Refactor/autotune/key #924
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
burn-wgpu/src/compute/tune_key.rs
Outdated
Comment on lines
11
to
15
| Matmul { | ||
| /// The key with specific matmul information | ||
| matmul_key: MatmulAutotuneKey, | ||
| }, | ||
| } |
There was a problem hiding this comment.
The name is not necessary:
WgpuAutotuneKey {
Matmul(MatmulAutotuneKey),
}
burn-compute/src/tune/tuner.rs
Outdated
| pub struct Tuner<S, C> { | ||
| tune_cache: TuneCache<S>, | ||
| pub struct Tuner<S: ComputeServer, C> { | ||
| tune_cache: TuneCache<S::AutotuneKey>, | ||
| _server: PhantomData<S>, |
There was a problem hiding this comment.
I don't think this phantom data is necessary.
Comment on lines
7
to
13
| pub struct MatmulAutotuneKey { | ||
| round: bool, // True when all matmul dims are multiples of 64 | ||
| broadcast: bool, // True when there are differences in batch size | ||
| anchored_m: usize, | ||
| anchored_k: usize, | ||
| anchored_n: usize, | ||
| } |
There was a problem hiding this comment.
I think we wanted to include the round batch size as well !
There was a problem hiding this comment.
I'm not sure if you mean to have a bool for if batch sizes are all multiples of 64, or have an anchored batch size as usize (for all batch dims?)
There was a problem hiding this comment.
anchored batch size for all batch dims!
Codecov ReportAttention:
Additional details and impacted files@@ Coverage Diff @@
## main #924 +/- ##
==========================================
- Coverage 86.54% 86.52% -0.03%
==========================================
Files 469 471 +2
Lines 44191 44238 +47
==========================================
+ Hits 38244 38275 +31
- Misses 5947 5963 +16
☔ View full report in Codecov by Sentry. |
nathanielsimard
approved these changes
Nov 3, 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.
Refactored the autotune key (wgpu) as an enum and a struct. It should now be more robust, clear and efficient.