-
Notifications
You must be signed in to change notification settings - Fork 217
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Allow term change in MCccfraft.cfg and disable abs model refinement #6498
Conversation
GitHub actions seems to have lost the full log after the timeout, but the result of this experiment is that Mccfraft becomes impractically slow, at least for CI purposes: To clarify, this is for 483b547 |
While 3 nodes with
|
+1 for creating a configuration with 1 configuration, 2 nodes, and one term change. Zooming out, here are dimensions I can think off by by we can slice and dice the space space:
|
@lemmy 100%, that is what I was contemplating. We won't quite be able to do everything, so I think we need to pick what's "interesting" and do a bit of trial and error to work out which of those run in a sensible amount of time, and how many we can run concurrently before the cost of a CI run starts to become interesting itself. Edit: is there a way to avoid the proliferation of 99%-identical configs are we do this? |
Going round robin over a collection of models in different CI runs or a scheduled (e.g. weekend) run of bigger models would provide a good compromise.
There is nothing in TLC right now, but I will look into it. |
Replaced by #6504 |
Experiment following discussion in #6493 (comment).
Tried to run locally, but this seems very slow either way, so offloading to the CI. On runs lasting a few minutes, it looks like refinement lowers the state exploration rate from ~1.2m to ~800k.
Goal: find out if it is practical to raise the term limit in MCccfraft.tla.