GitLab

Avoids double global loads.  Strided loops are unrolled which lets store results in array which compiler will use registers for since the index access is constant.   Updated to handle large reductions so which results with a better stable diffusion result

Local Threads of multiples 32 were introduced in #1348
But LocalThreads that are not multiple of 64 are causing correctness issues.

Updated each source file in the repo with the existing license.

There is significant improvement on larger tensors with half almost 50% faster:

lens: [1024, 384, 768]
gpu::code_object[code_object=13832,symbol_name=kernel,global=39321600,local=256,]: 1.16685ms
gpu::reduce_sum[axes={2}]: 1.73126ms
Also for non-trivial layouts this can sometimes be over 2x faster:

lens: [64, 1024, 768, 4]
gpu::code_object[code_object=13832,symbol_name=kernel,global=39321600,local=256,]: 1.1706ms
gpu::reduce_sum[axes={1}]: 2.63375ms
Of course if the stride becomes larger this speed improvement diminishes due to poor memory access patterns. A lane_reduce instead of a block_reduce is needed for such type of kernels. I plan to address that in a future PR.

Finally, this also includes a MIGRAPHX_GPU_DUMP_ASM env variable which will print out the assembly when the kernel compiles.

Checks wavefront size, then changes implementation and number of threads for DPP reduce

* code backup

* clang format

* change corresponding tool files

* clang format
Co-authored-by: mvermeulen <5479696+mvermeulen@users.noreply.github.com>

* Add all_targets cmake target

* Rename target

* Add ref target

* Rename tests

* Refactor compiler target

* Formatting

* Verify for every target

* Formatting

* Add verify test suite

* Formatting

* Add initial test programs

* Formatting

* Add rnn tests

* Formatting

* Validate gpu

* Formatting

* Remove old gpu tests

* Fix gpu tests

* Fix ref error

* Fix tidy issues

* Formatting

* Tidy fixes

* Fix header in python api

* Rename to ref

* Use ref in verify_onnx

* Fix tidy issue

* Build with verbose on

* Fix typo

* Remove verbose

* rename some cpu prefix to ref
Co-authored-by: Shucai Xiao <Shucai.Xiao@amd.com>