Exploring the effect of block shapes on the performance of sparse kernels

Karakasis, V; Goumas, G; Koziris, N

dc.contributor.author	Karakasis, V	en
dc.contributor.author	Goumas, G	en
dc.contributor.author	Koziris, N	en
dc.date.accessioned	2014-03-01T02:46:09Z
dc.date.available	2014-03-01T02:46:09Z
dc.date.issued	2009	en
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/32574
dc.subject	Experimental Evaluation	en
dc.subject	Memory Bandwidth	en
dc.subject	Sparse Matrix	en
dc.subject	Compressed Sparse Row	en
dc.subject.other	Compressed sparse row	en
dc.subject.other	Experimental evaluation	en
dc.subject.other	Memory bandwidths	en
dc.subject.other	Memory wall	en
dc.subject.other	Micro architectures	en
dc.subject.other	Multithreaded	en
dc.subject.other	Performance variations	en
dc.subject.other	Sparse kernels	en
dc.subject.other	Sparse matrix-vector multiplication	en
dc.subject.other	Storage formats	en
dc.subject.other	Vectorization	en
dc.subject.other	Shape memory effect	en
dc.subject.other	Distributed parameter networks	en
dc.title	Exploring the effect of block shapes on the performance of sparse kernels	en
heal.type	conferenceItem	en
heal.identifier.primary	10.1109/IPDPS.2009.5161159	en
heal.identifier.secondary	http://dx.doi.org/10.1109/IPDPS.2009.5161159	en
heal.identifier.secondary	5161159	en
heal.publicationDate	2009	en
heal.abstract	In this paper we explore the impact of the block shape on blocked and vectorized versions of the Sparse Matrix-Vector Multiplication (SpMV) kernel and build upon previous work by performing an extensive experimental evaluation of the most widespread blocking storage format, namely Block Compressed Sparse Row (BCSR) format, on a set of modern commodity microarchitectures. We evaluate the merit of vectorization on the memory-bound blocked SpMV kernel and report the results for single- and multithreaded (both SMP and NUMA) configurations. The performance of blocked SpMV can significantly vary with the block shape, despite similar memory bandwidth demands for different blocks. This is further accentuated when vectorizing the kernel. When moving to multiple cores, the memory wall problem becomes even more evident and may overwhelm any benefit from optimizations targeting the computational part of the kernel. In this paper we explore and discuss the architectural characteristics of modern commodity architectures that are responsible for these performance variations between block shapes. © 2009 IEEE.	en
heal.journalName	IPDPS 2009 - Proceedings of the 2009 IEEE International Parallel and Distributed Processing Symposium	en
dc.identifier.doi	10.1109/IPDPS.2009.5161159	en