Using Dynamic Tracing Sampling to Measure Long Running Programs Export

@inproceedings{citeulike:6102430, abstract = {Detailed cache simulation can be useful to both system developers and application writers to understand an application's performance. However, measuring long running programs can be extremely slow. In this paper we present a technique to use dynamic sampling of trace snippets throughout an application's execution. We demonstrate that our approach improves accuracy compared to sampling a few timesteps at the beginning of execution by judiciously choosing the frequency, as well as the points in the control flow, at which samples are collected. Our approach is validated using the SIGMA tracing and simulation framework for the IBM Power family of processors.}, address = {Washington, DC, USA}, author = {Odom, Jeffrey and Hollingsworth, Jeffrey K. and DeRose, Luiz and Ekanadham, Kattamuri and Sbaraglia, Simone}, booktitle = {SC '05: Proceedings of the 2005 ACM/IEEE conference on Supercomputing}, citeulike-article-id = {6102430}, citeulike-linkout-0 = {http://portal.acm.org/citation.cfm?id=1105760.1105824}, citeulike-linkout-1 = {http://dx.doi.org/10.1109/SC.2005.77}, doi = {10.1109/SC.2005.77}, isbn = {1-59593-061-2}, keywords = {high-performance\_computing, performance\_analysis}, pages = {59+}, posted-at = {2009-11-12 17:09:37}, priority = {2}, publisher = {IEEE Computer Society}, title = {Using Dynamic Tracing Sampling to Measure Long Running Programs}, url = {http://dx.doi.org/10.1109/SC.2005.77}, year = {2005} }

TY - CONF ID - citeulike:6102430 L3 - citeulike-article-id:6102430 N2 - Detailed cache simulation can be useful to both system developers and application writers to understand an application's performance. However, measuring long running programs can be extremely slow. In this paper we present a technique to use dynamic sampling of trace snippets throughout an application's execution. We demonstrate that our approach improves accuracy compared to sampling a few timesteps at the beginning of execution by judiciously choosing the frequency, as well as the points in the control flow, at which samples are collected. Our approach is validated using the SIGMA tracing and simulation framework for the IBM Power family of processors. SN - 1-59593-061-2 AD - Washington, DC, USA TI - Using Dynamic Tracing Sampling to Measure Long Running Programs PB - IEEE Computer Society T2 - SC '05: Proceedings of the 2005 ACM/IEEE conference on Supercomputing SP - 59 KW - high-performance_computing KW - performance_analysis AU - Odom, Jeffrey AU - Hollingsworth, Jeffrey AU - DeRose, Luiz AU - Ekanadham, Kattamuri AU - Sbaraglia, Simone PY - 2005/// UR - http://dx.doi.org/10.1109/SC.2005.77 ER -