2 ====================================================================
6 * Added NO_SHARED flag to disable generating the shared library.
7 * Compile LAPACKE with ILP64 modle when INTERFACE64=1 (#158)
8 * Export LAPACK 3.4.2 symbols in shared library. (#147)
9 * Only detect the number of physical CPU cores on Mac OSX. (#157)
10 * Fixed NetBSD build. (#155)
11 * Fixed compilation with TARGET=GENERIC. (#160)
13 * Restore the original CPU affinity when calling
14 openblas_set_num_threads(1) (#153)
15 * Fixed a SEGFAULT bug in dgemv_t when m is very large.(#154)
18 ====================================================================
22 * Upgraded LAPACK to 3.4.2 version. (#145)
23 * Provided support for passing CFLAGS, FFLAGS, PFLAGS,
24 FPFLAGS to make. (#137)
25 * f77blas.h:compatibility for compilers without C99 complex
26 number support. (#141)
28 * Added NO_AVX flag. Check OS supporting AVX on runtime. (#139)
29 * Fixed zdot incompatibility ABI issue with GCC 4.7 on
30 Windows 32-bit. (#140)
32 * Fixed the generation of shared library bug.
33 * Fixed the detection bug on the Loongson 3A server.
34 ====================================================================
38 * Fixed LAPACK unstable bug about ?laswp. (#130)
39 * Fixed the shared library bug about unloading the library on
41 * Fixed the compilation failure on BlueGene/P (TARGET=PPC440FP2)
42 Please use gcc and IBM xlf. (#134)
44 * Supported goto_set_num_threads and openblas_set_num_threads
45 APIs in Windows. They can set the number of threads on runtime.
47 ====================================================================
51 * Fixed exporting DLL functions bug on Windows/MingW
52 * Support GNU Hurd (Thank Sylvestre Ledru)
53 * Support kfreebsd kernel (Thank Sylvestre Ledru)
55 * Support Intel Sandy Bridge 22nm desktop/mobile CPU
57 * Improve the detection of SPARC (Thank Sylvestre Ledru)
59 ====================================================================
64 * Fixed the SEGFAULT bug about hyper-theading
65 * Support AMD Bulldozer by using GotoBLAS2 AMD Barcelona codes
67 ====================================================================
71 * Removed the limitation (64) of numbers of CPU cores.
72 Now, it supports 256 cores at max.
73 * Supported clang compiler.
74 * Fixed some build bugs on FreeBSD
76 * Optimized Level-3 BLAS on Intel Sandy Bridge x86-64 by AVX instructions.
77 Please use gcc >= 4.6 or clang >=3.1.
78 * Support AMD Bobcat by using GotoBLAS2 AMD Barcelona codes.
80 ====================================================================
84 * Upgraded LAPACK to 3.4.1 version. (Thank Zaheer Chothia)
85 * Supported LAPACKE, a C interface to LAPACKE. (Thank Zaheer Chothia)
86 * Fixed the build bug (MD5 and download) on Mac OSX.
87 * Auto download CUnit 2.1.2-2 from SF.net with UTEST_CHECK=1.
88 * Fxied the compatibility issue for compilers without C99 complex number
91 * Auto-detect Intel Sandy Bridge Core i7-3xxx & Xeon E7 Westmere-EX.
92 * Test alpha=Nan in dscale.
93 * Fixed a SEGFAULT bug in samax on x86 windows.
95 ====================================================================
99 * Set soname of shared library on Linux.
100 * Added LIBNAMESUFFIX flag in Makefile.rule. The user can use
101 this flag to control the library name, e.g. libopenblas.a,
102 libopenblas_ifort.a or libopenblas_omp.a.
103 * Added GEMM_MULTITHREAD_THRESHOLD flag in Makefile.rule.
104 The lib use single thread in GEMM function with small matrices.
106 * Used GEMV SSE/SSE2 kernels on x86 32-bit.
107 * Exported CBLAS functions in Windows DLL.
109 * Completed Level-3 BLAS optimization on Loongson 3A CPU.
110 * Improved GEMV performance on Loongson 3A CPU.
111 * Improved Level-3 BLAS performance on Loongson 3B CPU. (EXPERIMENT)
113 ====================================================================
117 * Fixed missing "#include <sched.h>" bug on Mac OS X.
118 Thank Mike Nolta for the patch.
119 * Upgraded LAPACK to 3.4.0 version
120 * Fixed a bug on Mac OS X. Don't require SystemStubs on OS X.
