-
Notifications
You must be signed in to change notification settings - Fork 489
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
Feat/shared sum #2737
Feat/shared sum #2737
Conversation
) -> Result<JitTensor<Run>, String> { | ||
crate::kernel::reduce::reduce::<Run, E, E, Sum>( | ||
input, | ||
crate::kernel::reduce::ReduceStrategy::Autotune, |
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.
Just to make sure i understand, calling sum_chained during the autotune of sum-full will trigger the autotune of sum-dim, which may trigger several autotunes during the chain if the various dims are not similar?
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.
Exactly!
/// Key for reduce dim operations | ||
Reduce(ReduceAutotuneKey), | ||
/// Key for sum operations | ||
Sum(SumAutotuneKey), |
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.
Is SumDim a Reduce or a Sum? It's getting a bit confusing
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.
In the reduce kernel, there is no function or trait or type or anything really called SumDim or something similar.
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #2737 +/- ##
==========================================
- Coverage 83.91% 83.58% -0.34%
==========================================
Files 821 824 +3
Lines 106890 108340 +1450
==========================================
+ Hits 89702 90553 +851
- Misses 17188 17787 +599 ☔ View full report in Codecov by Sentry. |
I imported the
shared_sum
reduction from cubecl and added it with autotune.