Flexible Batched Sparse Matrix-Vector Product on GPUs

TitleFlexible Batched Sparse Matrix-Vector Product on GPUs
Publication TypeConference Paper
Year of Publication2017
AuthorsAnzt, H., G. Collins, J. Dongarra, G. Flegar, and E. S. Quintana-Orti
Conference Name8th Workshop on Latest Advances in Scalable Algorithms for Large-Scale Systems (ScalA '17)
Date Published2017-11
PublisherACM Press
Conference LocationDenver, CO

We propose a variety of batched routines for concurrently processing a large collection of small-size, independent sparse matrix-vector products (SpMV) on graphics processing units (GPUs). These batched SpMV kernels are designed to be flexible in order to handle a batch of matrices which differ in size, nonzero count, and nonzero distribution. Furthermore, they support three most commonly used sparse storage formats: CSR, COO and ELL. Our experimental results on a state-of-the-art GPU reveal performance improvements of up to 25X compared to non-batched SpMV routines.

Project Tags: 
External Publication Flag: