3179

GEMM on a GPU

Jonathan Monette
University of Texas, Austin
Undergraduate Research Forum at the University of Texas at Austin. April 2010
BibTeX

Download Download (PDF)   View View   Source Source   

2434

views

The Matrix-Matrix Multiplication is the most important operation in High-Performance Linear Algebra. If your application can cast most of its computation in terms of the level-3 BLAS operations, the application can achieve very high-performance levels. For this reason the Basic Linear Algebra Subprograms(BLAS) tend to heavily optimize this operation. With Graphics Processing Units(GPUs) on the rise in the field of highperformance computing, exposing the parallelism in this operation becomes increasingly more important.
No votes yet.
Please wait...

* * *

* * *

HGPU group © 2010-2025 hgpu.org

All rights belong to the respective authors

Contact us:

contact@hpgu.org