Upgrade Arm xGEMM and Convolution tests #102
Open
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
Update and modernize Arm DeepBench code for easier testing with automated infrastructure. Additionally, convolution and gemmlowp tests do not use OMP, so added in thread-count specification to the command line parameters to allow controller how many threads are spawned. Signed-off-by: Geoffrey Blake (Geoffrey.Blake@arm.com)
|
Is there anybody willing to review these patches? |
added 2 commits
December 19, 2018 08:26
Adds in support for Arm Compute Library to be tested for xGEMM operations in addition to convolution. Also cleans up the Arm Makefile, cblas test and fixes up the conv_problems for Arm to not hit a segmentation fault. Signed-off-by: Geoffrey Blake (Geoffrey.Blake@arm.com)
Signed-off-by: Geoffrey Blake (Geoffrey.Blake@arm.com)
geoffreyblake
force-pushed
the
upgrade-arm-gemm-conv-tests
branch
from
520c66e
to
0843d86
Compare
|
Looking for a review. Any takers? Arm would like to get this included as we have other partners who use Deepbench for benchmarking their parts. |
|
@geoffreyblake As there have only been only 2 commits to DeepBench in the past 6 months, I assume the DeepBench project is abandoned. Please consider using CK-NNTest based on a similar idea but using CK for portability, reproducibility and crowdsourcing. (We started CK-NNTest as part of our collaboration with the Arm ML group, and effectively used it to optimize deep learning operators in the Arm Compute Library.) |
|
Thanks @psyhtest , I will take a look into that repository as a possible landing spot for this code. |
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Outdated suggestions cannot be applied.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.
This pull request upgrades the Arm xGEMM tests with new library bindings to the Arm Performance Library, Arm Compute Library and OpenBLAS library to test and compare with the existing GemmLowp library. It also upgrades how the tests are instantiated with new command line arguments to specify individual matrices, or the built in tests, and # of threads.
Output from GEMM and Convolution tests now output in a machine readable CSV format instead of just human readable text.
These changes are particularly useful for inserting Deepbench into automated testing infrastructures to get finer control and allow automatic parsing of test results.