Declare correct sharding in MetropolisSamplerState.init #1776
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
inailuig
reviewed
Apr 22, 2024
There was a problem hiding this comment.
I think the other changes are unnecessary.
n_steps_proc was already fully replicated and is increased on every device independently, and there is no need to make it a separate variable on every device and then sum.
For avoiding the communication see #1777
|
Ah! You are right. The replication does not incur into communication. Ok. I will revert to only include that single line |
inailuig
reviewed
Apr 23, 2024
PhilipVinc
changed the title
jwnys
pushed a commit
to jwnys/netket
that referenced
this pull request
Apr 23, 2024
…etket#1776) Declare correct sharding for the counter of accepted moves inside of MetropolisSamplerState.__init__ . The previous code was working fine but gave rise to additional communications among nodes.
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Outdated suggestions cannot be applied.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.
While reviewing #1769 I realised that our sampler states do not properly initialise some variables to be sharded, resulting in some (even if small) cross-GPU communication at every sampling step.
This should fix a few issues I noticed.