An approach to solve the cache thrashing problem Conference Paper uri icon

abstract

  • 1991 IEEE. Cache or local memory thrashing problem arises very often in parallel processing architectures where each processor has its local cache or memory and a write-back protocol is employed for cache coherence. To solve the problem that large amount of data moving back and forth between the caches or local memories in different processors, techniques associated with parallel compiler need to be developed. Based on the relations between array element accesses and enclosed loop indices in a nested parallel construct, we present in this paper some approach to reduce the data movement between the caches or local memories for parallel programs. By analyzing the array subscript expressions, the compilers let the processor execute the corresponding iterations of parallel loops in terms of the data in its cache or local memory at execution time. It benefits, particularly, the parallel programs in which a parallel loop is enclosed by a sequential loop and array elements are repeatedly used in different iterations of the parallel loop.

name of conference

  • The Fifth International Parallel Processing Symposium

published proceedings

  • [1991] Proceedings. The Fifth International Parallel Processing Symposium

author list (cited authors)

  • Zhixi Fang, .., Mi Lu, .., & Hua Lin.

publication date

  • January 1, 1991 11:11 AM