Issue No. 04 - April (2001 vol. 12)
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/71.920589
<p><b>Abstract</b>—In many real applications, for example, those with frequent and irregular communication patterns or those using large messages, network contention and contention for message processing resources can be a significant part of the total execution time. This paper presents a new cost model, called LoGPC, that extends the LogP [<ref rid="bibL04049" type="bib">9</ref>] and LogGP [<ref rid="bibL04044" type="bib">4</ref>] models to account for the impact of network contention and network interface DMA behavior on the performance of message passing programs. We validate LoGPC by analyzing three applications implemented with Active Messages [<ref rid="bibL040411" type="bib">11</ref>], [<ref rid="bibL040419" type="bib">19</ref>] on the MIT Alewife multiprocessor. Our analysis shows that network contention accounts for up to 50 percent of the total execution time. In addition, we show that the impact of communication locality on the communication costs is at most a factor of two on Alewife. Finally, we use the model to identify trade-offs between synchronous and asynchronous message passing styles.</p>
Multiprocessors, modeling, pipelining, contention, network.
M. I. Frank and C. A. Moritz, "LoGPC: Modeling Network Contention in Message-Passing Programs," in IEEE Transactions on Parallel & Distributed Systems, vol. 12, no. , pp. 404-415, 2001.