[amdgpu] D2D memcpy via streams and HSA #69977
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
jhuber6
reviewed
Oct 23, 2023
It does, accessed via getAgent. In order to pass the DeviceTy I'd have to reorder a bunch of classes in the file or otherwise deal with the type being incomplete at the point it is used. Since this function is local to amdgpu (cuda doesn't have it), passing the agent seems better than the reordering mess plus immediately calling getAgent in the prologue. |
|
Simpler version is at #69977, this is a response to review comments asking to implement it on the streams api instead |
jdoerfert
approved these changes
Oct 23, 2023
| AMDGPUStreamTy *Stream = nullptr; | ||
| if (auto Err = getStream(AsyncInfoWrapper, Stream)) | ||
| return Err; | ||
| if (Size < 0) |
There was a problem hiding this comment.
Ha, yes. Done
JonChesterfield
force-pushed
the
jc_d2d_memcpy_stream_trunk
branch
from
c6e16d8 to
f0ace3c
Compare
jhuber6
approved these changes
Oct 23, 2023
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Outdated suggestions cannot be applied.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.
hsa_amd_memory_async_copy can handle device to device copies if passed the corresponding parameters.
No functional change - currently D2D copy goes through a fallback in libomptarget that stages through a host malloc, after this it goes directly through HSA.
Works under exactly the situations that HSA works. Verified locally on a performance benchmark. Hoping to attract further testing from internal developers after it lands.