| Autor: |
Aktemur, Baris, Yildiz, Asim, Kamin, Sam |
| Zdroj: |
2012 20th Signal Processing & Communications Applications Conference (SIU); 1/ 1/2012, p1-4, 4p |
| Abstrakt: |
In this work we present the design of a library which generates high-performance sparse matrix-vector multiplication code that is specialized according to the contents of a given matrix. The library targets code generation in particular for matrices of large size as used in signal processing applications, scientific computation and finite element analysis. The generated code is selected out of many choices by empirical optimization to obtain the best code tuned for a particular machine. [ABSTRACT FROM PUBLISHER] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
