This repository has been archived by the owner on Nov 17, 2023. It is now read-only.
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Outdated suggestions cannot be applied.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.
Description
This PR serves to implement the remaining operators from the GEMM category in opperf. To achieve this, I added a call to
run_performance_test
for thekhatri_rao
op within therun_gemm_operators_benchmarks
function and, sincekhatri_rao
has yet to be implemented for GPU, added a check to ensure that the context was GPU before running benchmarks forkhatri_rao
. Sincekhatri_rao
was the only remaining op in the GEMM ops category, these were the only required changes.Checklist
Essentials
Please feel free to remove inapplicable items for your PR.
Changes
Comments
Tested on p2.16xl w/ubuntu 16.04 and Mac OS with:
run_gemm_operators_benchmarks
- runs all gemm ops (dot
,batch_dot
, andkhatri_rao
) on relevant dataopperf.py
(full run of all ops)Performance Results
Group of operator test - all GEMM ops (CPU)
Full OpPerf test (CPU)
Group of operator test - all GEMM ops (GPU)
Full OpPerf test (GPU)
@apeforest @access2rohit @ChaiBapchya