Reduce on cuda #274
Merged
Reduce on cuda #274
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
louisfd
approved these changes
Nov 19, 2024
maxtremblay
commented
Nov 19, 2024
crates/cubecl-cpp/src/shared/base.rs
Outdated
Comment on lines
30
to
48
| fn bfloat162_type_name(f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result; | ||
| // warp instructions (all threads participating) | ||
| fn warp_shuffle(input: &CppVariable<Self>, id: &CppVariable<Self>) -> String; | ||
| fn warp_shuffle_indexed( | ||
| input: &CppVariable<Self>, | ||
| index: usize, | ||
| id: &CppVariable<Self>, | ||
| ) -> String; | ||
| fn warp_shuffle_xor(out: &CppVariable<Self>) -> String; | ||
| fn warp_shuffle_xor_indexed(out: &CppVariable<Self>, index: usize) -> String; | ||
| fn warp_shuffle_down(out: &CppVariable<Self>) -> String; | ||
| fn warp_shuffle_down_indexed(out: &CppVariable<Self>, index: usize) -> String; | ||
| fn warp_all(out: &CppVariable<Self>) -> String; | ||
| fn warp_all_indexed(out: &CppVariable<Self>, index: usize) -> String; | ||
| fn warp_any(out: &CppVariable<Self>) -> String; | ||
| fn warp_any_indexed(out: &CppVariable<Self>, index: usize) -> String; | ||
| // Matrix-Multiple Accumulate | ||
| fn mma_namespace() -> &'static str; | ||
| } |
There was a problem hiding this comment.
Before merging, I want to double check you are ok with the approach used here. An alternative is to merge the indexed and regular methods into a single method with signature
fn do_something(&CppVariable<Self>, index: Option<usize>) -> String;Or to only support IndexedVariable
fn do_something(&IndexedVariable) -> String;
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Outdated suggestions cannot be applied.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.
The initial goal was to support testing the reduction algorithms on CUDA. This lead to some work on the warp operations in the compiler to support vectorization. I also added a bunch of related tests.