Low Thread-Count Gustavson: A Multithreaded Algorithm for Sparse Matrix-Matrix Multiplication Using Perfect Hashing

Sparse matrix-matrix multiplication is a critical kernel for several scientific computing applications, especially the setup phase of algebraic multigrid. The MPI+X programming model, which is growing in popularity, requires that such kernels be implemented in a way that exploits on-node parallelism...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:2018 IEEE/ACM 9th Workshop on Latest Advances in Scalable Algorithms for Large-Scale Systems (scalA) S. 57 - 64
Hauptverfasser: Elliott, James J., Siefert, Christopher M.
Format: Tagungsbericht
Sprache:Englisch
Veröffentlicht: IEEE 01.11.2018
Schlagworte:
algebraic-multigrid
Dynamic scheduling
Heuristic algorithms
Instruction sets
Kernel
Memory management
OpenMP
Partitioning algorithms
Sparse matrices
sparse-matrix-matrix-multiplciation
spgemm
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Sparse matrix-matrix multiplication is a critical kernel for several scientific computing applications, especially the setup phase of algebraic multigrid. The MPI+X programming model, which is growing in popularity, requires that such kernels be implemented in a way that exploits on-node parallelism. We present a single-pass OpenMP variant of Gustavson’s sparse matrix matrix multiplication algorithm designed for architectures (e.g. CPU or Intel Xeon Phi) with reasonably large memory and modest thread counts (tens of threads, not thousands). These assumptions allow us to exploit perfect hashing and dynamic memory allocation to achieve performance improvements of up to 2x over third-party kernels for matrices derived from algebraic multigrid setup.
DOI:10.1109/ScalA.2018.00011