3 [![Join the chat at https://gitter.im/xianyi/OpenBLAS](https://badges.gitter.im/Join%20Chat.svg)](https://gitter.im/xianyi/OpenBLAS?utm_source=badge&utm_medium=badge&utm_campaign=pr-badge&utm_content=badge)
5 Travis CI: [![Build Status](https://travis-ci.org/xianyi/OpenBLAS.svg?branch=develop)](https://travis-ci.org/xianyi/OpenBLAS)
7 AppVeyor: [![Build status](https://ci.appveyor.com/api/projects/status/09sohd35n8nkkx64/branch/develop?svg=true)](https://ci.appveyor.com/project/xianyi/openblas/branch/develop)
11 OpenBLAS is an optimized BLAS library based on GotoBLAS2 1.13 BSD version.
13 Please read the documentation on the OpenBLAS wiki pages: <http://github.com/xianyi/OpenBLAS/wiki>.
17 We provide official binary packages for the following platform:
21 You can download them from [file hosting on sourceforge.net](https://sourceforge.net/projects/openblas/files/).
23 ## Installation from Source
25 Download from project homepage, http://xianyi.github.com/OpenBLAS/, or check out the code
26 using Git from https://github.com/xianyi/OpenBLAS.git.
30 Building OpenBLAS requires the following to be installed:
33 * A C compiler, e.g. GCC or Clang
34 * A Fortran compiler (optional, for LAPACK)
35 * IBM MASS (optional, see below)
39 Simply invoking `make` (or `gmake` on BSD) will detect the CPU automatically.
40 To set a specific target CPU, use `make TARGET=xxx`, e.g. `make TARGET=NEHALEM`.
41 The full target list is in the file `TargetList.txt`.
45 Set `CC` and `FC` to point to the cross toolchains, and set `HOSTCC` to your host C compiler.
46 The target must be specified explicitly when cross compiling.
50 * On an x86 box, compile this library for a loongson3a CPU:
52 make BINARY=64 CC=mips64el-unknown-linux-gnu-gcc FC=mips64el-unknown-linux-gnu-gfortran HOSTCC=gcc TARGET=LOONGSON3A
55 * On an x86 box, compile this library for a loongson3a CPU with loongcc (based on Open64) compiler:
57 make CC=loongcc FC=loongf95 HOSTCC=gcc TARGET=LOONGSON3A CROSS=1 CROSS_SUFFIX=mips64el-st-linux-gnu- NO_LAPACKE=1 NO_SHARED=1 BINARY=32
62 A debug version can be built using `make DEBUG=1`.
64 ### Compile with MASS support on Power CPU (optional)
66 The [IBM MASS](http://www-01.ibm.com/software/awdtools/mass/linux/mass-linux.html) library
67 consists of a set of mathematical functions for C, C++, and Fortran applications that are
68 are tuned for optimum performance on POWER architectures.
69 OpenBLAS with MASS requires a 64-bit, little-endian OS on POWER.
70 The library can be installed as shown:
74 wget -q http://public.dhe.ibm.com/software/server/POWER/Linux/xl-compiler/eval/ppc64le/ubuntu/public.gpg -O- | sudo apt-key add -
75 echo "deb http://public.dhe.ibm.com/software/server/POWER/Linux/xl-compiler/eval/ppc64le/ubuntu/ trusty main" | sudo tee /etc/apt/sources.list.d/ibm-xl-compiler-eval.list
77 sudo apt-get install libxlmass-devel.8.1.5
82 wget http://public.dhe.ibm.com/software/server/POWER/Linux/xl-compiler/eval/ppc64le/rhel7/repodata/repomd.xml.key
83 sudo rpm --import repomd.xml.key
84 wget http://public.dhe.ibm.com/software/server/POWER/Linux/xl-compiler/eval/ppc64le/rhel7/ibm-xl-compiler-eval.repo
85 sudo cp ibm-xl-compiler-eval.repo /etc/yum.repos.d/
86 sudo yum install libxlmass-devel.8.1.5
89 After installing the MASS library, compile OpenBLAS with `USE_MASS=1`.
90 For example, to compile on Power8 with MASS support: `make USE_MASS=1 TARGET=POWER8`.
92 ### Install to a specific directory (optional)
94 Use `PREFIX=` when invoking `make`, for example
97 make install PREFIX=your_installation_directory
100 The default installation directory is `/opt/OpenBLAS`.
102 ## Supported CPUs and Operating Systems
104 Please read `GotoBLAS_01Readme.txt`.
106 ### Additional supported CPUs
110 - **Intel Xeon 56xx (Westmere)**: Used GotoBLAS2 Nehalem codes.
111 - **Intel Sandy Bridge**: Optimized Level-3 and Level-2 BLAS with AVX on x86-64.
112 - **Intel Haswell**: Optimized Level-3 and Level-2 BLAS with AVX2 and FMA on x86-64.
113 - **AMD Bobcat**: Used GotoBLAS2 Barcelona codes.
