GitLab

Bumps [tensorflow](https://github.com/tensorflow/tensorflow) from 2.6.4 to 2.7.2.
- [Release notes](https://github.com/tensorflow/tensorflow/releases)
- [Changelog](https://github.com/tensorflow/tensorflow/blob/master/RELEASE.md)
- [Commits](https://github.com/tensorflow/tensorflow/compare/v2.6.4...v2.7.2

)

---
updated-dependencies:
- dependency-name: tensorflow
  dependency-type: direct:production
...
Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
Co-authored-by: Chris Austen <causten@users.noreply.github.com>

Following up on issue #1166 and PR #1220. Using the same approach as in #1220 for parallelizing the eval calls, we can significantly reduce the time spent on eliminate_contiguous pass.

Addressing issue #1166 - propagate_constant pass currently uses a recursive approach to find all instructions in a module that can be evaluated to a literal and performs the replacement in the same call.

New approach:

Perform single pass though instructions in the module to determine which instructions can be evaluated
Evaluate selected instructions in parallel
Replace the selected instructions with the corresponding literal

* Upgrade to cppcheck 2.8

raw is the download for the file, blob is the url for the github page.

Bumps [tensorflow](https://github.com/tensorflow/tensorflow) from 2.5.3 to 2.6.4.
- [Release notes](https://github.com/tensorflow/tensorflow/releases)
- [Changelog](https://github.com/tensorflow/tensorflow/blob/master/RELEASE.md)
- [Commits](https://github.com/tensorflow/tensorflow/compare/v2.5.3...v2.6.4

)

---
updated-dependencies:
- dependency-name: tensorflow
  dependency-type: direct:production
...
Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
Co-authored-by: Chris Austen <causten@users.noreply.github.com>

* Improve applicable batched gemms for bert

Remove std references in runtime compilation since these are not available when using hiprtc and the headers may not be available on the system

* Fuse gemm add with pointwise fusions

As described in #1196, the ONNX mean parser does not work correctly for integral types. This update fixes the issue by handling integral types separately, where summation is performed before division. Additional test cases have also been added for handling integral types.

For clarity on kernel names found when profiling. The new names are set to the order of the ops being compiled. For example: add + relu = add_relu_kernel.

Updated variable names according to #1193

Our documentation indicates a user with sudo can run the install_prereqs.sh file. Turns out that the file is not complete enough to run on Ubuntu 18.04/20.04 independently. I updated the file to resolve the failures.

resolves #1191

Fuse layernorm and added triadd_layernorm fusion.  This is a prep performance booster

ONNX Models changed from master to main. Changing path reflect the proper location

Expose add_literal method in C/C++ api

Improves performance for add_gelu.  In bert it is 4x faster and for mul_add it is 50% faster than what we current have.

Move to CI containers to rocm 5.0.2
upgrade to 20.04
free up some more file space in github action environments

Add compile tests for gpu math functions

Fixes the #error when using cppcheck. This no longer suppresses cppcheck errors when including those errors. This fixes the cppcheck errors that was there already.

Helps avoid dangling references. This also deprecates the constructors that didnt take a lifetime annotation since its ambiguous the lifetime.

Release branch created for ROCm 5.2 so moving develop branch to 2.3

Add ref and gpu implementations for ONNX op GatherND

Resolves #1032

With reductions such as {2048, 2, 1456} on axes 1, this is 23x faster than using our new block_reduce, and its even over 100x faster than our original reduce_sum:

# lane
gpu::code_object[code_object=13736,symbol_name=kernel,global=2981888,local=1024,]: 0.0672928ms
# block
gpu::code_object[code_object=13800,symbol_name=kernel,global=39321600,local=64,]: 1.46072ms
# original
gpu::reduce_sum[axes={1}]: 6.73456ms
There is some basic logic to pick between lane and block reduce automatically.

* expose get_queue method

Implements the ReverseSequence ONNX operator as a parser.

This parser can only handle a constant sequence_lens input. This is the same as what is handled for TensorRT as far as I can tell.
We could handle a variable sequence_lens input; that would require ref and GPU implementations of the operator.
The ONNX backend tests are disabled because this does not handle variable sequence_lens.

Refactored the reference implementation of pooling to something like what was done for roialign. Moved the reference implementation of pooling from targets/ref/lowering.cpp to pooling.hpp.
Removed cpu_pooling, instead using reference pooling in pooling.hpp
Added reference implementation of Lp Norm pooling and the global version
Added tests for the Lp Norm Pooling