David Dixon

Minor CMake cleanup to prevent redundant work in client builds

3

The CMake support in Tensile (namely TensileConfig.cmake) is doing unnecessary work. In turn, the configure/build times in downstream projects can take up to 20% longer than is necessary. The unnecessary...

The goal of this change set is to move the default compiler from hipcc to amdclang++. The most significant difference between hipcc and amdclang are the flags used when invoking...

The goal of this PR is to refactor the logic associated with the `generateLogicDataAndSolutions` into smaller testable functions with documentation. This will enable finer grained profile results that will help...

Disable failing rhel9 tests

2

The rhel9 tests are failing on develop post-merge due to a link error when building the client via: ``` tox run -e ci -- -m unit ``` and also when...

Adds example documentation for TensileCreateLibrary.

@AlexBrownAMD observed that tensilelt was significantly slower than tensile when running as follows: ```bash ./Tensile/Bin/Tensile test.yaml tensile-out ``` Profiling revealed that `writeBenchmarkFiles` was called by tensilelt leading to the additional...

Adds a decorator to make profiling easier.

Cheery picks commit from develop that resolves SLES test failures.

Master solution libraries (both .dat and .yaml) take considerable time to write to disk. This operation is current serialized. This PR - parallelizes writing the master solution libraries - adds...

This PR removes support for the `--short-names` option which reduces complexity and risk for regression. The corresponding naming functions were removed as well.