Matrix Algebra on GPU and Multi-core Architectures (MAGMA) is a collection of next-generation linear algebra libraries for heterogeneous computing. The MAGMA package supports interfaces for current linear algebra packages and standards (e.g., LAPACK and BLAS) to enable computational scientists to easily port any linear algebra–reliant software components to heterogeneous computing systems. MAGMA enables applications to fully exploit the power of current hybrid systems of many-core CPUs and multi-GPUs/coprocessors to deliver the fastest possible time to accurate solution within given energy constraints.
MAGMA features LAPACK-compliant routines for multi-core CPUs enhanced with NVIDIA or AMD GPUs. MAGMA 2.6.2 now includes more than 400 routines that cover one-sided dense matrix factorizations and solvers, two-sided factorizations, and eigen/singular-value problem solvers, as well as a subset of highly optimized BLAS for GPUs. A MagmaDNN package has been added and further enhanced to provide high-performance data analytics, including functionalities for machine learning applications that use MAGMA as their computational back end. The MAGMA Sparse and MAGMA Batched packages have been included since MAGMA 1.6.
Find out more at http://icl.eecs.utk.edu/magma/
In Collaboration With
- King Abdullah University of Science and Technology
- University of California Berkeley
- University of Colorado Denver
With Support From