[Auto para] Relaunch with auto mapping function #37326
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
… auto_para_mapping
|
Thanks for your contribution! |
… auto_para_graph
… auto_para_launch
… auto_para_launch
… auto_para_launch
aoyulong
force-pushed
the
auto_para_launch
branch
from
01b8b9c to
8d4199c
Compare
kuizhiqing
reviewed
Dec 1, 2021
Comment on lines
+306
to
+307
| original_args = sys.argv[1:] | ||
| os.environ["PADDLE_ORIGINAL_CMD_ARGS"] = " ".join(original_args) |
There was a problem hiding this comment.
This has been dealt with by shlex.split on line 154 of the above parallelizer.py.
Comment on lines
+169
to
+179
| "--cluster_topo_path", | ||
| type=str, | ||
| default=None, | ||
| help="A json format file will be stored in this path which is used" | ||
| "to represent the cluster topology information for auto parallel.") | ||
| collective_group.add_argument( | ||
| "--rank_mapping_path", | ||
| type=str, | ||
| default=None, | ||
| help="A json format file will be stored in this path which is used" | ||
| "to map processes to machines for auto parallel.") |
There was a problem hiding this comment.
Add one more config file is expensive, is it possible to use xxx_config to hold all ? may be paddle_config and you hold some sections ?
There was a problem hiding this comment.
The rank_mapping file will be automatically generated by our framework in the pre-launch analysis pass and must not be exposed to users.
… auto_para_launch
… auto_para_launch
… auto_para_launch
XieYunshen
approved these changes
Dec 7, 2021
fuyinno4
approved these changes
Dec 7, 2021
fuyinno4
approved these changes
Dec 7, 2021
Zjq9409
pushed a commit
to Zjq9409/Paddle
that referenced
this pull request
Dec 10, 2021
* [Auto Parallel] Add the unified cluster representation * [Auto Parallel] Add the graph class for physical mapping * [Auto Parallel] Add the simple physical mapper * Set the timeout of the mapper * Merge the upstream develop unittests cmake files * Fix a bug of the process group * Remove mapper unittest from platforms which is not GPU * Move the instantiation of process group after resharding * Add the local id for devices * Update the rank mapping format * [Auto Parallel] Relaunch with the rank mapping file * Remove the unnecessary json file * Avoid entering get_device_proc_info for auto mapping * Correct the mapper unit test * Add some comments * Remove the related files about mapping * Update the unittest for auto mapping * Remove unused rank_mapping unittest * Improve the unittest coverage * Improve the unittest coverage * Improve the unittest of relaunch * Fix the unittest problem in CI * Improve the unittest of relaunch * Remove unnecessary statements * Update the unittest cmakefile * Correct the cmakefile of auto parallel unittests * Modify codes based on the new elastic change * Use the GPUs exclusively in the unittest * Correct the cmakefile * Set the timeout of the unittest
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.
PR types
New features
PR changes
Others
Describe
The pr will relaunch the distributed training based on the rank mapping file produced by the auto mapping function. So there will be two times of launching.