Title :
Comparison of scalable parallel matrix multiplication libraries
Author :
Huss-Lederman, Steven ; Jacobson, Elaine M. ; Tsao, Anna
Author_Institution :
Supercomputing Res. Center, Bowie, MD, USA
Abstract :
This paper compares two general library routines for performing parallel distributed matrix multiplication. The PUMMA algorithm utilities block scattered data layout, whereas BiMMeR utilizes virtual 2-D torus wrap. The algorithmic differences resulting from these different layouts are discussed us well as the general issues associated with different data layouts for library routines. Results on the Intel Delta for the two matrix multiplication algorithms are presented
Keywords :
data structures; mathematics computing; matrix algebra; parallel algorithms; BiMMeR; Intel Delta; PUMMA algorithm; block scattered data layout; matrix multiplication algorithms; parallel distributed matrix multiplication; scalable parallel matrix multiplication libraries; virtual 2-D torus wrap; Broadcasting; Distributed computing; Drives; Jacobian matrices; Kernel; Libraries; Matrix decomposition; Packaging machines; Scattering; Topology;
Conference_Titel :
Scalable Parallel Libraries Conference, 1993., Proceedings of the
Conference_Location :
Mississippi State, MS
Print_ISBN :
0-8186-4980-1
DOI :
10.1109/SPLC.1993.365573
