CUDA Accelerated Linpack on Clusters - Nvidia
CUDA Accelerated Linpack on Clusters - Nvidia
CUDA Accelerated Linpack on Clusters - Nvidia
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
LINPACK Benchmark<br />
Solve a dense NxN linear system:<br />
Ax=b<br />
Soluti<strong>on</strong> is obtained by Gaussian<br />
eliminati<strong>on</strong> with partial pivoting<br />
Floating point workload:<br />
2/3 N^3 + 2 N^2<br />
(LU decompositi<strong>on</strong>) (back solve)<br />
Factorize the current block (red), update the green and yellow parts when d<strong>on</strong>e<br />
The bigger the problem size N is, the more time is spent in the update (DGEMM)