[compiler toolkit] specify passes through config #2006
Merged
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
meta-cla
bot
added
the
CLA Signed
yiming0416
force-pushed
the
yiming/control_pass_with_config
branch
2 times, most recently
from
7a9bd50 to
1fdba5c
Compare
yiming0416
force-pushed
the
yiming/control_pass_with_config
branch
from
1fdba5c to
be49e8c
Compare
|
|
||
| **SimpleFSDP + TP + auto-bucketing** | ||
| ```shell | ||
| NGPU=8 CONFIG_FILE=./torchtitan/models/llama3/train_configs/debug_model.toml ./run_train.sh --model.name compiler_toolkit.llama3 --parallelism.data_parallel_shard_degree=2 --parallelism.tensor_parallel_degree=4 --job.custom_config_module=torchtitan.experiments.compiler_toolkit.job_config --compile.passes autobucketing_reordering |
Contributor
There was a problem hiding this comment.
I don't think they need to be exposed via command line args.
The key feature here should be passes can be specified via an API.
Contributor
Author
There was a problem hiding this comment.
@SherlockNoMad If we don't expose via command line args, how do users/developers trigger runs with different config passes?
I would imagine the use case to be the following:
- Run CLI command with config 1 (no passes)
- Run CLI command with config 2 (some passes turned on)
- compare the loss curves / profiling traces with run1 and run2 to verify the effect of passes.
SherlockNoMad
approved these changes
Nov 10, 2025
Contributor
SherlockNoMad
left a comment
SherlockNoMad
left a comment
There was a problem hiding this comment.
approve to unblock.
ahoffman-aws
pushed a commit
to drcanchi-aws/torchtitan
that referenced
this pull request
Nov 11, 2025
We should be able to control what passes to run in the compiler. This PR uses the config compile.passes to indicate in a list of graph passes to apply on the captured gm. By default, no pass is applied. Users can specify what passes to apply. Currently there are `autobucketing_reordering_pass` and `regional_inductor_pass`. ``` NGPU=8 CONFIG_FILE=./torchtitan/models/llama3/train_configs/debug_model.toml ./run_train.sh --model.name compiler_toolkit.llama3 --parallelism.data_parallel_shard_degree=2 --parallelism.tensor_parallel_degree=4 --job.custom_config_module=torchtitan.experiments.compiler_toolkit.job_config --compile.passes autobucketing_reordering,regional_inductor ``` Also updated CI to include this new config
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.
We should be able to control what passes to run in the compiler. This PR uses the config compile.passes to indicate in a list of graph passes to apply on the captured gm.
By default, no pass is applied. Users can specify what passes to apply.
Currently there are
autobucketing_reordering_passandregional_inductor_pass.Also updated CI to include this new config