BLAS and LAPACK: Difference between revisions

[http://www.netlib.org/blas/ BLAS (Basic Linear Algebra Subprogram] and [http://www.netlib.org/lapack/ LAPACK (Linear Algebra PACK)] are two of the most commonly used libraries in advanced research computing. They are used for vector and matrix operations that are so commonly found in a plethora of algorithms. More importantly, they are more than libraries, as they define a standard programming interface. A programming interface is a set of function definitions that can be called to accomplish specific computation, for example the dot product of two vectors of double precision numbers, or the matrix product of two hermitian matrices of complex numbers.  

Beside the reference implementation done by Netlib, there exist a large number of implementations of these two standards. The performance of these implementations can vary widely depending on the hardware that is running them. For example, it is well established that the implementation provided by the [https://www.intel.com/content/www/us/en/developer/tools/oneapi/onemkl.html Intel Math Kernel Library (Intel MKL)] performs best in most situations on Intel processors. That implementation is however proprietary, and in some situations, it is preferred to use the open source implementation [https://github.com/xianyi/OpenBLAS OpenBLAS]. In 2018, AMD also released its own implementation, named [https://developer.amd.com/amd-aocl/blas-library/ AMD BLIS], which performs better on AMD processors. Previously, you may have known [https://www.tacc.utexas.edu/research-development/tacc-software/gotoblas2 gotoblas] and [https://github.com/math-atlas/math-atlas ATLAS BLAS], but those projects are no longer maintained.

For the past few years, we have been recommending to use Intel MKL as a reference implementation. This recommendation was driven by the fact that we only had Intel processors in our clusters. This was change with the arrival of Narval, which hosts AMD processors. We now recommend using FlexiBLAS when compiling code. Our FlexiBLAS module is configured such that Intel MKL will be used by default except when using AMD processors, in which case AMD BLIS will be used, offering optimal performance.  

Unfortunately, FlexiBLAS is relatively new, and not all build systems will recognize it by default. This can generally be fixed by setting the linking options to use <tt>-lflexiblas</tt> for BLAS and for LAPACK. You will typically find these options in your Makefile, or be able to pass them as parameters to <tt>configure</tt> or <tt>cmake</tt>.  

The big benefit of using FlexiBLAS is to be able to change the implementation backend at run time by setting the environment variable <tt>FLEXIBLAS</tt>. At the time of this writing, four implementations are available: <tt>netlib</tt>, <tt>blis</tt>, <tt>imkl</tt> and <tt>openblas</tt>, but the full list can be obtained by running the command

On [[Narval]], we have set <tt>FLEXIBLAS=blis</tt> to use AMD's implementation by default, while on other clusters, <tt>FLEXIBLAS</tt> is left undefined, which defaults to using Intel MKL.

If you are using a non-Intel compiler, for example the Gnu compiler collection, then you will need to explicitly list the necessary MKL libraries during the link phase. Intel provides a tool called the [https://software.intel.com/en-us/articles/intel-mkl-link-line-advisor MKL Link Advisor] to help you find the correct compiler and linker options.