Scaling parallel I/O performance through I/O delegate and caching system Export

@inproceedings{citeulike:3576123, abstract = {Increasingly complex scientific applications require massive parallelism to achieve the goals of fidelity and high computational performance. Such applications periodically offload checkpointing data to file system for post-processing and program resumption. As a side effect of high degree of parallelism, I/O contention at servers doesn't allow overall performance to scale with increasing number of processors. To bridge the gap between parallel computational and I/O performance, we propose a portable MPI-IO layer where certain tasks, such as file caching, consistency control, and collective I/O optimization are delegated to a small set of compute nodes, collectively termed as I/O Delegate nodes. A collective cache design is incorporated to resolve cache coherence and hence alleviates the lock contention at I/O servers. By using popular parallel I/O benchmark and application I/O kernels, our experimental evaluation indicates considerable performance improvement with a small percentage of compute resources reserved for I/O.}, address = {Piscataway, NJ, USA}, author = {Nisar, Arifa and Liao, Wei K. and Choudhary, Alok}, booktitle = {SC '08: Proceedings of the 2008 ACM/IEEE conference on Supercomputing}, citeulike-article-id = {3576123}, citeulike-linkout-0 = {http://portal.acm.org/citation.cfm?id=1413380}, citeulike-linkout-1 = {http://dx.doi.org/10.1145/1413370.1413380}, doi = {10.1145/1413370.1413380}, isbn = {978-1-4244-2835-9}, keywords = {caching, io, mpi-io, parallel}, location = {Austin, Texas}, pages = {1--12}, posted-at = {2008-11-19 01:53:07}, priority = {2}, publisher = {IEEE Press}, title = {Scaling parallel I/O performance through I/O delegate and caching system}, url = {http://dx.doi.org/10.1145/1413370.1413380}, year = {2008} }

TY - CONF ID - citeulike:3576123 L3 - citeulike-article-id:3576123 N2 - Increasingly complex scientific applications require massive parallelism to achieve the goals of fidelity and high computational performance. Such applications periodically offload checkpointing data to file system for post-processing and program resumption. As a side effect of high degree of parallelism, I/O contention at servers doesn't allow overall performance to scale with increasing number of processors. To bridge the gap between parallel computational and I/O performance, we propose a portable MPI-IO layer where certain tasks, such as file caching, consistency control, and collective I/O optimization are delegated to a small set of compute nodes, collectively termed as I/O Delegate nodes. A collective cache design is incorporated to resolve cache coherence and hence alleviates the lock contention at I/O servers. By using popular parallel I/O benchmark and application I/O kernels, our experimental evaluation indicates considerable performance improvement with a small percentage of compute resources reserved for I/O. CY - Austin, Texas SN - 978-1-4244-2835-9 AD - Piscataway, NJ, USA EP - 12 TI - Scaling parallel I/O performance through I/O delegate and caching system PB - IEEE Press T2 - SC '08: Proceedings of the 2008 ACM/IEEE conference on Supercomputing SP - 1 KW - caching KW - io KW - mpi-io KW - parallel AU - Nisar, Arifa AU - Liao, Wei AU - Choudhary, Alok PY - 2008/// UR - http://dx.doi.org/10.1145/1413370.1413380 ER -