Merge branch 'release-v0.1alpha2' into develop

Changelog.txt

@@ -1,7 +1,8 @@
 OpenBLAS ChangeLog
 ====================================================================
-Version 0.1 alpha2(in development)
+Version 0.1 alpha2
-0;136;0c
+23-Jun-2011
+
 common:
 	* Fixed blasint undefined bug in <cblas.h> file. Other software 
 	  could include this header successfully(Refs issue #13 on github)
@@ -31,6 +32,8 @@ x86/x86_64:
 
 MIPS64:
 	* Fixed #28 a wrong result of dsdot on Loongson3A/MIPS64. 
+	* Optimized single/double precision BLAS Level3 on Loongson3A/MIPS64. (Refs #2)
+	* Optimized single/double precision axpy function on Loongson3A/MIPS64. (Refs #3)
 
 ====================================================================
 Version 0.1 alpha1

Makefile

@@ -74,7 +74,7 @@ ifeq ($(OSNAME), Darwin)
 endif
 ifeq ($(OSNAME), WINNT)
 	$(MAKE) -C exports dll
-#	-ln -fs $(LIBDLLNAME) libopenblas.dll
+	-ln -fs $(LIBDLLNAME) libopenblas.dll
 endif
 ifeq ($(OSNAME), CYGWIN_NT)
 	$(MAKE) -C exports dll

README

@@ -72,6 +72,7 @@ Please see Changelog.txt to obtain the differences between GotoBLAS2 1.13 BSD ve
 9.Known Issues
 * The number of CPUs/Cores should less than or equal to 8*sizeof(unsigned long). On 64 bits, the limit 
   is 64. On 32 bits, it is 32.
+* This library is not compatible with EKOPath Compiler Suite 4.0.10 (http://www.pathscale.com/ekopath-compiler-suite). However, Path64 (https://github.com/path64/compiler) could compile the codes successfully.
 
 10. Specification of Git Branches
 We used the git branching model in this article (http://nvie.com/posts/a-successful-git-branching-model/). 
@@ -79,4 +80,4 @@ Now, there are 4 branches in github.com.
   * The master branch. This a main branch to reflect a production-ready state.
   * The develop branch. This a main branch to reflect a state with the latest delivered development changes for the next release.
   * The loongson3a branch. This is a feature branch. We develop Loongson3A codes on this branch. We will merge this feature to develop branch in future.
-  * The gh-pages branch. This is for web pages
+  * The gh-pages branch. This is for web pages

common_mips64.h

@@ -220,6 +220,11 @@ REALNAME: ;\
 
 #define BUFFER_SIZE     ( 8 << 20)
 
+#if defined(LOONGSON3A)
+#define PAGESIZE	(16UL << 10)
+#define FIXED_PAGESIZE	(16UL << 10)
+#endif
+
 #ifndef PAGESIZE
 #define PAGESIZE	(64UL << 10)
 #endif

driver/others/blas_server_omp.c

@@ -38,7 +38,7 @@
 
 #include <stdio.h>
 #include <stdlib.h>
-#include <sys/mman.h>
+//#include <sys/mman.h>
 #include "common.h"
 
 #ifndef USE_OPENMP

exports/Makefile

@@ -53,18 +53,19 @@ dyn : $(LIBDYNNAME)
 zip : dll
 	zip $(LIBZIPNAME) $(LIBDLLNAME) $(LIBNAME)
 
-dll  : libgoto2.dll
+dll  : ../$(LIBDLLNAME)
+#libgoto2.dll
 
 dll2 : libgoto2_shared.dll
 
-libgoto2.dll : ../$(LIBNAME) libgoto2.def dllinit.$(SUFFIX)
+../$(LIBDLLNAME) : ../$(LIBNAME) libgoto2.def dllinit.$(SUFFIX)
 	$(RANLIB) ../$(LIBNAME)
 ifeq ($(BINARY32), 1)
-	$(DLLWRAP) -o $(@F) --def libgoto2.def \
+	$(DLLWRAP) -o ../$(LIBDLLNAME) --def libgoto2.def \
 	--entry _dllinit@12 -s dllinit.$(SUFFIX) --dllname $(@F) ../$(LIBNAME) $(FEXTRALIB)
 	-lib /machine:i386 /def:libgoto2.def
 else
-	$(DLLWRAP) -o $(@F) --def libgoto2.def \
+	$(DLLWRAP) -o ../$(LIBDLLNAME) --def libgoto2.def \
 	--entry _dllinit    -s dllinit.$(SUFFIX) --dllname $(@F) ../$(LIBNAME) $(FEXTRALIB)
 	-lib /machine:X64 /def:libgoto2.def 
 endif

kernel/mips64/KERNEL

@@ -91,15 +91,37 @@ ifndef ZGEMM_BETA
 ZGEMM_BETA = ../generic/zgemm_beta.c
 endif
 