121 SystemStubs does not exist on Lion. Thank Stefan Karpinski.
122 * Improved README with using OpenMP. Check the internal threads
123 count less than or equal to omp_get_max_threads()
125 * Auto-detect Intel Core i6/i7 (Sandy Bridge) CPU with Nehalem assembly kernels
126 * Fixed some bugs on MingW 64-bit including zgemv, cdot, zdot.
128 ====================================================================
132 * Fixed a bug about installation. The header file "fblas77.h"
134 * Fixed #61 a building bug about setting TARGET and DYNAMIC_ARCH.
135 * Try to handle absolute path of shared library in OSX. (#57)
136 Thank Dr Kane O'Donnell.
137 * Changed the installation folder layout to $(PREFIX)/include and
141 * Fixed #58 zdot/xdot SEGFAULT bug with GCC-4.6 on x86. According
142 to i386 calling convention, The callee should remove the first
143 hidden parameter.Thank Mr. John for this patch.
145 ====================================================================
150 * Added DTB_ENTRIES into dynamic arch setting parameters. Now,
151 it can read DTB_ENTRIES on runtime. (Refs issue #55 on github)
153 ====================================================================
158 * Fixed a building bug when DYNAMIC_ARCH=1 & INTERFACE64=1.
159 (Refs issue #44 on github)
161 ====================================================================
166 * Stop the build and output the error message when detecting
167 fortran compiler failed. (Refs issue #42 on github)
169 ====================================================================
174 * Fixed blasint undefined bug in <cblas.h> file. Other software
175 could include this header successfully(Refs issue #13 on github)
176 * Fixed the SEGFAULT bug on 64 cores. On SMP server, the number
177 of CPUs or cores should be less than or equal to 64.(Refs issue #14
179 * Support "void goto_set_num_threads(int num_threads)" and "void
180 openblas_set_num_threads(int num_threads)" when USE_OPENMP=1
181 * Added extern "C" to support C++. Thank Tasio for the patch(Refs
183 * Provided an error message when the arch is not supported.(Refs
185 * Fixed issue #23. Fixed a bug of f_check script about generating link flags.
186 * Added openblas_set_num_threads for Fortran.
187 * Fixed #25 a wrong result of rotmg.
188 * Fixed a bug about detecting underscore prefix in c_check.
189 * Print the wall time (cycles) with enabling FUNCTION_PROFILE
190 * Fixed #35 a build bug with NO_LAPACK=1 & DYNAMIC_ARCH=1
191 * Added install target. You can use "make install". (Refs #20)
195 * Fixed #28 a wrong result of dsdot on x86_64.
196 * Fixed #32 a SEGFAULT bug of zdotc with gcc-4.6.
197 * Fixed #33 ztrmm bug on Nehalem.
198 * Walk round #27 the low performance axpy issue with small imput size & multithreads.
201 * Fixed #28 a wrong result of dsdot on Loongson3A/MIPS64.
202 * Optimized single/double precision BLAS Level3 on Loongson3A/MIPS64. (Refs #2)
203 * Optimized single/double precision axpy function on Loongson3A/MIPS64. (Refs #3)
205 ====================================================================
210 * Support "make NO_LAPACK=1" to build the library without
212 * Fixed randomly SEGFAULT when nodemask==NULL with above Linux 2.6.34.
213 Thank Mr.Ei-ji Nakama providing this patch. (Refs issue #12 on github)
214 * Added DEBUG=1 rule in Makefile.rule to build debug version.
215 * Disable compiling quad precision in reference BLAS library(netlib BLAS).
216 * Added unit testcases in utest/ subdir. Used CUnit framework.
217 * Supported OPENBLAS_* & GOTO_* environment variables (Pleas see README)
218 * Imported GotoBLAS2 1.13 BSD version
221 * On x86 32bits, fixed a bug in zdot_sse2.S line 191. This would casue
222 zdotu & zdotc failures.Instead,Walk around it. (Refs issue #8 #9 on github)
223 * Modified ?axpy functions to return same netlib BLAS results
224 when incx==0 or incy==0 (Refs issue #7 on github)
225 * Modified ?swap functions to return same netlib BLAS results
226 when incx==0 or incy==0 (Refs issue #6 on github)
227 * Modified ?rot functions to return same netlib BLAS results
228 when incx==0 or incy==0 (Refs issue #4 on github)
229 * Detect Intel Westmere,Intel Clarkdale and Intel Arrandale
230 to use Nehalem codes.
231 * Fixed a typo bug about compiling dynamic ARCH library.
233 * Improve daxpy performance on ICT Loongson 3A.
234 * Supported ICT Loongson 3A CPU (Refs issue #1 on github)
235 ====================================================================