Operating the OpenPiton regression

Step one in working the OpenPiton regression is to comply with the steps outlined within the github repo and full the method efficiently. 

The following step is to obtain the design and verification recordsdata. Directions are offered within the github repo. As soon as downloaded, there are three easy setup duties to carry out:

  1. Arrange the PITON_ROOT setting variable (%export PITON_ROOT=<location of root of OpenPiton extracted recordsdata>) 

  2. Arrange the simulator dwelling (%export ICARUS_HOME=/usr). The scripts offered to you within the github repo already maintain putting in Icarus on the machines provisioned. This exhibits yet one more benefit of cloud: simplified machine configuration.

  3. Lastly, supply your required settings (%supply $PITON_ROOT/piton/piton_settings.bash)

For the verification run, we used the one tile setup for OpenPiton, the regression script ‘sims’ offered within the OpenPiton bundle and the ‘tile1_mini’ regression. We tried two runs—sequential and parallel. The parallel runs have been managed by SLURM.

We invoked the sequential run utilizing the next command:

%sims -sim_type=icv -group=tile1_mini

And the distributed run utilizing this command:

%sims -sim_type=icv -group=tile1_mini -slurm -sim_q_command=sbatch

Outcomes 

The ‘tile1_mini’ regression has 46 assessments. Operating all 46 tile1_mini assessments sequentially took a mean of 120 minutes. The parallel run for tile1_mini with 10 auto-scaled SLURM nodes accomplished in 21 minutes—a 6X enchancment!



Leave a Reply

Your email address will not be published. Required fields are marked *