2 ====================================================================
6 * Improved OpenMP performance slightly. (d744c9)
7 * Improved cblas.h compatibility with Intel MKL.(#185)
8 * Fixed the overflowing bug in single thread cholesky factorization.
9 * Fixed the overflowing buffer bug of multithreading hbmv and sbmv.(#174)
12 * Added AMD Bulldozer x86-64 S/DGEMM AVX kernels. (Thank Werner Saar)
13 We will tune the performance in future.
14 * Auto-detect Intel Xeon E7540.
15 * Fixed the overflowing buffer bug of gemv. (#173)
16 * Fixed the bug of s/cdot about invalid reading NAN on x86_64. (#189)
20 ====================================================================
24 * Added NO_SHARED flag to disable generating the shared library.
25 * Compile LAPACKE with ILP64 modle when INTERFACE64=1 (#158)
26 * Export LAPACK 3.4.2 symbols in shared library. (#147)
27 * Only detect the number of physical CPU cores on Mac OSX. (#157)
28 * Fixed NetBSD build. (#155)
29 * Fixed compilation with TARGET=GENERIC. (#160)
31 * Restore the original CPU affinity when calling
32 openblas_set_num_threads(1) (#153)
33 * Fixed a SEGFAULT bug in dgemv_t when m is very large.(#154)
36 ====================================================================
40 * Upgraded LAPACK to 3.4.2 version. (#145)
41 * Provided support for passing CFLAGS, FFLAGS, PFLAGS,
42 FPFLAGS to make. (#137)
43 * f77blas.h:compatibility for compilers without C99 complex
44 number support. (#141)
46 * Added NO_AVX flag. Check OS supporting AVX on runtime. (#139)
47 * Fixed zdot incompatibility ABI issue with GCC 4.7 on
48 Windows 32-bit. (#140)
50 * Fixed the generation of shared library bug.
51 * Fixed the detection bug on the Loongson 3A server.
52 ====================================================================
56 * Fixed LAPACK unstable bug about ?laswp. (#130)
57 * Fixed the shared library bug about unloading the library on
59 * Fixed the compilation failure on BlueGene/P (TARGET=PPC440FP2)
60 Please use gcc and IBM xlf. (#134)
62 * Supported goto_set_num_threads and openblas_set_num_threads
63 APIs in Windows. They can set the number of threads on runtime.
65 ====================================================================
69 * Fixed exporting DLL functions bug on Windows/MingW
70 * Support GNU Hurd (Thank Sylvestre Ledru)
71 * Support kfreebsd kernel (Thank Sylvestre Ledru)
73 * Support Intel Sandy Bridge 22nm desktop/mobile CPU
75 * Improve the detection of SPARC (Thank Sylvestre Ledru)
77 ====================================================================
82 * Fixed the SEGFAULT bug about hyper-theading
83 * Support AMD Bulldozer by using GotoBLAS2 AMD Barcelona codes
85 ====================================================================
89 * Removed the limitation (64) of numbers of CPU cores.
90 Now, it supports 256 cores at max.
91 * Supported clang compiler.
92 * Fixed some build bugs on FreeBSD
94 * Optimized Level-3 BLAS on Intel Sandy Bridge x86-64 by AVX instructions.
95 Please use gcc >= 4.6 or clang >=3.1.
96 * Support AMD Bobcat by using GotoBLAS2 AMD Barcelona codes.
98 ====================================================================
102 * Upgraded LAPACK to 3.4.1 version. (Thank Zaheer Chothia)
103 * Supported LAPACKE, a C interface to LAPACKE. (Thank Zaheer Chothia)
104 * Fixed the build bug (MD5 and download) on Mac OSX.
105 * Auto download CUnit 2.1.2-2 from SF.net with UTEST_CHECK=1.
106 * Fxied the compatibility issue for compilers without C99 complex number
109 * Auto-detect Intel Sandy Bridge Core i7-3xxx & Xeon E7 Westmere-EX.
110 * Test alpha=Nan in dscale.
111 * Fixed a SEGFAULT bug in samax on x86 windows.
113 ====================================================================
117 * Set soname of shared library on Linux.
118 * Added LIBNAMESUFFIX flag in Makefile.rule. The user can use
119 this flag to control the library name, e.g. libopenblas.a,
120 libopenblas_ifort.a or libopenblas_omp.a.
121 * Added GEMM_MULTITHREAD_THRESHOLD flag in Makefile.rule.
122 The lib use single thread in GEMM function with small matrices.
124 * Used GEMV SSE/SSE2 kernels on x86 32-bit.
125 * Exported CBLAS functions in Windows DLL.
127 * Completed Level-3 BLAS optimization on Loongson 3A CPU.
128 * Improved GEMV performance on Loongson 3A CPU.
129 * Improved Level-3 BLAS performance on Loongson 3B CPU. (EXPERIMENT)
131 ====================================================================
135 * Fixed missing "#include <sched.h>" bug on Mac OS X.
