A Comparison of the Performance of SaP::GPU and Intel's Math Kernel Library (MKL) for Solving Dense Banded Linear Systems
Ang Li, Omkar Deshmukh, Radu Serban, Dan Negruţ · 2015
SaP::GPU is a solver developed in the Simulation Based Engineering Lab (SBEL) [1] to solve large banded and sparse linear systems on the GPU. This report contributes the performance comparison of the banded solver of SaP::GPU and Intel’s Math Kernel Library [2] on a large set of synthetic problems. The results of several numerical experiments indicate that when it is used in conjunction with large dense banded matrices, SaP::GPU is two to five times faster than the latest version of the MKL dense solver when the latter is run on the Haswell, Ivy Bridge, or Phi architectures.