[RUNTIME][CLML] OpenCLML tuning and profiling enhanced #13843
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
|
Thanks for contributing to TVM! Please refer to the contributing guidelines https://tvm.apache.org/docs/contribute/ for useful information and tips. Please request code reviews from Reviewers by @-ing them in a comment. Generated by tvm-bot |
srkreddy1238
force-pushed
the
clml_tuning
branch
3 times, most recently
from
199755d to
4f672d5
Compare
Tuning cache bin is serialized through DMLC::Stream to support multiple CLML sub graphs with in a tvm module. Individual tuning cache blobs are saved to same output file. New API on OpenCLWorkspace to enable or disable profiling on command queue rather doing this only when Timer is invoked. This is required to perform CLML operator tuning. CLML layer profiling now uses OpenCL Timer interface. This PR also fix avoiding pad operator offloading at the very first layer (to be specific before at least one convolution layer) due to the limitation of CLML pad operator is concerned about layout. Please refer to CLML SDK documentation for more details.
srkreddy1238
force-pushed
the
clml_tuning
branch
from
4f672d5 to
9960020
Compare
Co-authored-by: Egor Churaev <egor.churaev@gmail.com>
srkreddy1238
force-pushed
the
clml_tuning
branch
from
a8b79f5 to
535d1e0
Compare
fzi-peccia
pushed a commit
to fzi-peccia/tvm
that referenced
this pull request
Mar 27, 2023
* [RUNTIME][CLML] OpenCLML tuning and profiling enhanced Tuning cache bin is serialized through DMLC::Stream to support multiple CLML sub graphs with in a tvm module. Individual tuning cache blobs are saved to same output file. New API on OpenCLWorkspace to enable or disable profiling on command queue rather doing this only when Timer is invoked. This is required to perform CLML operator tuning. CLML layer profiling now uses OpenCL Timer interface. This PR also fix avoiding pad operator offloading at the very first layer (to be specific before at least one convolution layer) due to the limitation of CLML pad operator is concerned about layout. Please refer to CLML SDK documentation for more details. * Update src/runtime/opencl/opencl_common.h Co-authored-by: Egor Churaev <egor.churaev@gmail.com> * * review comments --------- Co-authored-by: Egor Churaev <egor.churaev@gmail.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.
Tuning cache bin is serialized through DMLC::Stream to support multiple CLML sub graphs with in a tvm module. Individual tuning cache blobs are saved to same output file.
New API on OpenCLWorkspace to enable or disable profiling on command queue rather doing this only when Timer is invoked. This is required to perform CLML operator tuning.
CLML layer profiling now uses OpenCL Timer interface.
This PR also fix avoiding pad operator offloading at the very first layer (to be specific before at least one convolution layer) due to the limitation of CLML pad operator is concerned about layout. Please refer to CLML SDK documentation for more details.
Co-Authored-By: Krishna Raju Vegiraju quic_kvegiraju@quicinc.com