Randomized Cache Placement for Eliminating Conflicts

Nigel Topham, Antonio González

Research output: Contribution to journalArticlepeer-review

Abstract

Applications with regular patterns of memory access can experience high levels of cache conflict misses. In shared-memory multiprocessors conflict misses can be increased significantly by the data transpositions required for parallelization. Techniques such as blocking which are introduced within a single thread to improve locality, can result in yet more conflict misses. The tension between minimizing cache conflicts and the other transformations needed for efficientparallelization leads to complex optimization problems for parallelizing compilers. This paper shows how the introduction of a pseudorandom element into the cache index function can effectively eliminate repetitive conflict misses and produce a cache where miss ratio depends solely on working set behavior. We examine the impact of pseudorandom cache indexing on processor cycle times and present practical solutions to some of the major implementation issues for this type of cache. Our conclusions are supported by simulations of a superscalar out-of-order processor executing the SPEC95 benchmarks, as well as from cache simulations of individual loop kernels to illustrate specific effects. We present measurements of Instructions committed Per Cycle (IPC) when comparing the performance of different cache architectures on whole-program benchmarks such as the SPEC95 suite.
Original languageEnglish
Pages (from-to)185-192
Number of pages8
JournalIEEE Transactions on Computers
Volume48
Issue number2
DOIs
Publication statusPublished - 1 Feb 1999

Keywords

  • Conflict avoidance, cache architectures, performance evaluation.

Fingerprint Dive into the research topics of 'Randomized Cache Placement for Eliminating Conflicts'. Together they form a unique fingerprint.

Cite this