Fallback to standard mesh on neuron backend for incompatible multi-granule meshes #1146
Conversation
This file contains hidden or 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
|
Rebased the PR |
ruomingp
approved these changes
May 21, 2025
axlearn/common/utils.py
Outdated
| if ( | ||
| device_platform == "gpu" | ||
| device_platform in ("gpu", "neuron") |
There was a problem hiding this comment.
Suggested change
| device_platform in ("gpu", "neuron") | |
| device_platform != "tpu" |
There was a problem hiding this comment.
Made this change, thank you!
axlearn/common/utils.py
Outdated
| @@ -1743,13 +1743,15 @@ def create_device_mesh( | |||
| assert num_devices % num_granules == 0, "Number of devices should divide number of granules." | |||
| num_devices_per_granule = num_devices // num_granules | |||
|
|
|||
| # Fallback to a standard mesh if on GPU with incompatible multi-granule mesh. | |||
| # Fallback to a standard mesh if on GPU or neuron with incompatible multi-granule mesh. | |||
There was a problem hiding this comment.
Suggested change
| # Fallback to a standard mesh if on GPU or neuron with incompatible multi-granule mesh. | |
| # Fallback to a standard mesh with incompatible multi-granule mesh if not on TPU. |
There was a problem hiding this comment.
Made this change, thank you!
- Switch to a new mesh for neuron-(trn2|trn2n).48xlarge-64 with better scale-out performance.
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Outdated suggestions cannot be applied.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.
Allow fallback to standard mesh for multi-granule mesh as such a mesh provides better performance on TRN2
neuron-(trn2|trn2n).48xlarge-64on70B-FujiV2as it provides better scale-out performance.