A GPU Algorithm for Matrix-Matrix Multiplication using OpenGL Compute Shader
Park SangKun · Korean Journal of Computational Design and Engineering · 2018
In this paper we present a GPU algorithm for dense or sparse matrix-matrix multiplication using OpenGL compute shader, which can play a very important role as a fundamental building block for many high-performance computing applications. It includes the matrix product algorithm for dense-dense, sparse-dense, sparse-sparse matrix multiplication, and the matrix conversion algorithm from sparse to dense. Experimental results on NVIDIA Quad 4000 show that the proposed algorithm runs 184 times faster than CPU algorithm and achieves performance of 71 GFLOPS in single precision for dense or sparse matrices with size 4,096. Such performance proves that our algorithm is practical for real applications.