
The upcoming OpenMP tools interface (OMPT) has been designed as a portable interface for performance analysis tools. It provides access to OpenMP-related information at program runtime and can thus extend the analysis capabilities of current performance tools. This paper compares the functionality and convenience of OMPT with OPARI2 for event-based performance analysis. For this purpose, we integrated OMPT into the measurement infrastructure Score-P, which previously accessed OpenMP-related information using only source-level instrumentation with OPARI2. For comparison, we performed Score-P measurements of the NAS Parallel Benchmark suite and the LULESH code with OPARI2 instrumentation and with OMPT. In each case, we determined the overhead and evaluated the output. We found that the measurement overhead is dominated by the measurement system, while the contribution of the event source remains negligible. Moreover, OMPT and OPARI2 provide complementary views of the performance behavior. Whereas OPARI2 maintains a strictly source-code-centric perspective that reflects OpenMP standard abstractions, OMPT mirrors the behavior of the OpenMP runtime and exposes compiler optimizations.
| selected citations These citations are derived from selected sources. This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically). | 8 | |
| popularity This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network. | Average | |
| influence This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically). | Top 10% | |
| impulse This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network. | Average |
