AlphaSparse: Generating High Performance SpMV Codes Directly from Sparse Matrices

TL;DR

AlphaSparse automatically generates high-performance, machine-designed sparse matrix formats and SpMV kernels directly from input matrix patterns and hardware details, outperforming existing methods significantly.

Contribution

It introduces a novel framework that creates custom sparse formats and kernels automatically, surpassing human-designed approaches in efficiency.

Findings

Achieves 3.2× average speedup over state-of-the-art formats.

Attains 1.5× average improvement over traditional auto-tuning implementations.

Demonstrates effectiveness on 843 real-world matrices.

Abstract

Sparse Matrix-Vector multiplication (SpMV) is an essential computational kernel in many application scenarios. Tens of sparse matrix formats and implementations have been proposed to compress the memory storage and speed up SpMV performance. We develop AlphaSparse, a superset of all existing works that goes beyond the scope of human-designed format(s) and implementation(s). AlphaSparse automatically \emph{creates novel machine-designed formats and SpMV kernel implementations} entirely from the knowledge of input sparsity patterns and hardware architectures. Based on our proposed Operator Graph that expresses the path of SpMV format and kernel design, AlphaSparse consists of three main components: Designer, Format \& Kernel Generator, and Search Engine. It takes an arbitrary sparse matrix as input while outputs the performant machine-designed format and SpMV implementation. By extensively evaluating 843 matrices from SuiteSparse Matrix Collection, AlphaSparse achieves significant performance improvement by 3.2$\times$ on average compared to five state-of-the-art artificial formats and 1.5$\times$ on average (up to 2.7$\times$) over the up-to-date implementation of traditional auto-tuning philosophy.