2 ====================================================================
6 * Fixed LAPACK unstable bug about ?laswp. (#130)
7 * Fixed the shared library bug about unloading the library on
9 * Fixed the compilation failure on BlueGene/P (TARGET=PPC440FP2)
10 Please use gcc and IBM xlf. (#134)
12 * Supported goto_set_num_threads and openblas_set_num_threads
13 APIs in Windows. They can set the number of threads on runtime.
15 ====================================================================
19 * Fixed exporting DLL functions bug on Windows/MingW
20 * Support GNU Hurd (Thank Sylvestre Ledru)
21 * Support kfreebsd kernel (Thank Sylvestre Ledru)
23 * Support Intel Sandy Bridge 22nm desktop/mobile CPU
25 * Improve the detection of SPARC (Thank Sylvestre Ledru)
27 ====================================================================
32 * Fixed the SEGFAULT bug about hyper-theading
33 * Support AMD Bulldozer by using GotoBLAS2 AMD Barcelona codes
35 ====================================================================
39 * Removed the limitation (64) of numbers of CPU cores.
40 Now, it supports 256 cores at max.
41 * Supported clang compiler.
42 * Fixed some build bugs on FreeBSD
44 * Optimized Level-3 BLAS on Intel Sandy Bridge x86-64 by AVX instructions.
45 Please use gcc >= 4.6 or clang >=3.1.
46 * Support AMD Bobcat by using GotoBLAS2 AMD Barcelona codes.
48 ====================================================================
52 * Upgraded LAPACK to 3.4.1 version. (Thank Zaheer Chothia)
53 * Supported LAPACKE, a C interface to LAPACKE. (Thank Zaheer Chothia)
54 * Fixed the build bug (MD5 and download) on Mac OSX.
55 * Auto download CUnit 2.1.2-2 from SF.net with UTEST_CHECK=1.
56 * Fxied the compatibility issue for compilers without C99 complex number
59 * Auto-detect Intel Sandy Bridge Core i7-3xxx & Xeon E7 Westmere-EX.
60 * Test alpha=Nan in dscale.
61 * Fixed a SEGFAULT bug in samax on x86 windows.
63 ====================================================================
67 * Set soname of shared library on Linux.
68 * Added LIBNAMESUFFIX flag in Makefile.rule. The user can use
69 this flag to control the library name, e.g. libopenblas.a,
70 libopenblas_ifort.a or libopenblas_omp.a.
71 * Added GEMM_MULTITHREAD_THRESHOLD flag in Makefile.rule.
72 The lib use single thread in GEMM function with small matrices.
74 * Used GEMV SSE/SSE2 kernels on x86 32-bit.
75 * Exported CBLAS functions in Windows DLL.
77 * Completed Level-3 BLAS optimization on Loongson 3A CPU.
78 * Improved GEMV performance on Loongson 3A CPU.
79 * Improved Level-3 BLAS performance on Loongson 3B CPU. (EXPERIMENT)
81 ====================================================================
85 * Fixed missing "#include <sched.h>" bug on Mac OS X.
86 Thank Mike Nolta for the patch.
87 * Upgraded LAPACK to 3.4.0 version
88 * Fixed a bug on Mac OS X. Don't require SystemStubs on OS X.
89 SystemStubs does not exist on Lion. Thank Stefan Karpinski.
90 * Improved README with using OpenMP. Check the internal threads
91 count less than or equal to omp_get_max_threads()
93 * Auto-detect Intel Core i6/i7 (Sandy Bridge) CPU with Nehalem assembly kernels
94 * Fixed some bugs on MingW 64-bit including zgemv, cdot, zdot.
96 ====================================================================
100 * Fixed a bug about installation. The header file "fblas77.h"
102 * Fixed #61 a building bug about setting TARGET and DYNAMIC_ARCH.
