Contribution to a conference proceedings/Contribution to a book

FZJ-2019-05220

http://join2-wiki.gsi.de/foswiki/pub/Main/Artwork/join2_logo100x88.png Score-P and OMPT: Navigating the Perils of Callback-Driven Parallel Runtime Introspection

Feld, C. (Corresponding author)FZJ* ; Convent, S. ; Hermanns, M.-A.FZJ* ; Protze, J. ; Geimer, M.FZJ* ; Mohr, B.FZJ*

2019
Springer International Publishing Cham
ISBN: 978-3-030-28595-1 (print), 978-3-030-28596-8 (electronic)

This record in other databases:  

Please use a persistent id in citations: http://hdl.handle.net/2128/23232  doi:10.1007/978-3-030-28596-8_2

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.

---

Contributing Institute(s):

1. Jülich Supercomputing Center (JSC)

Research Program(s):

1. 511 - Computational Science and Mathematical Methods (POF3-511) (POF3-511)
2. Scalable Performance Analysis of Large-Scale Parallel Applications (jzam11_20191101) (jzam11_20191101)
3. ATMLPP - ATML Parallel Performance (ATMLPP) (ATMLPP)

Appears in the scientific report 2019

---

Database coverage:
; Nationallizenz ; SCOPUS

---