Automatically Tuned Linear Algebra Software - ATLAS
Top500 Fastest Installed Computers
Performance Development
Architectures
Chip Technology
Excerpt from TOP500
High-Performance Computing Directions
Where Does the Performance Go? or Why Should I Cares About the Memory Hierarchy?
Memory Hierarchy
Performance Software
Performance Issues - Cache &Bandwidth
How To Get Performance From Commodity Processors?
ATLAS
Why ATLAS Is Fast?
Why Adaptive Programs?
Adaptive Approach for Level 3 BLAS
Code Generation Strategy
ATLAS 500x500 DGEMM Across Various Architectures
500 x 500 LU Right-Looking
Recursive Approach for Other Level 3 BLAS
500x500 Recursive BLAS on UltraSparc 2200
500x500 Recursive BLAS on 433Mhz DEC 21164
Multithreaded BLAS for Performance
Ultra Sparc 2200 - Matrix Mult. (200 MHz 400 Mflop/s Peak)
Future Plans for ATLAS
Contributors to These Ideas
Email: dongarra@cs.utk.edu
Home Page: http://www.netlib.org/utk/people/JackDongarra/