Efficient Algorithm Design of Optimizing SpMV on GPU
Genshen Chu, Yuanjie He, Lingyu Dong, Zhezhao Ding, Dandan Chen, He Bai, Xuesong Wang, Changjun Hu
Abstract
Sparse matrix-vector multiplication (SpMV) is a fundamental building block for various numerical computing applications. However, most existing GPU-SpMV approaches may suffer from either long preprocessing overhead, load imbalance, format conversion, bad memory access patterns. In this paper, we proposed two new SpMV algorithms:flat andline-enhance, as well as their implementations, for GPU systems to overcome the above shortcomings. Our algorithms work directly on the CSR sparse matrix format. To achieve high performance: 1) for load balance, theflat algorithm uses non-zero splitting andline-enhance uses a mix of row and non-zero splitting; 2) memory access patterns are designed for both algorithms for data loading, storing and reduction steps; and 3) an adaptive approach is proposed to select appropriate algorithm and parameters based on matrix characteristics.