[HIP] Implement workaround for hipMemset2D #1395
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
Codecov ReportAll modified and coverable lines are covered by tests ✅
❗ Your organization needs to install the Codecov GitHub app to enable full functionality. Additional details and impacted files@@ Coverage Diff @@
## main #1395 +/- ##
==========================================
- Coverage 14.82% 12.51% -2.32%
==========================================
Files 250 239 -11
Lines 36220 35949 -271
Branches 4094 4076 -18
==========================================
- Hits 5369 4498 -871
- Misses 30800 31447 +647
+ Partials 51 4 -47 ☔ View full report in Codecov by Sentry. |
GeorgeWeb
reviewed
Feb 29, 2024
konradkusiak97
force-pushed
the
improvedQueueFill
branch
2 times, most recently
from
e400d8f
to
7a05c32
Compare
GeorgeWeb
approved these changes
Mar 5, 2024
konradkusiak97
force-pushed
the
improvedQueueFill
branch
from
e6dfc62
to
f277422
Compare
sarnex
pushed a commit
to intel/llvm
that referenced
this pull request
Apr 12, 2024
sarnex
pushed a commit
to intel/llvm
that referenced
this pull request
May 2, 2024
This PR changes the `queue.fill()` implementation to make use of the native functions for a specific backend. It also unifies that implementation with the one for memset, since it is just an 8-bit subset operation of fill. In the CUDA case, both memset and fill are currently calling `urEnqueueUSMFill` which depending on the size of the filling pattern calls either `cuMemsetD8Async`, `cuMemsetD16Async`, `cuMemsetD32Async` or `commonMemSetLargePattern`. Before this patch memset was using the same thing, just beforehand setting patternSize always to 1 byte which resulted in calling `cuMemsetD8Async`. In other backends, the behaviour is analogous. The fill method was just invoking a `parallel_for` to fill the memory with the pattern which was making this operation quite slow. This PR depends on: - oneapi-src/unified-runtime#1395 - oneapi-src/unified-runtime#1412
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.
There is an issue with
hipMemset2Din ROCm prior to 6.0.0 version and this PR adds a workaround for it incommonMemSetLargePattern.The issue appears only when using a pointer to host pinned memory from
hipHostMalloc. I believe such a case hasn't been used until trying to refactor the USM fill (intel/llvm#12702).Testing intel/llvm: intel/llvm#12898