136 Thank Mike Nolta for the patch.
137 * Upgraded LAPACK to 3.4.0 version
138 * Fixed a bug on Mac OS X. Don't require SystemStubs on OS X.
139 SystemStubs does not exist on Lion. Thank Stefan Karpinski.
140 * Improved README with using OpenMP. Check the internal threads
141 count less than or equal to omp_get_max_threads()
143 * Auto-detect Intel Core i6/i7 (Sandy Bridge) CPU with Nehalem assembly kernels
144 * Fixed some bugs on MingW 64-bit including zgemv, cdot, zdot.
146 ====================================================================
150 * Fixed a bug about installation. The header file "fblas77.h"
152 * Fixed #61 a building bug about setting TARGET and DYNAMIC_ARCH.
153 * Try to handle absolute path of shared library in OSX. (#57)
154 Thank Dr Kane O'Donnell.
155 * Changed the installation folder layout to $(PREFIX)/include and
159 * Fixed #58 zdot/xdot SEGFAULT bug with GCC-4.6 on x86. According
160 to i386 calling convention, The callee should remove the first
161 hidden parameter.Thank Mr. John for this patch.
163 ====================================================================
168 * Added DTB_ENTRIES into dynamic arch setting parameters. Now,
169 it can read DTB_ENTRIES on runtime. (Refs issue #55 on github)
171 ====================================================================
176 * Fixed a building bug when DYNAMIC_ARCH=1 & INTERFACE64=1.
177 (Refs issue #44 on github)
179 ====================================================================
184 * Stop the build and output the error message when detecting
185 fortran compiler failed. (Refs issue #42 on github)
187 ====================================================================
192 * Fixed blasint undefined bug in <cblas.h> file. Other software
193 could include this header successfully(Refs issue #13 on github)
194 * Fixed the SEGFAULT bug on 64 cores. On SMP server, the number
195 of CPUs or cores should be less than or equal to 64.(Refs issue #14
197 * Support "void goto_set_num_threads(int num_threads)" and "void
198 openblas_set_num_threads(int num_threads)" when USE_OPENMP=1
199 * Added extern "C" to support C++. Thank Tasio for the patch(Refs
201 * Provided an error message when the arch is not supported.(Refs
203 * Fixed issue #23. Fixed a bug of f_check script about generating link flags.
204 * Added openblas_set_num_threads for Fortran.
205 * Fixed #25 a wrong result of rotmg.
206 * Fixed a bug about detecting underscore prefix in c_check.
207 * Print the wall time (cycles) with enabling FUNCTION_PROFILE
208 * Fixed #35 a build bug with NO_LAPACK=1 & DYNAMIC_ARCH=1
209 * Added install target. You can use "make install". (Refs #20)
213 * Fixed #28 a wrong result of dsdot on x86_64.
214 * Fixed #32 a SEGFAULT bug of zdotc with gcc-4.6.
215 * Fixed #33 ztrmm bug on Nehalem.
216 * Walk round #27 the low performance axpy issue with small imput size & multithreads.
219 * Fixed #28 a wrong result of dsdot on Loongson3A/MIPS64.
220 * Optimized single/double precision BLAS Level3 on Loongson3A/MIPS64. (Refs #2)
221 * Optimized single/double precision axpy function on Loongson3A/MIPS64. (Refs #3)
223 ====================================================================
228 * Support "make NO_LAPACK=1" to build the library without
230 * Fixed randomly SEGFAULT when nodemask==NULL with above Linux 2.6.34.
231 Thank Mr.Ei-ji Nakama providing this patch. (Refs issue #12 on github)
232 * Added DEBUG=1 rule in Makefile.rule to build debug version.
233 * Disable compiling quad precision in reference BLAS library(netlib BLAS).
234 * Added unit testcases in utest/ subdir. Used CUnit framework.
235 * Supported OPENBLAS_* & GOTO_* environment variables (Pleas see README)
236 * Imported GotoBLAS2 1.13 BSD version
239 * On x86 32bits, fixed a bug in zdot_sse2.S line 191. This would casue
240 zdotu & zdotc failures.Instead,Walk around it. (Refs issue #8 #9 on github)
241 * Modified ?axpy functions to return same netlib BLAS results
242 when incx==0 or incy==0 (Refs issue #7 on github)
243 * Modified ?swap functions to return same netlib BLAS results
244 when incx==0 or incy==0 (Refs issue #6 on github)
245 * Modified ?rot functions to return same netlib BLAS results
246 when incx==0 or incy==0 (Refs issue #4 on github)
247 * Detect Intel Westmere,Intel Clarkdale and Intel Arrandale
248 to use Nehalem codes.
249 * Fixed a typo bug about compiling dynamic ARCH library.
251 * Improve daxpy performance on ICT Loongson 3A.
252 * Supported ICT Loongson 3A CPU (Refs issue #1 on github)
253 ====================================================================