Verification Challenges in Sparse Matrix Vector Multiplication in High Performance Computing: Part I

2025/10/15 by Zhang, Junchao
#Distributed #FOS: Computer and information sciences #Logic in Computer Science (cs.LO) #Mathematical Software (cs.MS) #Parallel #and Cluster Computing (cs.DC)

paper · doi:10.48550/arxiv.2510.13427

Abstract

Sparse matrix vector multiplication (SpMV) is a fundamental kernel in scientific codes that rely on iterative solvers. In this first part of our work, we present both a sequential and a basic MPI parallel implementations of SpMV, aiming to provide a challenge problem for the scientific software verification community. The implementations are described in the context of the PETSc library.

Related

• Algorithms for Parallel Shared-Memory Sparse Matrix-Vector Multiplication on Unstructured Matrices
• Towards Efficient Sparse Matrix Vector Multiplication on Real Processing-In-Memory Systems
• Performance Enhancement Strategies for Sparse Matrix-Vector Multiplication (SpMV) and Iterative Linear Solvers
• A survey of sparse matrix-vector multiplication performance on large\n matrices
• Performance evaluation of accelerated real and complex multiple-precision sparse matrix-vector multiplication