ScaLAPACK is a library of high-performance linear algebra routines for parallel distributed memory machines. It depends on external libraries including BLAS and LAPACK for Linear Algebra computations. AMD’s optimized version of ScaLAPACK enables using the BLIS and libFLAME libraries with optimized dense-matrix functions and solvers for AMD processors.
Highlights of AMD ScaLAPACK 3.1
- Enabled ILP64 support
Highlights of AMD ScaLAPACK 3.0
- Improved LU Factorization for real double precision datatype on AMD EPYCTM processors
- Updated CMake build system. It can be used for building both static and shared libraries
- New API to get version number of the library, GET_AOCL_SCALAPACK_VERSION()
The package containing Scalapack binaries, examples and documentation are available in the Downloads section below.
Source code for AMD ScaLAPACK will be available shortly on GitHub https://github.com/amd/aocl-scalapack
Refer here for prior versions of ScaLAPACK documentation and downloads.