magma: Matrix Algebra on GPU and Multicore Architectures

Magma matrix classes and methods for parallel processing of matrix algebra operations. Operations are performed with algorithms developed by the MAGMA research project. MAGMA aims to achieve the fastest possible linear algebra libraries on hybrid multicore CPU and GPU architectures by exploiting their massive parallelism and minimizing communication latencies.

Version: 1.3.0-2
Depends: R (≥ 2.11.0), methods
OS_type: unix
Published: 2013-04-03
Author: Brian J Smith
Maintainer: Brian J Smith <brian-j-smith at uiowa.edu>
License: GPL-3
URL: http://icl.cs.utk.edu/magma
NeedsCompilation: yes
SystemRequirements: MAGMA shared libraries (== release 1.3.0), LAPACK shared library, NVIDIA CUDA Toolkit (>= release 4.0)
Citation: magma citation info
Materials: ChangeLogINSTALL
In views: HighPerformanceComputing
CRAN checks: magma results

Downloads:

Reference manual: magma.pdf
Package source: magma_1.3.0-2.tar.gz
OS X binary: not available, see check log.
Windows binary: not available, see ReadMe.
Old sources: magma archive