Multithreaded sparse matrix-matrix multiplication for many-core and GPU architectures

Journal Article · Mon Jul 09 00:00:00 EDT 2018 · Parallel Computing

DOI:https://doi.org/10.1016/j.parco.2018.06.009· OSTI ID:1466997

Deveci, Mehmet ^[1]; Trott, Christian ^[1]; Rajamanickam, Sivasankaran ^[1]

---

1. Sandia National Lab. (SNL-NM), Albuquerque, NM (United States)

Sparse matrix-matrix multiplication is a key kernel that has applications in several domains such as scientific computing and graph analysis. Several algorithms have been studied in the past for this foundational kernel. In this work, we develop parallel algorithms for sparse matrix-matrix multiplication with a focus on performance portability across different high performance computing architectures. The performance of these algorithms depend on the data structures used in them. We compare different types of accumulators in these algorithms and demonstrate the performance difference between these data structures. Furthermore, we develop a meta-algorithm, kkSpGEMM, to choose the right algorithm and data structure based on the characteristics of the problem. We show performance comparisons on three architectures and demonstrate the need for the community to develop two phase sparse matrix-matrix multiplication implementations for efficient reuse of the data structures involved.

• Research Organization: Sandia National Lab. (SNL-NM), Albuquerque, NM (United States)
• Sponsoring Organization: USDOE National Nuclear Security Administration (NNSA)
• Grant/Contract Number: AC04-94AL85000; NA-0003525
• OSTI ID: 1466997
• Alternate ID(s): OSTI ID: 1548025
• Report Number(s): SAND2017-13679J; 659607
• Journal Information: Parallel Computing, Vol. 78, Issue C; ISSN 0167-8191
• Publisher: Elsevier
• Country of Publication: United States
• Language: English

Related Subjects

97 MATHEMATICS AND COMPUTING
Sparse matrix sparse matrix multiplication
KNLs
GPUs
SpGEMM