Compressed Multirow Storage Format for Sparse Matrices on Graphics Processing Units

Zbigniew Koza, Maciej Matyka, Sebastian Szkoda, Łukasz Mirosław
2014 SIAM Journal on Scientific Computing  
A new format for storing sparse matrices is proposed for efficient sparse matrix-vector (SpMV) product calculation on modern graphics processing units (GPUs). This format extends the standard compressed row storage (CRS) format and can be quickly converted to and from it. Computational performance of two SpMV kernels for the new format is determined for over 130 sparse matrices on Fermi-class and Kepler-class GPUs and compared with that of five existing generic algorithms and industrial
more » ... industrial implementations, including Nvidia cuSparse CSR and HYB kernels. We found the speedup of up to $\approx 60%$ over the best of the five alternative kernels.
doi:10.1137/120900216 fatcat:epbpxogrcvfhjobuyanno7ajou