Running Benchmarks¶
This page gives some examples on how to use the benchmark tool.
To get started, install the benchmark tool as described here, call btool init
and copy (or symlink) the two executables clasp-3.4.0 and runlim
into the ./programs folder.
This example assumes that you want to run a benchmark that shall be started using simple
bash scripts.
After finishing the preparation described above, run:
$ btool gen ./runscripts/runscript-seq.xml
This creates a set of start scripts in the ./output folder.
To start the benchmark, run:
$ ./output/clasp-big/houat/start.py
Once the benchmark is finished, run:
$ btool eval ./runscripts/runscript-seq.xml | btool conv -o result.xlsx
Finally, open the file in your favourite spreadsheet tool:
$ xdg-open result.xlsx
This example assumes that you want to run a benchmark on a cluster.
After finishing the preparation described above, run:
$ btool gen ./runscripts/runscript-dist.xml
This creates a set of start scripts in the ./output folder.
To start the benchmark, run (on the cluster):
$ ./output/clasp-one-as/hpc/start.sh
Once the benchmark is finished, run:
$ btool eval ./runscripts/runscript-dist.xml | btool conv -o result.xlsx
Finally, open the file in your favourite spreadsheet tool:
$ xdg-open result.xlsx