POC: Define memory flags and use it for ppln protocols #10540
+359
−110
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
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Outdated suggestions cannot be applied.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.
What?
Define memory type flags to distinguish between memory of the same type but with different capabilities or properties. For example, to differentiate between legacy pinned CUDA memory and fabric VMM memory (the latter being suitable for MNNVL).
Why?
This allows selecting device memory pipeline protocols for all relevant cases (e.g., for
osu_bw H D). Currently, sending host memory to legacy pinned memory does not select a pipeline protocol because the receiver responds with a normal RTR instead of RTR mtype. As a result, the sender fails to unpack cuda_ipc key (since it is legacy pinned memory) and cannot select cuda-ipc for the transfer.How?
This PR is PoC demonstrating the use of memory flags. It will be split into multiple PRs, starting with a separate PR for UCT API changes. Thus, no need to review it