GitLab

* Use smaller width for lse_accum dist tensor

* Update pipeline comment

* Fix wrong distribution for lse_accum

* Remove duplicate dim in lse_accum dist encoding

* Decide fmha splitkv combine kernel kBlockSize by kM0

* Remove assumption of MPerThread=1

* Add log<4> & log<8> specialization

* Enlarge occupancy array

* Fix vector size for small tile

* Add support for kMaxSplits=8

* Re-format gemm.hpp

* Use 16x16x16 warp gemm for fwd_splitkv

* Centralize policy code changes

* Leave fp8/bf8 tile settings unchanged

This reverts commit 02236580.

* Add kQKHeaddimForGemmN and kVHeaddimForGemmN in order to support headdim 96

* Remove the using of MakeKRegBlockDescriptor and MakeVRegBlockDescriptor

* Fix in bwd_piple_default_policy

* Remove kQKHeaddim and rename kQKHeaddimForGemmN to kQKHeaddim in the bwd kernel and pipelines

* Replace kVHeaddimForGemmN by kVHeaddim and kDoDvHeaddim

* Update to hd96 tile settings

* Add smoke test scripts for fmha-bwd hd96

* Revert "Add smoke test scripts for fmha-bwd hd96"

This reverts commit 7ca7e1a93dc65eb99ce3ff4e82693589830e42a2.

* Remove hd96 tile settings in fmha_bwd codegen to save compiling

* Fix lost code line in bwd_pipeline_default_policy

* Merge kDoDvHeaddim/kPadHeadDimDoDv to kVHeaddim/kPadHeadDimV and remove TileFmhaBwdTraits

* Rename KRegSliceBlockDescriptor/VRegSliceBlockDescriptor to KRegBlockDescriptor/VRegBlockDescriptor

* tiny adjustments

---------
Co-authored-by: Po Yen Chen <PoYen.Chen@amd.com>
Co-authored-by: danyao12 <Dan.Yao@amd.com>

* [CK_TILE] Add block universal gemm pipeline policy

* Fixes

* fixes2

* Fixes3

* fixeS

* Add non_native_vector_type

* Add a test

* Add non-native vector type

* Fix CTOR

* Fix non-native vector type of 1

* Fix CTORs

* Use vector_type to cover non-native implementation as well

* Update the test

* Format

* Format

* Fix copyright years

* Remove BoolVecT so far

* Add AsType test cases

* Update assert error message

* Remove redundant type

* Update naming

* Add complex half type with tests

* Add tests for vector reshaping

* Add missing alignas

* Update test/data_type/test_custom_type.cpp
Co-authored-by: Adam Osewski <19374865+aosewski@users.noreply.github.com>

* Compare custom types to built-in types

* Add default constructor test

* Add an alignment test

---------
Co-authored-by: Illia Silin <98187287+illsilin@users.noreply.github.com>
Co-authored-by: Adam Osewski <19374865+aosewski@users.noreply.github.com>
Co-authored-by: Po Yen Chen <PoYen.Chen@amd.com>

* Add transpose scale amax example

* fixes

* Tune reduce instance

2. Move construction of tensor_view and tile_window to operator()

* decouple the calling from gemm_pipeline

* clang format

* ake the cshuffle compilable

* modify Mhe reference on gpu and cpu. Correaccess of cshuffle

* fix the cpu reference code

* Complete the in tile shuffle logic

* restructure the kernel template input

* change the naming pattern of ck_tile gemm pipeline

* Re-format files using remod.py

* Solve the fmha conflict with gemm

* Comment Addressed from Carlus

---------
Co-authored-by: Po Yen, Chen <PoYen.Chen@amd.com>

* Fix text alignment of ArgParser::print()

* Update example README files

* Clarify make-ck-dev.sh <arch> usage

* Only keep some of the argument from '-?' output

* Undo command line output changes in README

* Only keep existing argument on doc and update description

* Fix text alignment

* Make cmake-ck-*.sh compatible with 'sh' command

* Simplify the codes in splitkv_combine pipeline

* Always set kPadSeqLenK=true for fmha splitkv kernels

* Change in Oacc Alignment and TileDistribution to be more adaptable to tile sizes

---------
Co-authored-by: Po Yen Chen <PoYen.Chen@amd.com>

* update build logic with GPU_ARCHS

* fix the GPU_ARCHS build for codegen

* unset GPU_TARGETS when GPU_ARCHS are set

Co-authored-by: Po Yen Chen <PoYen.Chen@amd.com>

* Fix compile error

* Add one pass pipeline

* Extract creating tile_window to operator()

* clang format

* reduce duplicated code

* do not hardcode

* Support padding in layernorm

---------
Co-authored-by: Po Yen Chen <PoYen.Chen@amd.com>

* Adding seed and offset pointer support to the philox random number generator.

* Separating seed and offset pointer checks with different condition statements.

* Changes include, adding support for device seed and offset pointers, union is used to store seed/offset values and device pointers to minimize device SGPRs.

* Correcting a typo in the readme file

* Re-format files using remod.py

* Use STL type for API parameters

* Use simpler struct design for drop_seed & drop_offset

* Undo unnecessary changes

* Sync kargs style for fmha_fwd.hpp/.cpp

* Use templated union to reduce code

* Use structured binding to make code more readable

---------
Co-authored-by: Sudhir Kylasa <sukylasa@amd.com>
Co-authored-by: Po Yen Chen <PoYen.Chen@amd.com>

Without this change, the following diagnostic is generated:
  a template argument list is expected after a name prefixed by the template
  keyword [-Wmissing-template-arg-list-after-template-kw]

See C++17 spec [temp.names] p5.

* add missing vector header

* Re-format header using remod.py

---------
Co-authored-by: Po Yen, Chen <PoYen.Chen@amd.com>

* Use same layout for o_acc and o tensor

* Use better param names in partitioner

* Remove redundant kargs 'max_seqlen_q'

* Use better param names in splitkv kernel

* Add comment for additional kernel arguments

* Sync empty loop early return logics between pipelines

* Pass more arguments to cmake in scripts

* Align backslashes

* Fix wrong o_acc tensor view strides

* Change o_acc layout if o_perm=0

* Handle whole row masked via attn_bias

* Use use vector width = 1 for o_acc

* Use more even split sizes

* [CK_TILE] Image to Column kernel

* Fixes

* Vector loads and stores

* Fixes

* Fixes

* change test dir name

* add barriers

* tail bias barriers

* adjust bf16/hd256 tol

* continue adjust bf16/hd256 tol

* fix clang20 compilation errors for gfx90a

* fix clang20 compilation errors for gfx11 targets