2.6. Benchmark Utility¶
What does the benchmark utility do?¶
The benchmark utility uses the `submit`_ utility to submit multiple HPC jobs to a cluster. Please read the `submit`_ guide to understand how to submit a job using `submit`_.
After understanding the `submit`_, creating multiple simulations using
Benchmark is as simple as creating a file named job.csv,
which is a comma
separated values file where the headers of each column are parameters for
`submit`_ and each row contains the specific values unique to each
run.
An example csv:
num_cores, solver_type, solver_version, job_name
24, mpp_lsdyna, r712sp, test_dyna_r712sp_24
48, mpp_lsdyna, r712sp, test_dyna_r712sp_48
24, mpp_lsdyna, r8sp, test_dyna_r8sp_24
48, mpp_lsdyna, r8sp, test_dyna_r8sp_48
The above csv will use an existing job.json
file containing all of the
parameters required to run `submit`_ and will change the number of
cores and solver version for each simulation. Creating four different
simulations two for version r712sp and two for r8sp, each version will have a
24 core job and a 48 core job.