s1, s1, s2
+#else
+	ADD	s1, s1, s2
 #endif
 	j	$31
 	NOP
--- a/utest/ctest.h
+++ b/utest/ctest.h
@@ -84,7 +84,7 @@ struct ctest {
 #endif

 #if _MSC_VER < 1900
-#define snprintf _snprintf_s
+#define snprintf _snprintf
 #endif

 #ifndef __cplusplus
Author	SHA1	Message	Date
Martin Kroeker	e8a68ef261	Merge pull request #1702 from xianyi/develop Merge develop for 0.3.2	2018-07-30 07:25:01 +02:00
Martin Kroeker	64826a0d7d	Merge branch 'release-0.3.0' into develop	2018-07-29 22:37:09 +02:00
Martin Kroeker	25f2d25cfe	Merge pull request #1697 from martin-frbg/issue1696 Do not treat WIndows UWB builds as cross-compiling	2018-07-25 19:55:29 +02:00
Martin Kroeker	73131fa30a	Do not treat WIndows UWB builds as cross-compiling	2018-07-24 17:46:33 +02:00
Martin Kroeker	66fcdd5be8	Merge pull request #1695 from martin-frbg/issue1692 Unset memory table entry, not just the local pointer to it on shutdown	2018-07-22 16:34:09 +02:00
Martin Kroeker	43ac839c16	Unset memory table entry, not just the temporary pointer to it on shutdown to fix crash with multiple instances of OpenBLAS, #1692	2018-07-22 09:19:19 +02:00
Martin Kroeker	7ba5936ecd	Merge pull request #1688 from martin-frbg/issue1673 Temporarily disable special handling of OPENMP thread memory allocation	2018-07-19 19:03:45 +02:00
Martin Kroeker	b14f44d2ad	Temporarily disable special handling of OPENMP thread memory allocation for issue #1673	2018-07-19 08:57:56 +02:00
Martin Kroeker	e71d70ba87	Merge pull request #1681 from martin-frbg/issue1671 Add cpu identification via mfpvr call for the BSDs	2018-07-16 22:47:05 +02:00
Martin Kroeker	d671870f5f	Merge pull request #1684 from martin-frbg/issue1672 Work around utest failures in the MIPS64 SICORTEX target	2018-07-16 22:46:49 +02:00
Martin Kroeker	4e103c822c	typo fix	2018-07-16 12:56:39 +02:00
Martin Kroeker	d2142760e0	Fix precision problem in DSDOT	2018-07-15 17:11:40 +02:00
Martin Kroeker	2fbfc64da8	Use C kernels for default c/zAXPY, xROT, c/zSWAP	2018-07-15 17:09:55 +02:00
Martin Kroeker	8d5b33b6be	Add cpu identification via mfpvr call for the BSDs fixes #1671	2018-07-12 23:39:00 +02:00
Martin Kroeker	36aea5ce2d	Merge pull request #1680 from martin-frbg/snprint Fix wrong redefinitions of snprintf for older MSVC	2018-07-12 14:05:13 +02:00
Martin Kroeker	1309711e24	Fix declaration of snprintf for older MSVC _snprintf_s takes an additional (size) argument, so is no direct replacement. (Note that this code is currently unused - the two instances of snprintf here are within ifdef blocks that are not compiled for MSVC)	2018-07-12 11:47:52 +02:00
Martin Kroeker	571e9de2ac	Fix definition of snprintf for MSVC MS _snprintf_s takes an additional argument for the size of the buffer, so is not a direct replacement (utest/ctest.h from which I copied was wrong)	2018-07-12 11:42:25 +02:00
Martin Kroeker	448ed15115	Merge pull request #1678 from martin-frbg/issue1677 Define snprintf for older versions of MSVC	2018-07-12 09:21:34 +02:00
Martin Kroeker	045fb5ea2c	Define snprintf for older versions of MSVC for #1677	2018-07-12 07:30:58 +02:00
Martin Kroeker	4dd70d98d7	Merge pull request #1667 from xianyi/revert-1642-develop Revert "Rewrite &= -> = and simplify the initial blocking phase."	2018-07-04 08:27:21 +02:00
Martin Kroeker	504310eeb9	Merge pull request #1665 from martin-frbg/cpuid-ryzen2 Add cpuid for AMD Ryzen 2	2018-07-04 08:19:40 +02:00
Martin Kroeker	ea1f39518f	Merge pull request #1663 from martin-frbg/issue1641 Double MAX_ALLOCATING_THREADS to fix segfaults with Go and Octave	2018-07-04 08:19:11 +02:00
Martin Kroeker	d0ec4325cf	Add cpuid for AMD Ryzen 2	2018-07-03 21:03:24 +02:00
Martin Kroeker	3f73e8b8cf	Add cpuid for AMD Ryzen 2 for #1664	2018-07-03 21:01:35 +02:00
Martin Kroeker	a49203b48c	Double MAX_ALLOCATING_THREADS to fix segfaults with Go and Octave for #1641	2018-07-03 17:35:54 +02:00
Martin Kroeker	c6aec89d10	Merge pull request #1657 from martin-frbg/release-0.3.0 Release 0.3.1	2018-07-01 12:03:07 +02:00
Martin Kroeker	e6d7711199	remove dev suffix from version number	2018-07-01 11:59:47 +02:00
Martin Kroeker	7a914347c5	remove dev suffix from version number	2018-07-01 11:58:57 +02:00
Martin Kroeker	3a8f0a6a1f	Merge pull request #1656 from xianyi/develop Update the 0.3 branch from develop	2018-07-01 11:55:21 +02:00
Martin Kroeker	939452ea9d	Merge pull request #1570 from xianyi/develop Update release-0.3.0 branch to match develop	2018-05-23 15:12:20 +02:00