-
Notifications
You must be signed in to change notification settings - Fork 11
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add several minor fixes #32
Conversation
MC952-arch
commented
Feb 7, 2025
- Invoke event block call in WorkFlagCX::wait func
- Add plugin/torch/run_hetero.sh script for heterogeneous test
- Reuse sync stream as a substitution of barrier op for better performance
plugin/torch/example.py
Outdated
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Please refactor this file in a more modular way in the future.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
OK
if (!coalesced_) | ||
{ | ||
event_->block(device_id_); | ||
} | ||
if (isBarrierOp_) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
What is the reason for this code working now?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
What is the reason for this code working now?
It seems that a sync stream call is required after flagcxHeteroSend/Recv ops. Such synchronization is added in all C2C algorithms for the moment. Here we adopt the same way as ProcessGroupNCCL does, blocking current stream on the flagcx stream.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Something will be Improved in the future.