A multithreaded communication engine for multicore architectures Export

@inproceedings{citeulike:3255762, abstract = {The current trend in clusters leads towards an increase of the number of cores per node. As a result, an increasing number of parallel applications is mixing message passing and multithreading as an attempt to better match the underlying architecture's structure. This naturally raises the problem of designing efficient, multithreaded implementations of MPL In this paper, we present the design of a multithreaded communication engine able to exploit idle cores to speed up communications in two ways: it can move CPU- intensive operations out of the critical path (e.g. PIO transfers off load), and is able to let rendezvous transfers progress asynchronously. We have implemented these methods in the PM2 software suite, evaluated their behavior in typical cases, and we have observed good performance results in overlapping communication and computation.}, author = {Trahay, F. and Brunet, E. and Denis, A. and Namyst, R.}, booktitle = {Parallel and Distributed Processing, 2008. IPDPS 2008. IEEE International Symposium on}, citeulike-article-id = {3255762}, citeulike-linkout-0 = {http://dx.doi.org/10.1109/IPDPS.2008.4536139}, citeulike-linkout-1 = {http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=4536139}, doi = {10.1109/IPDPS.2008.4536139}, journal = {Parallel and Distributed Processing, 2008. IPDPS 2008. IEEE International Symposium on}, keywords = {multi-core, multi-thread, scheduling}, pages = {1--7}, posted-at = {2008-09-15 03:41:22}, priority = {2}, title = {A multithreaded communication engine for multicore architectures}, url = {http://dx.doi.org/10.1109/IPDPS.2008.4536139}, year = {2008} }

TY - CONF ID - citeulike:3255762 L3 - citeulike-article-id:3255762 N2 - The current trend in clusters leads towards an increase of the number of cores per node. As a result, an increasing number of parallel applications is mixing message passing and multithreading as an attempt to better match the underlying architecture's structure. This naturally raises the problem of designing efficient, multithreaded implementations of MPL In this paper, we present the design of a multithreaded communication engine able to exploit idle cores to speed up communications in two ways: it can move CPU- intensive operations out of the critical path (e.g. PIO transfers off load), and is able to let rendezvous transfers progress asynchronously. We have implemented these methods in the PM2 software suite, evaluated their behavior in typical cases, and we have observed good performance results in overlapping communication and computation. JF - Parallel and Distributed Processing, 2008. IPDPS 2008. IEEE International Symposium on EP - 7 TI - A multithreaded communication engine for multicore architectures T2 - Parallel and Distributed Processing, 2008. IPDPS 2008. IEEE International Symposium on SP - 1 KW - multi-core KW - multi-thread KW - scheduling AU - Trahay, F AU - Brunet, E AU - Denis, A AU - Namyst, R PY - 2008/// UR - http://dx.doi.org/10.1109/IPDPS.2008.4536139 ER -