| 1 |
Source: jube |
| 2 |
Section: science |
| 3 |
Priority: optional |
| 4 |
Maintainer: Debian HPC Team <[email protected]> |
| 5 |
Uploaders: |
| 6 |
Andreas Beckmann <[email protected]>, |
| 7 |
Build-Depends: |
| 8 |
debhelper-compat (= 13), |
| 9 |
dh-python, |
| 10 |
python3-all, |
| 11 |
python3-setuptools, |
| 12 |
Standards-Version: 4.7.2 |
| 13 |
Rules-Requires-Root: no |
| 14 |
Homepage: https://www.fz-juelich.de/jsc/jube |
| 15 |
Vcs-Browser: https://salsa.debian.org/hpc-team/jube |
| 16 |
Vcs-Git: https://salsa.debian.org/hpc-team/jube.git |
| 17 |
|
| 18 |
Package: jube |
| 19 |
Architecture: all |
| 20 |
Depends: |
| 21 |
${misc:Depends}, |
| 22 |
${python3:Depends}, |
| 23 |
Description: JUBE Benchmarking Environment |
| 24 |
Automating benchmarks is important for reproducibility and |
| 25 |
hence comparability which is the major intent when |
| 26 |
performing benchmarks. Furthermore managing different |
| 27 |
combinations of parameters is error-prone and often |
| 28 |
results in significant amounts work especially if the |
| 29 |
parameter space gets large. |
| 30 |
. |
| 31 |
In order to alleviate these problems JUBE helps performing |
| 32 |
and analyzing benchmarks in a systematic way. It allows |
| 33 |
custom work flows to be able to adapt to new architectures. |
| 34 |
. |
| 35 |
For each benchmark application the benchmark data is written |
| 36 |
out in a certain format that enables JUBE to deduct the |
| 37 |
desired information. This data can be parsed by automatic |
| 38 |
pre- and post-processing scripts that draw information, |
| 39 |
and store it more densely for manual interpretation. |
| 40 |
. |
| 41 |
The JUBE benchmarking environment provides a script based |
| 42 |
framework to easily create benchmark sets, run those sets |
| 43 |
on different computer systems and evaluate the results. It |
| 44 |
is actively developed by the Juelich Supercomputing Centre |
| 45 |
of Forschungszentrum Juelich, Germany. |