[Reproducibility] Score-P and OMPT: Smoothing the bumpy road to OpenMP performance measurement

[This tarball contains reproducibility material for the paper "Score-P and OMPT: Smoothing the bumpy road to OpenMP performance measurement" which was accepted for 15th International Parallel Tools Workshop 2024]

Abstract: The OpenMP API is a widely used interface for high-level parallel pro-
gramming. With large adoption in the HPC community, OpenMP is a key compo-
nent in leveraging node-level parallelism in applications and frameworks. Herewith,
a need for OpenMP-aware performance measurement and analysis tools arose. In
version 5.0 of the OpenMP specification, the OpenMP Tools Interface (OMPT) was
introduced, providing means to collect information about the application’s use of
OpenMP directives and routines.
Although provided with a detailed specification, understanding and correctly han-
dling the CPU execution model event sequence dispatched from runtimes requires
analysis of events, their parameters and executing threads. To facilitate this, we de-
veloped a freely available OMPT tool that allows for logging events and metadata for
post-mortem inspection. Analyzing the output of this tool enabled us to implement
an OMPT tool for the performance measurement infrastructure Score-P.
As vendors interpret the OMPT specification, various checks were developed to
detect deviations and work around them. Accompanying the development of OMPT
itself, the overhead being introduced was always a concern. To assess this overhead
in various contemporary runtimes, we used the EPCC OpenMP MicroBenchmark
Suite and the SPEChpc 2021 Benchmark Suites, without an OMPT tool, with a
NOOP tool, and with the Score-P OMPT tool attached.