[hipblaslt] allow skipping kernel launch for matrix initialization #1431
Conversation
This file contains hidden or 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
same as used for GEMM kernel.
KKyang
reviewed
Sep 3, 2025
| @@ -126,6 +128,12 @@ void hipblaslt_init_device(ABC_dims abc, | |||
| size_t stride, | |||
| size_t batch_count) | |||
| { | |||
| if(TensileLite::Debug::Instance().skipKernelLaunch()) | |||
| { | |||
| std::cout << "DEBUG: Skip execution of kernel for matrix initialization" << std::endl; | |||
There was a problem hiding this comment.
Same reason why you would want to skip GEMM kernel execution (both with TENSILE_DB2=0x1). Perhaps for testing host code when a specific GPU is not available. This was requested by @bethune-bryant
There was a problem hiding this comment.
We shouldn't use TensileLite::Debug outside tensilelite.
There was a problem hiding this comment.
Tensile runtime env shouldn't be used by hipblaslt-bench tool. Don't mix those up.
eidenyoshida
pushed a commit
that referenced
this pull request
Sep 3, 2025
- Fix groovy multi-gpu test script. [ROCm/rocFFT commit: 68037a6]
KKyang
requested changes
Sep 4, 2025
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Outdated suggestions cannot be applied.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.
Motivation
This allows skipping the kernel for matrix initialization in hipblaslt-bench using the environment variable
TENSILE_DB2=0x1, which is also used to skip GEMM kernel execution.Useful for testing host code if a specific GPU is not available.
Submission Checklist