Chao Liu

CK Tile Programming Interface and some examples

Pointwise kernel choose grid size based on number of CU

2

https://github.com/ROCmSoftwarePlatform/composable_kernel/pull/261#discussion_r883726267

- [ ] https://github.com/ROCmSoftwarePlatform/composable_kernel/pull/395#discussion_r964097258 - [ ] https://github.com/ROCmSoftwarePlatform/composable_kernel/pull/395#discussion_r964097985 - [ ] https://github.com/ROCmSoftwarePlatform/composable_kernel/pull/395#discussion_r964098118 - [ ] https://github.com/ROCmSoftwarePlatform/composable_kernel/pull/399#discussion_r964049341

major code refactor

1

# major - [x] Add license - [x] use absolute path for header - [x] fix header dependency: https://github.com/ROCmSoftwarePlatform/composable_kernel/issues/170 - [ ] Tensor operation naming: grid/block/warp/thread-level operation, conv/gemm/reduce/elementwise operation -...

Naming for Reduction Operation

4

_Originally posted by @j4yan in https://github.com/ROCmSoftwarePlatform/composable_kernel/pull/128#discussion_r832741425_ # IdentityValue This value is a mathematical property of reduction type, and should be deterministic and not specified by user of reduction https://docs.oracle.com/javase/tutorial/collections/streams/reduction.html #...

Post-merge issue of PR #644

3

The corresponding logic in #644 should be ok, but some threads may not do copy thread-wise copy. This logic may be better: https://github.com/ROCmSoftwarePlatform/MIOpen/pull/644#discussion_r545192142 This logic need to be inspected: https://github.com/ROCmSoftwarePlatform/MIOpen/pull/644#pullrequestreview-554817964

Use shorter name for iGEMM kernel wrapper

1

@asroy I mean this `gridwise_convolution_backward_data_implicit_gemm` is always the same (except "backward_data" of course), so I would replace it with smth like `gw_conv_fwd/bwd/wrw_igemm` or even `conv_igemm_gw_fwd/bwd/wrw`... _Originally posted by @atamazov in...

For some reason, if CI passed previously for a old commit, github will allow "merge" option, even new CI of new commit is not finished.

- [ ] Have a PR that fix every Device Op class Issue brought up by @j4yan https://github.com/ROCmSoftwarePlatform/composable_kernel/pull/128#discussion_r832736874 Example Fix (need to be applied for all Device Op: ``` ---...