Performance Optimization and Modeling of Blocked Sparse Kernels

Alfredo Buttari 1 Victor Eijkhout 2 Julien Langou 3 Salvatore Filippone 4
1 IRIT - Institut de recherche en informatique de Toulouse
2 Texas Advanced Computing Center, The University of Texas at Austin, Austin, Texas
Texas Advanced Computing Center
3 Department of Mathematical and Statistical Sciences
4 University of Rome "Tor Vergeta"
Abstract : We present a method for automatically selecting optimal implementations of sparse matrix-vector operations. Our software “AcCELS” (Accelerated Compress-storage Elements for Linear Solvers) involves a setup phase that probes machine characteristics, and a run-time phase where stored characteristics are combined with a measure of the actual sparse matrix to find the optimal kernel implementation. We present a performance model that is shown to be accurate over a large range of matrices.
Document type :
Journal articles
Complete list of metadatas
https://hal.archives-ouvertes.fr/hal-02421133
Contributor : Alfredo Buttari <>
Submitted on : Friday, December 20, 2019 - 11:54:14 AM
Last modification on : Friday, January 10, 2020 - 9:09:27 PM

Links full text

openaccess

Identifiers

Collections

UNIV-TLSE3 | UT1-CAPITOLE | SMS | UNIV-TLSE2 | CNRS

Citation

Alfredo Buttari, Victor Eijkhout, Julien Langou, Salvatore Filippone. Performance Optimization and Modeling of Blocked Sparse Kernels. International Journal of High Performance Computing Applications, SAGE Publications, 2016, 21 (4), pp.467-484. ⟨10.1177/1094342007083801⟩. ⟨hal-02421133⟩

Export

BibTeX TEI DC DCterms EndNote

Share

Metrics

Record views

25

Contact

Informations

CCSD