On-line automated performance diagnosis on thousands of processes
Proceedings of the eleventh ACM SIGPLAN symposium on Principles and practice of parallel programming
ompP: a profiling tool for OpenMP
IWOMP'05/IWOMP'06 Proceedings of the 2005 and 2006 international conference on OpenMP shared memory parallel programming
Performance analysis of shared-memory parallel applications using performance properties
HPCC'05 Proceedings of the First international conference on High Performance Computing and Communications
Boosting Application-Specific Parallel I/O Optimization Using IOSIG
CCGRID '12 Proceedings of the 2012 12th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing (ccgrid 2012)
Hi-index | 0.00 |
Automatic performance tools must of course be testedas to whether they perform their task correctly. Becauseperformance tools are meta-programs, tool testing is morecomplex than ordinary program testing and comprises atleast three aspects. First, it must be ensured that the tools doneither alter the semantics nor distort the run-time behaviorof the application under investigation. Next, it must be verified that the tools collect the correct performance data asrequired by their specification. Finally, it must be checkedthat the tools indeed perform their intended tasks and detectrelevant performance problems. Focusing on the latter (correctness) aspect, testing can be done using synthetic testfunctions with controllable performance properties, and/orreal world applications with known performance behavior.A systematic test suite can be built from synthetic test functions and other components, possibly with the help of toolsto assist the user in putting the pieces together into executable test programs. Clearly, such a test suite can behighly useful to builders of performance analysis tools. Itis surprising that up till now, no systematic effort has beenundertaken to provide such a suite.In this paper we discuss the initial design of a test suitefor checking the correctness (in the above sense) of automatic performance analysis tools. In particular, we describe a collection of synthetic test functions which allowsto easily construct both simple and more complex test programs with desired performance properties.