103 * Try to handle absolute path of shared library in OSX. (#57)
104 Thank Dr Kane O'Donnell.
105 * Changed the installation folder layout to $(PREFIX)/include and
109 * Fixed #58 zdot/xdot SEGFAULT bug with GCC-4.6 on x86. According
110 to i386 calling convention, The callee should remove the first
111 hidden parameter.Thank Mr. John for this patch.
113 ====================================================================
118 * Added DTB_ENTRIES into dynamic arch setting parameters. Now,
119 it can read DTB_ENTRIES on runtime. (Refs issue #55 on github)
121 ====================================================================
126 * Fixed a building bug when DYNAMIC_ARCH=1 & INTERFACE64=1.
127 (Refs issue #44 on github)
129 ====================================================================
134 * Stop the build and output the error message when detecting
135 fortran compiler failed. (Refs issue #42 on github)
137 ====================================================================
142 * Fixed blasint undefined bug in <cblas.h> file. Other software
143 could include this header successfully(Refs issue #13 on github)
144 * Fixed the SEGFAULT bug on 64 cores. On SMP server, the number
145 of CPUs or cores should be less than or equal to 64.(Refs issue #14
147 * Support "void goto_set_num_threads(int num_threads)" and "void
148 openblas_set_num_threads(int num_threads)" when USE_OPENMP=1
149 * Added extern "C" to support C++. Thank Tasio for the patch(Refs
151 * Provided an error message when the arch is not supported.(Refs
153 * Fixed issue #23. Fixed a bug of f_check script about generating link flags.
154 * Added openblas_set_num_threads for Fortran.
155 * Fixed #25 a wrong result of rotmg.
156 * Fixed a bug about detecting underscore prefix in c_check.
157 * Print the wall time (cycles) with enabling FUNCTION_PROFILE
158 * Fixed #35 a build bug with NO_LAPACK=1 & DYNAMIC_ARCH=1
159 * Added install target. You can use "make install". (Refs #20)
163 * Fixed #28 a wrong result of dsdot on x86_64.
164 * Fixed #32 a SEGFAULT bug of zdotc with gcc-4.6.
165 * Fixed #33 ztrmm bug on Nehalem.
166 * Walk round #27 the low performance axpy issue with small imput size & multithreads.
169 * Fixed #28 a wrong result of dsdot on Loongson3A/MIPS64.
170 * Optimized single/double precision BLAS Level3 on Loongson3A/MIPS64. (Refs #2)
171 * Optimized single/double precision axpy function on Loongson3A/MIPS64. (Refs #3)
173 ====================================================================
178 * Support "make NO_LAPACK=1" to build the library without
180 * Fixed randomly SEGFAULT when nodemask==NULL with above Linux 2.6.34.
181 Thank Mr.Ei-ji Nakama providing this patch. (Refs issue #12 on github)
182 * Added DEBUG=1 rule in Makefile.rule to build debug version.
183 * Disable compiling quad precision in reference BLAS library(netlib BLAS).
184 * Added unit testcases in utest/ subdir. Used CUnit framework.
185 * Supported OPENBLAS_* & GOTO_* environment variables (Pleas see README)
186 * Imported GotoBLAS2 1.13 BSD version
189 * On x86 32bits, fixed a bug in zdot_sse2.S line 191. This would casue
190 zdotu & zdotc failures.Instead,Walk around it. (Refs issue #8 #9 on github)
191 * Modified ?axpy functions to return same netlib BLAS results
192 when incx==0 or incy==0 (Refs issue #7 on github)
193 * Modified ?swap functions to return same netlib BLAS results
194 when incx==0 or incy==0 (Refs issue #6 on github)
195 * Modified ?rot functions to return same netlib BLAS results
196 when incx==0 or incy==0 (Refs issue #4 on github)
197 * Detect Intel Westmere,Intel Clarkdale and Intel Arrandale
198 to use Nehalem codes.
199 * Fixed a typo bug about compiling dynamic ARCH library.
201 * Improve daxpy performance on ICT Loongson 3A.
202 * Supported ICT Loongson 3A CPU (Refs issue #1 on github)
203 ====================================================================