2 ====================================================================
6 * Fixed missing "#include <sched.h>" bug on Mac OS X.
7 Thank Mike Nolta for the patch.
8 * Upgraded LAPACK to 3.4.0 version
9 * Fixed a bug on Mac OS X. Don't require SystemStubs on OS X.
10 SystemStubs does not exist on Lion. Thank Stefan Karpinski.
11 * Improved README with using OpenMP. Check the internal threads
12 count less than or equal to omp_get_max_threads()
14 * Auto-detect Intel Core i6/i7 (Sandy Bridge) CPU with Nehalem assembly kernels
15 * Fixed some bugs on MingW 64-bit including zgemv, cdot, zdot.
17 ====================================================================
21 * Fixed a bug about installation. The header file "fblas77.h"
23 * Fixed #61 a building bug about setting TARGET and DYNAMIC_ARCH.
24 * Try to handle absolute path of shared library in OSX. (#57)
25 Thank Dr Kane O'Donnell.
26 * Changed the installation folder layout to $(PREFIX)/include and
30 * Fixed #58 zdot/xdot SEGFAULT bug with GCC-4.6 on x86. According
31 to i386 calling convention, The callee should remove the first
32 hidden parameter.Thank Mr. John for this patch.
34 ====================================================================
39 * Added DTB_ENTRIES into dynamic arch setting parameters. Now,
40 it can read DTB_ENTRIES on runtime. (Refs issue #55 on github)
42 ====================================================================
47 * Fixed a building bug when DYNAMIC_ARCH=1 & INTERFACE64=1.
48 (Refs issue #44 on github)
50 ====================================================================
55 * Stop the build and output the error message when detecting
56 fortran compiler failed. (Refs issue #42 on github)
58 ====================================================================
63 * Fixed blasint undefined bug in <cblas.h> file. Other software
64 could include this header successfully(Refs issue #13 on github)
65 * Fixed the SEGFAULT bug on 64 cores. On SMP server, the number
66 of CPUs or cores should be less than or equal to 64.(Refs issue #14
68 * Support "void goto_set_num_threads(int num_threads)" and "void
69 openblas_set_num_threads(int num_threads)" when USE_OPENMP=1
70 * Added extern "C" to support C++. Thank Tasio for the patch(Refs
72 * Provided an error message when the arch is not supported.(Refs
74 * Fixed issue #23. Fixed a bug of f_check script about generating link flags.
75 * Added openblas_set_num_threads for Fortran.
76 * Fixed #25 a wrong result of rotmg.
77 * Fixed a bug about detecting underscore prefix in c_check.
78 * Print the wall time (cycles) with enabling FUNCTION_PROFILE
79 * Fixed #35 a build bug with NO_LAPACK=1 & DYNAMIC_ARCH=1
80 * Added install target. You can use "make install". (Refs #20)
84 * Fixed #28 a wrong result of dsdot on x86_64.
85 * Fixed #32 a SEGFAULT bug of zdotc with gcc-4.6.
86 * Fixed #33 ztrmm bug on Nehalem.
87 * Walk round #27 the low performance axpy issue with small imput size & multithreads.
90 * Fixed #28 a wrong result of dsdot on Loongson3A/MIPS64.
91 * Optimized single/double precision BLAS Level3 on Loongson3A/MIPS64. (Refs #2)
92 * Optimized single/double precision axpy function on Loongson3A/MIPS64. (Refs #3)
94 ====================================================================
99 * Support "make NO_LAPACK=1" to build the library without
101 * Fixed randomly SEGFAULT when nodemask==NULL with above Linux 2.6.34.
102 Thank Mr.Ei-ji Nakama providing this patch. (Refs issue #12 on github)
103 * Added DEBUG=1 rule in Makefile.rule to build debug version.
104 * Disable compiling quad precision in reference BLAS library(netlib BLAS).
105 * Added unit testcases in utest/ subdir. Used CUnit framework.
106 * Supported OPENBLAS_* & GOTO_* environment variables (Pleas see README)
107 * Imported GotoBLAS2 1.13 BSD version
110 * On x86 32bits, fixed a bug in zdot_sse2.S line 191. This would casue
111 zdotu & zdotc failures.Instead,Walk around it. (Refs issue #8 #9 on github)
112 * Modified ?axpy functions to return same netlib BLAS results
113 when incx==0 or incy==0 (Refs issue #7 on github)
114 * Modified ?swap functions to return same netlib BLAS results
115 when incx==0 or incy==0 (Refs issue #6 on github)
116 * Modified ?rot functions to return same netlib BLAS results
117 when incx==0 or incy==0 (Refs issue #4 on github)
118 * Detect Intel Westmere,Intel Clarkdale and Intel Arrandale
119 to use Nehalem codes.
120 * Fixed a typo bug about compiling dynamic ARCH library.
122 * Improve daxpy performance on ICT Loongson 3A.
123 * Supported ICT Loongson 3A CPU (Refs issue #1 on github)
124 ====================================================================