Stop overdubbing CUDA math functions #249
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
|
bors try |
|
As I mentioned on slack, things are a bit more complicated since we don't want to overdub these, but we still have to use Cassette for the time being. |
vchuravy
reviewed
May 11, 2021
tryBuild failed: |
This reverts commit da035c1.
|
Can you also add a test case for pow? |
|
bors try |
|
Not sure if it's what you had in mind but I added a test for |
|
bors try |
|
bors r+ |
vchuravy
pushed a commit
that referenced
this pull request
May 13, 2021
249: Stop overdubbing CUDA math functions r=vchuravy a=ali-ramadhan I think with CUDA.jl v3 there is no longer a need to overdub/rewrite CUDA specific math functions. If anything `CUDA.pow` doesn't exist anymore. Co-authored-by: Ali Ramadhan <ali.hh.ramadhan@gmail.com> Co-authored-by: ali.hh.ramadhan@gmail.com <ali.hh.ramadhan@gmail.com>
Merged
vchuravy
reviewed
May 13, 2021
| @@ -28,6 +32,11 @@ function compiler_testsuite() | |||
| @test !any(stmt->(stmt isa Expr) && stmt.head == :invoke, CI.code) | |||
| end | |||
|
|
|||
| let (CI, rt) = @ka_code_typed square(CPU())(zeros(1), zeros(1), ndrange=1) | |||
There was a problem hiding this comment.
I noticed to late that this is just CPU() and thus doesn't check the CUDAKernels implementation
bors bot
added a commit
that referenced
this pull request
May 14, 2021
251: Backport #249 r=vchuravy a=vchuravy Backport #249. @ali-ramadhan I noticed that you only tested `CPU()` execution, but the changes are limited to CUDAKernels. Co-authored-by: bors[bot] <26634292+bors[bot]@users.noreply.github.com> Co-authored-by: Valentin Churavy <vchuravy@users.noreply.github.com>
bors bot
added a commit
that referenced
this pull request
May 15, 2021
251: Backport #249 r=vchuravy a=vchuravy Backport #249. @ali-ramadhan I noticed that you only tested `CPU()` execution, but the changes are limited to CUDAKernels. Co-authored-by: bors[bot] <26634292+bors[bot]@users.noreply.github.com> Co-authored-by: Valentin Churavy <vchuravy@users.noreply.github.com>
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.
I think with CUDA.jl v3 there is no longer a need to overdub/rewrite CUDA specific math functions.
If anything
CUDA.powdoesn't exist anymore.