Mixed Precision Iterative Refinement Techniques for the Solution of Dense Linear Systems

Abstract : By using a combination of 32-bit and 64-bit floating point arithmetic, the performance of many dense and sparse linear algebra algorithms can be significantly enhanced while maintaining the 64-bit accuracy of the resulting solution. The approach presented here can apply not only to conventional processors but also to exotic technologies such as Field Programmable Gate Arrays (FPGA), Graphical Processing Units (GPU), and the Cell BE processor. Results on modern processor architectures and the Cell BE are presented.
Complete list of metadatas

https://hal.archives-ouvertes.fr/hal-02421121
Contributor : Alfredo Buttari <>
Submitted on : Friday, December 20, 2019 - 11:51:01 AM
Last modification on : Friday, January 10, 2020 - 9:09:27 PM

Links full text

Identifiers

Collections

Citation

Alfredo Buttari, Jack Dongarra, Julie Langou, Julien Langou, Piotr Luszczek, et al.. Mixed Precision Iterative Refinement Techniques for the Solution of Dense Linear Systems. International Journal of High Performance Computing Applications, SAGE Publications, 2016, 21 (4), pp.457-466. ⟨10.1177/1094342007084026⟩. ⟨hal-02421121⟩

Share

Metrics

Record views

23