Please use this identifier to cite or link to this item:
|Title:||Accelerating sparse matrix-vector multiplication on GPUs using bit-representation-optimized schemes|
Sparse matrix format
|Citation:||Tang, W.T., Tan, W.J., Ray, R., Wong, Y.W., Chen, W., Kuo, S.-H., Goh, R.S.M., Turner, S.J., Wong, W.-F. (2013). Accelerating sparse matrix-vector multiplication on GPUs using bit-representation-optimized schemes. International Conference for High Performance Computing, Networking, Storage and Analysis, SC : -. ScholarBank@NUS Repository. https://doi.org/10.1145/2503210.2503234|
|Abstract:||The sparse matrix-vector (SpMV) multiplication routine is an important building block used in many iterative algo-rithms for solving scientific and engineering problems. One of the main challenges of SpMV is its memory-boundedness. Although compression has been proposed previously to im-prove SpMV performance on CPUs, its use has not been demonstrated on the GPU because of the serial nature of many compression and decompression schemes. In this pa-per, we introduce a family of bit-representation-optimized (BRO) compression schemes for representing sparse matrices on GPUs. The proposed schemes, BRO-ELL, BRO-COO, and BRO-HYB, perform compression on index data and help to speed up SpMV on GPUs through reduction of memory trafic. Furthermore, we formulate a BRO-aware matrix re-ordering scheme as a data clustering problem and use it to increase compression ratios. With the proposed schemes, ex-periments show that average speedups of 1.5 compared to ELLPACK and HYB can be achieved for SpMV on GPUs. Copyright 2013 ACM.|
|Source Title:||International Conference for High Performance Computing, Networking, Storage and Analysis, SC|
|Appears in Collections:||Staff Publications|
Show full item record
Files in This Item:
There are no files associated with this item.
checked on Dec 12, 2018
WEB OF SCIENCETM
checked on Dec 4, 2018
checked on Nov 23, 2018
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.