DMCA
Cautious, machine-independent performance tuning for shared-memory multiprocessors (1996)
Cached
Download Links
| Venue: | Euro-Par 96 - Second European Conference on Parallel Processing |
| Citations: | 1 - 1 self |
BibTeX
@INPROCEEDINGS{Talbot96cautious,machine-independent,
author = {Sarah A. M. Talbot and Andrew J. Bennett and Paul H. J. Kelly},
title = {Cautious, machine-independent performance tuning for shared-memory multiprocessors},
booktitle = {Euro-Par 96 - Second European Conference on Parallel Processing},
year = {1996},
pages = {106--113},
publisher = {Springer-Verlag}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
Abstract. Coherent-cache shared-memory architectures often give disappointing performance which can be alleviated by manual tuning. We describe a new trace analysis tool, clarissa, which helps diagnose problems and pinpoint their causes. Unusually, clarissa works by analysing potential contention, instead of measuring predicted contention by simulating a speci c memory system design. This is important because, after tuning, the software will be executed on di erent inputs and di erent con gurations. The goal is to produce a program with robustly good performance. This paper explains the principle behind cautious trace analysis, describes our implementation, and presents our experience of using the tool. 1
Keyphrases
shared-memory multiprocessor machine-independent performance good performance di erent input coherent-cache shared-memory architecture potential contention disappointing performance manual tuning new trace analysis tool speci memory system design di erent con gurations clarissa work cautious trace analysis
