Score-P and OMPT: Navigating the perils of callback-driven parallel runtime introspection

6Citations
Citations of this article
5Readers
Mendeley users who have this article in their library.
Get full text

Abstract

Event-based performance analysis aims at modeling the behavior of parallel applications through a series of state transitions during execution. Different approaches to obtain such transition points for OpenMP programs include source-level instrumentation (e.g., OPARI) and callback-driven runtime support (e.g., OMPT). In this paper, we revisit a previous evaluation and comparison of OPARI and an LLVM OMPT implementation—now updated to the OpenMP 5.0 specification—in the context of Score-P. We describe the challenges faced while trying to use OMPT as a drop-in replacement for the existing instrumentation-based approach and the changes in event order that could not be avoided. Furthermore, we provide details on Score-P measurements using OPARI and OMPT as event sources with the EPCC and SPEC OpenMP benchmark suites.

Cite

CITATION STYLE

APA

Feld, C., Convent, S., Hermanns, M. A., Protze, J., Geimer, M., & Mohr, B. (2019). Score-P and OMPT: Navigating the perils of callback-driven parallel runtime introspection. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 11718 LNCS, pp. 21–35). Springer Verlag. https://doi.org/10.1007/978-3-030-28596-8_2

Register to see more suggestions

Mendeley helps you to discover research relevant for your work.

Already have an account?

Save time finding and organizing research with Mendeley

Sign up for free