This title appears in the Scientific Report :
2017
Please use the identifier:
http://dx.doi.org/10.1109/LDAV.2016.7874340 in citations.
Correlating sub-phenomena in performance data in the frequency domain
Correlating sub-phenomena in performance data in the frequency domain
Finding and understanding correlated performance behaviour of the individual functions of massively parallel high-performance computing (HPC) applications is a time-consuming task. In this poster, we propose filtered correlation analysis for automatically locating interdependencies in call-path perf...
Saved in:
Personal Name(s): | Vierjahn, Tom (Corresponding author) |
---|---|
Hermanns, Marc-Andre / Mohr, Bernd / Muller, Matthias S. / Kuhlen, Torsten W. / Hentschel, Bernd | |
Contributing Institute: |
JARA - HPC; JARA-HPC Jülich Supercomputing Center; JSC |
Imprint: |
IEEE
2016
|
Physical Description: |
105-106 |
ISBN: |
978-1-5090-5659-0 |
DOI: |
10.1109/LDAV.2016.7874340 |
Conference: | 2016 IEEE 6th Symposium on Large Data Analysis and Visualization, Baltimore, MD (USA), 2016-10-23 - 2016-10-28 |
Document Type: |
Contribution to a book Contribution to a conference proceedings |
Research Program: |
Computational Science and Mathematical Methods |
Publikationsportal JuSER |
Finding and understanding correlated performance behaviour of the individual functions of massively parallel high-performance computing (HPC) applications is a time-consuming task. In this poster, we propose filtered correlation analysis for automatically locating interdependencies in call-path performance profiles. Transforming the data into the frequency domain splits a performance phenomenon into sub-phenomena to be correlated |