114 - **AMD Bulldozer**: x86-64 ?GEMM FMA4 kernels. (Thanks to Werner Saar)
115 - **AMD PILEDRIVER**: Uses Bulldozer codes with some optimizations.
116 - **AMD STEAMROLLER**: Uses Bulldozer codes with some optimizations.
120 - **ICT Loongson 3A**: Optimized Level-3 BLAS and the part of Level-1,2.
121 - **ICT Loongson 3B**: Experimental
125 - **ARMv6**: Optimized BLAS for vfpv2 and vfpv3-d16 (e.g. BCM2835, Cortex M0+)
126 - **ARMv7**: Optimized BLAS for vfpv3-d32 (e.g. Cortex A8, A9 and A15)
130 - **ARMv8**: Experimental
131 - **ARM Cortex-A57**: Experimental
135 - **POWER8**: Optmized Level-3 BLAS and some Level-1, only with `USE_OPENMP=1`
137 #### IBM zEnterprise System
139 - **Z13**: Optimized Level-3 BLAS and Level-1,2 (double precision)
144 - **MinGW or Visual Studio (CMake)/Windows**: Please read <https://github.com/xianyi/OpenBLAS/wiki/How-to-use-OpenBLAS-in-Microsoft-Visual-Studio>.
145 - **Darwin/macOS**: Experimental. Although GotoBLAS2 supports Darwin, we are not macOS experts.
146 - **FreeBSD**: Supported by the community. We don't actively test the library on this OS.
147 - **OpenBSD**: Supported by the community. We don't actively test the library on this OS.
148 - **DragonFly BSD**: Supported by the community. We don't actively test the library on this OS.
149 - **Android**: Supported by the community. Please read <https://github.com/xianyi/OpenBLAS/wiki/How-to-build-OpenBLAS-for-Android>.
153 Statically link with `libopenblas.a` or dynamically link with `-lopenblas` if OpenBLAS was
154 compiled as a shared library.
156 ### Setting the number of threads using environment variables
158 Environment variables are used to specify a maximum number of threads.
162 export OPENBLAS_NUM_THREADS=4
163 export GOTO_NUM_THREADS=4
164 export OMP_NUM_THREADS=4
167 The priorities are `OPENBLAS_NUM_THREADS` > `GOTO_NUM_THREADS` > `OMP_NUM_THREADS`.
169 If you compile this library with `USE_OPENMP=1`, you should set the `OMP_NUM_THREADS`
170 environment variable; OpenBLAS ignores `OPENBLAS_NUM_THREADS` and `GOTO_NUM_THREADS` when
171 compiled with `USE_OPENMP=1`.
173 ### Setting the number of threads at runtime
175 We provide the following functions to control the number of threads at runtime:
178 void goto_set_num_threads(int num_threads);
179 void openblas_set_num_threads(int num_threads);
182 If you compile this library with `USE_OPENMP=1`, you should use the above functions too.
186 Please submit an issue in https://github.com/xianyi/OpenBLAS/issues.
190 * OpenBLAS users mailing list: https://groups.google.com/forum/#!forum/openblas-users
191 * OpenBLAS developers mailing list: https://groups.google.com/forum/#!forum/openblas-dev
195 Please see Changelog.txt to view the differences between OpenBLAS and GotoBLAS2 1.13 BSD version.
199 * Please read the [FAQ](https://github.com/xianyi/OpenBLAS/wiki/Faq) first.
200 * Please use GCC version 4.6 and above to compile Sandy Bridge AVX kernels on Linux/MinGW/BSD.
201 * Please use Clang version 3.1 and above to compile the library on Sandy Bridge microarchitecture.
202 Clang 3.0 will generate the wrong AVX binary code.
203 * The number of CPUs/cores should less than or equal to 256. On Linux `x86_64` (`amd64`),
204 there is experimental support for up to 1024 CPUs/cores and 128 numa nodes if you build
205 the library with `BIGNUMA=1`.
206 * OpenBLAS does not set processor affinity by default.
207 On Linux, you can enable processor affinity by commenting out the line `NO_AFFINITY=1` in
208 Makefile.rule. However, note that this may cause
209 [a conflict with R parallel](https://stat.ethz.ch/pipermail/r-sig-hpc/2012-April/001348.html).
210 * On Loongson 3A, `make test` may fail with a `pthread_create` error (`EAGAIN`).
211 However, it will be okay when you run the same test case on the shell.
215 1. [Check for open issues](https://github.com/xianyi/OpenBLAS/issues) or open a fresh issue
216 to start a discussion around a feature idea or a bug.
217 2. Fork the [OpenBLAS](https://github.com/xianyi/OpenBLAS) repository to start making your changes.
218 3. Write a test which shows that the bug was fixed or that the feature works as expected.
219 4. Send a pull request. Make sure to add yourself to `CONTRIBUTORS.md`.
223 Please read [this wiki page](https://github.com/xianyi/OpenBLAS/wiki/Donation).