Difference between revisions of "Scalasca Trace Tools"
m (Add software infobox) |
m |
||
| (2 intermediate revisions by one other user not shown) | |||
| Line 1: | Line 1: | ||
| − | [[Category:HPC-Developer]] | + | [[Category:HPC-Developer]]<nowiki /> |
| + | [[Category:Benchmarking]]<nowiki /> | ||
| + | |||
{{Infobox Software | {{Infobox Software | ||
| name = Scalasca Trace Tools | | name = Scalasca Trace Tools | ||
| Line 10: | Line 12: | ||
| lang = C, C++, Fortran | | lang = C, C++, Fortran | ||
| arch = x86, Power, ARM | | arch = x86, Power, ARM | ||
| − | }} | + | }}<nowiki /> |
Scalasca is a trace-based performance analysis tool for parallel programs targeted at large-scale high-performance computing platforms, but it can be also used for small and medium-sized clusters. | Scalasca is a trace-based performance analysis tool for parallel programs targeted at large-scale high-performance computing platforms, but it can be also used for small and medium-sized clusters. | ||
Latest revision as of 12:27, 19 July 2024
| Software | |
|---|---|
| Name: | Scalasca Trace Tools
|
| Provider: | Forschungszentrum Jülich |
| License: | BSD-style |
| Homepage: | https://scalasca.org/ |
| E-Mail: | scalasca@fz-juelich.de |
| Concurrency: | parallel |
| Programming Model: | MPI, OpenMP, Pthreads |
| Programming Language: | C, C++, Fortran |
| Architecture: | x86, Power, ARM |
Scalasca is a trace-based performance analysis tool for parallel programs targeted at large-scale high-performance computing platforms, but it can be also used for small and medium-sized clusters. It focuses on identifying waiting time in communication and synchronization mechanisms including their sources in parallel applications. It is available under the New BSD open-source license.
Since version 2.0 is relies on the Score-P instrumentation and measurement framework to instrument applications and provide an OTF2 event trace. To retain scalability during the analysis of highly parallel applications, Scalasca performs its parallel analysis, using the same level of parallelism as was used for the measurement. The analysis report is provided as a Cube performance report.
It supports the analysis of applications using the Message Passing Interface (MPI), OpenMP, and POSIX threads. For further information please visit the official Scalasca website.