[batch] fix g2 machine memory calculation #14498
Merged
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
daniel-goldstein
suggested changes
Apr 26, 2024
There was a problem hiding this comment.
Leaving my comments from our discussion yesterday. This looks great and I'm really happy we caught this bug. Just a few comments on the organization.
It would also be great to have some unit testing around the problematic function: gcp_worker_memory_per_core_mib. Take a look at batch/test/test_utils.py, it contains a couple unit tests that just test the behavior of a couple of individual functions. You can run that test file with pytest batch/test/test_utils.py. This would be a good place to add tests for gcp_worker_memory_per_core_mib. We should cover at least a handful of n1s and g2s and some invalid / not-yet-supported machine types that should raise some sort of error.
added 3 commits
May 1, 2024 15:16
daniel-goldstein
suggested changes
May 2, 2024
added 2 commits
May 2, 2024 21:16
daniel-goldstein
approved these changes
May 3, 2024
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Outdated suggestions cannot be applied.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.
The memory that machines use is calculated based off of the number of cores the machine has, with a fix ratio set of cores to Gb of RAM. The code assumed a ratio of 3.75 cores: 1 Gb but this assumption does not hold outside of the n1 machine family. This PR changes the functions that calculate memory from number of cores to account for this by passing the machine_type as the main parameter rather than the worker type. Then, logic is added to use the appropriate ratio of 4:1 for the g2 family of machines.