+ifndef	STRSMKERNEL_LN
 STRSMKERNEL_LN	=  trsm_kernel_LN.S
-STRSMKERNEL_LT	=  trsm_kernel_LT.S
+endif
-STRSMKERNEL_RN	=  trsm_kernel_LT.S
-STRSMKERNEL_RT	=  trsm_kernel_RT.S
 
+ifndef	STRSMKERNEL_LT
+STRSMKERNEL_LT	=  trsm_kernel_LT.S
+endif
+
+ifndef	STRSMKERNEL_RN
+STRSMKERNEL_RN	=  trsm_kernel_LT.S
+endif
+
+ifndef	STRSMKERNEL_RT
+STRSMKERNEL_RT	=  trsm_kernel_RT.S
+endif
+
+ifndef DTRSMKERNEL_LN
 DTRSMKERNEL_LN	=  trsm_kernel_LN.S
+endif
+
+ifndef DTRSMKERNEL_LT
 DTRSMKERNEL_LT	=  trsm_kernel_LT.S
+endif
+
+ifndef DTRSMKERNEL_RN
 DTRSMKERNEL_RN	=  trsm_kernel_LT.S
+endif
+
+ifndef DTRSMKERNEL_RT
 DTRSMKERNEL_RT	=  trsm_kernel_RT.S
+endif
 
 CTRSMKERNEL_LN	=  ztrsm_kernel_LT.S
 CTRSMKERNEL_LT	=  ztrsm_kernel_LT.S

kernel/mips64/KERNEL.LOONGSON3A

@@ -1,2 +1,24 @@
 SAXPYKERNEL=axpy_loongson3a.S
 DAXPYKERNEL=daxpy_loongson3a_simd.S
+
+SGEMMKERNEL    =  sgemm_kernel_loongson3a.S
+SGEMMONCOPY    = ../generic/gemm_ncopy_4.c
+SGEMMOTCOPY    = ../generic/gemm_tcopy_4.c
+SGEMMONCOPYOBJ =  sgemm_oncopy.o
+SGEMMOTCOPYOBJ =  sgemm_otcopy.o
+
+DGEMMKERNEL    =  gemm_kernel_loongson3a.S
+DGEMMONCOPY    = ../generic/gemm_ncopy_4.c
+DGEMMOTCOPY    = ../generic/gemm_tcopy_4.c
+DGEMMONCOPYOBJ = dgemm_oncopy.o
+DGEMMOTCOPYOBJ = dgemm_otcopy.o
+
+STRSMKERNEL_LN	=  ../generic/trsm_kernel_LN.c
+STRSMKERNEL_LT	=  ../generic/trsm_kernel_LT.c
+STRSMKERNEL_RN	=  ../generic/trsm_kernel_RN.c
+STRSMKERNEL_RT	=  ../generic/trsm_kernel_RT.c
+
+DTRSMKERNEL_LN	= ../generic/trsm_kernel_LN.c
+DTRSMKERNEL_LT	= ../generic/trsm_kernel_LT.c
+DTRSMKERNEL_RN	= ../generic/trsm_kernel_RN.c
+DTRSMKERNEL_RT	= ../generic/trsm_kernel_RT.c

param.h

@@ -1480,27 +1480,29 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 #define GEMM_DEFAULT_OFFSET_B 0
 #define GEMM_DEFAULT_ALIGN 0x03fffUL
 
-#define SGEMM_DEFAULT_UNROLL_M  2
+#define SGEMM_DEFAULT_UNROLL_M  4
-#define SGEMM_DEFAULT_UNROLL_N  8
+#define SGEMM_DEFAULT_UNROLL_N  4
-#define DGEMM_DEFAULT_UNROLL_M  2
+
-#define DGEMM_DEFAULT_UNROLL_N  8
+#define DGEMM_DEFAULT_UNROLL_M  4
+#define DGEMM_DEFAULT_UNROLL_N  4
+
 #define CGEMM_DEFAULT_UNROLL_M  1
 #define CGEMM_DEFAULT_UNROLL_N  4
 #define ZGEMM_DEFAULT_UNROLL_M  1
 #define ZGEMM_DEFAULT_UNROLL_N  4
 
-#define SGEMM_DEFAULT_P 108
+#define SGEMM_DEFAULT_P 32
-#define DGEMM_DEFAULT_P 112
+#define DGEMM_DEFAULT_P	32 
 #define CGEMM_DEFAULT_P 108
 #define ZGEMM_DEFAULT_P 112
 
-#define SGEMM_DEFAULT_Q 288
+#define SGEMM_DEFAULT_Q 116
-#define DGEMM_DEFAULT_Q 144
+#define DGEMM_DEFAULT_Q 116
 #define CGEMM_DEFAULT_Q 144
 #define ZGEMM_DEFAULT_Q  72
 
-#define SGEMM_DEFAULT_R 2000
+#define SGEMM_DEFAULT_R 1000
-#define DGEMM_DEFAULT_R 2000
+#define DGEMM_DEFAULT_R 1000
 #define CGEMM_DEFAULT_R 2000
 #define ZGEMM_DEFAULT_R 2000