Evaluation of Dynamic Data Distributions on NUMA Shared Memory Multiprocessors

Dynamic data distributions offer a number of performance benefits, but require more sophisticated compiler support and incur run-time overhead. We investigate attainable benefits using a compiler system we developed for the Hector NUMA multiprocessor. We show that the benefits depend on a number of factors, includingdata size relative to the cache size, data access patterns, the degree of "NUMAness" of the multiprocessor system, and the extent to which data is reused. Programmers and compiler designers must take these factors into consideration.

• Type: Conference
• Year: 1996
• Where: PDPTA
• Authors: Tarek S. Abdelrahman, Kenneth L. Ma
• Keywords: Hector NUMA Multiprocessor, Incur Run-time Overhead, PDPTA 1996, PDPTA 2007, Sophisticated Compiler Support