2 ====================================================================
6 * Fixed a bug about installation. The header file "fblas77.h"
8 * Fixed #61 a building bug about setting TARGET and DYNAMIC_ARCH.
9 * Try to handle absolute path of shared library in OSX. (#57)
10 Thank Dr Kane O'Donnell.
11 * Changed the installation folder layout to $(PREFIX)/include and
15 * Fixed #58 zdot/xdot SEGFAULT bug with GCC-4.6 on x86. According
16 to i386 calling convention, The callee should remove the first
17 hidden parameter.Thank Mr. John for this patch.
19 ====================================================================
24 * Added DTB_ENTRIES into dynamic arch setting parameters. Now,
25 it can read DTB_ENTRIES on runtime. (Refs issue #55 on github)
27 ====================================================================
32 * Fixed a building bug when DYNAMIC_ARCH=1 & INTERFACE64=1.
33 (Refs issue #44 on github)
35 ====================================================================
40 * Stop the build and output the error message when detecting
41 fortran compiler failed. (Refs issue #42 on github)
43 ====================================================================
48 * Fixed blasint undefined bug in <cblas.h> file. Other software
49 could include this header successfully(Refs issue #13 on github)
50 * Fixed the SEGFAULT bug on 64 cores. On SMP server, the number
51 of CPUs or cores should be less than or equal to 64.(Refs issue #14
53 * Support "void goto_set_num_threads(int num_threads)" and "void
54 openblas_set_num_threads(int num_threads)" when USE_OPENMP=1
55 * Added extern "C" to support C++. Thank Tasio for the patch(Refs
57 * Provided an error message when the arch is not supported.(Refs
59 * Fixed issue #23. Fixed a bug of f_check script about generating link flags.
60 * Added openblas_set_num_threads for Fortran.
61 * Fixed #25 a wrong result of rotmg.
62 * Fixed a bug about detecting underscore prefix in c_check.
63 * Print the wall time (cycles) with enabling FUNCTION_PROFILE
64 * Fixed #35 a build bug with NO_LAPACK=1 & DYNAMIC_ARCH=1
65 * Added install target. You can use "make install". (Refs #20)
69 * Fixed #28 a wrong result of dsdot on x86_64.
70 * Fixed #32 a SEGFAULT bug of zdotc with gcc-4.6.
71 * Fixed #33 ztrmm bug on Nehalem.
72 * Walk round #27 the low performance axpy issue with small imput size & multithreads.
75 * Fixed #28 a wrong result of dsdot on Loongson3A/MIPS64.
76 * Optimized single/double precision BLAS Level3 on Loongson3A/MIPS64. (Refs #2)
77 * Optimized single/double precision axpy function on Loongson3A/MIPS64. (Refs #3)
79 ====================================================================
84 * Support "make NO_LAPACK=1" to build the library without
86 * Fixed randomly SEGFAULT when nodemask==NULL with above Linux 2.6.34.
87 Thank Mr.Ei-ji Nakama providing this patch. (Refs issue #12 on github)
88 * Added DEBUG=1 rule in Makefile.rule to build debug version.
89 * Disable compiling quad precision in reference BLAS library(netlib BLAS).
90 * Added unit testcases in utest/ subdir. Used CUnit framework.
91 * Supported OPENBLAS_* & GOTO_* environment variables (Pleas see README)
92 * Imported GotoBLAS2 1.13 BSD version
95 * On x86 32bits, fixed a bug in zdot_sse2.S line 191. This would casue
96 zdotu & zdotc failures.Instead,Walk around it. (Refs issue #8 #9 on github)
97 * Modified ?axpy functions to return same netlib BLAS results
98 when incx==0 or incy==0 (Refs issue #7 on github)
99 * Modified ?swap functions to return same netlib BLAS results
100 when incx==0 or incy==0 (Refs issue #6 on github)
101 * Modified ?rot functions to return same netlib BLAS results
102 when incx==0 or incy==0 (Refs issue #4 on github)
103 * Detect Intel Westmere,Intel Clarkdale and Intel Arrandale
104 to use Nehalem codes.
105 * Fixed a typo bug about compiling dynamic ARCH library.
107 * Improve daxpy performance on ICT Loongson 3A.
108 * Supported ICT Loongson 3A CPU (Refs issue #1 on github)
109 ====================================================================