This repository has been archived by the owner on Nov 17, 2023.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Outdated suggestions cannot be applied.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 all Linear Algebra operators in OpPerf. To achieve this, I created a file (
benchmark/opperf/nd_operations/linalg_operators.py
) with a function (run_linalg_operators_benchmarks
) to run all Linear Algebra ops. Within this function, I used calls torun_performance_test
for the linalg ops with specific requirements on input data values and a call to a new function I wrote (get_all_linalg_operators
) in conjunction withrun_op_benchmarks
for all other linalg ops. I also added a call torun_linalg_operators_benchmarks
inopperf.py
.I also disabled backwards runs on ops that do not support them (
linalg_gelqf
,linalg_slogdet
,moments
).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_linalg_operators_benchmarks
- runs all Linear Algebra ops on relevant dataopperf.py
(full run of all ops)Performance Results
Group of operator test - all linalg ops (CPU)
Full OpPerf test (CPU)
Group of operator test - all linalg ops (GPU)
Full OpPerf test (GPU)
@apeforest @access2rohit @ChaiBapchya