Allows workers and rsets to be set by detected GPUs. #939
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Uses set_workers_by_gpus. But consider setting rsets and honor workers if exists. Only for local comms - unless honor workers.
15 tasks
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Decide option name and type
set_workers_by_gpus.Sets resource_sets to no. of GPUs, and currently sets workers to rsets+1. But maybe we don't want to expose users to
resource_setsterminology, as intended to be simplified.But consider setting rsets and honor workers if exists. Only for local comms - unless honoring workers. rsets is more elegant as can establish basic configuration and then just have a multiplier.
Could allow the user to set up resources in their calling script, similar to the executor. This could include the configuration of a resource set. The user could also use the returned attributes in calling script if want - including
nworkers(when set by detection). If set by user, would not get re-calculated within libE. This would also allow checks before calling libE. You could specify system / code / env variables here - or get back whether a known system is detected. It may be also that you want to access some shared configuration file - and that could also be specified here (or a common location could be looked for). Would need to be added to YAML parsing - as with any additions. Then, do you set the above here, instead of libE_specs?