* refactor the block_gemm_areg_breg_creg_v1 and add the v2 policy with 2x2 warp gemm
* Finished the 2x2 warp gemm policy and the block selection mechanism
* Clang format
* address poyen's comment
* Address feedbacks
* Fixed the compilation issue
* Change the function name
[ROCm/composable_kernel commit: 5d671a5fc4]
* Observed a 2x perf improvement with kBlockSize = 256
* Using 512 threads may lead to redundant computations
[ROCm/composable_kernel commit: 0b8f117f1a]
* 1. enable bias feature that add bias before adding residual; 2. change block size from 128->64 when m<64 in fp16
* delete comment
* 1.remove fmha change 2.change buffer name from bias to xbias
* Now bias can be used independently from fadd
* change kbias to kxbias
---------
Co-authored-by: feli <felix.li@amd.com>
[ROCm/composable_kernel commit: d5c8a334ca]
* Update license year
* Add initial code to override decode problem
* Fix splitkv traits/args overriding error
* Reshape and transpose lse for decode
* Remove debug code
* Prettify example code
* Use better function name
* Add kMergeNumHeadGroupsSeqLenQ flag
Kernel user can use this switch to turn on/off optimization for
some problem sizes
* Add missing flag declarations
* Default turn off kMergeNumHeadGroupsSeqLenQ in codegen
* Group similar statements together
* Remove assumption of seqlen_q=1
* Remove kMergeNumHeadGroupsSeqLenQ from splitkv combine kernel
* Support kMergeNumHeadGroupsSeqLenQ=true in fmha splitkv kernel
* Run kMergeNumHeadGroupsSeqLenQ=true kernels when need
* Fix group mode block skip logics
* Undo changes of normal fwd kernel
* Update in GridSize() and using GridSize() for splitkv kernel (#1799)
---------
Co-authored-by: Qianfeng <qianfeng.zhang@amd.com>
[ROCm/composable_kernel commit: 24b12d04af]
* add no welford
* enable output raw
* raw of int8
* fix build
* fix smoke test err
* [ck_tile]layernorm: fix welford ok, set int8 and bf16 small N as default and others open by generate
* [cktile]layernorm, fix err commit files and remove uselss
* fix quant 8192 err & change norm_reduce class and file name
---------
Co-authored-by: coderfeli <coderfeli@163.com>
Co-authored-by: carlushuang <carlus.huang@amd.com>
[ROCm/composable_kernel commit: 4bc610416a]
* Remove using tile partitioner for fmha_fwd_kernel
* Remove using tile partitioner for fmha_fwd_splitkv and splitkv-combine kernels
* Remove using tile partitioner for fmha_fwd_appendkv kernel
* Unify the format of GetTileIndex
[ROCm/composable_kernel commit: 4e076909b6]
* [CK TILE] Add split K support in GEMM
* Updates
* Fixes
* rebase
* fix
* Fix
* fixes
* support for batched gemm
[ROCm/composable_kernel commit: af66494880]
* Add check for zero values
* Add static assertions
* Remove invalid option '-e' in smoke_test.sh
* Use correct path of smoke_test.sh
* Avoid zero-sized shared memory array
* Add warning comment
* Replace expr by integer_divide_ceil() call
* Use more readable constant names
* Write down assumption as static assertion
* Add more diagnostic error messages
* Fix wrong BlockWarps when using default pipeline policy
* Add more static assertions for A LDS desc
* Allow using vector size < 8 for data type fp16/bf16
* Align vector size between DRAM dist & LDS desc
* Remove no-longer used func decl
* Fix wrong displayed piepline name
* Undo policy template changes for tile_example_gemm_basic
* Add missing space and make error message stands out
* Unify print precision
* Add missing include directive <iomanip>
* Replace constant 64 by get_warp_size() call
* Replace constant 128 by named variable: BankLength
* Add kAMBlock/kBNBlock attributes
* Allow usig different A/B warp dist for multiple blocks
* Add helper function to get warp dist encodings
* Add 4x64x4 fp16 warp gemm attribute impl
* Complete the A/B warp dist encoding logic
* Fix wrong thread mapping for C matrix
* Use smaller vector size for small tile
* Add static assert to block unsupported warp gemm impl
* Extract common code out as helper method
* Add 4x64x16 fp16 warp gemm type alias
* Add comment to warning developers
* Undo WarpGemmAtrributeMfma<> changes
* Use more clear static assertion error message
* Add trivial wrapper to get warp dstr encodings
* Only transpose warp gemm result if it's square
* Fix compilation error
* Support multi-block warp gemm (on N direction)
* Remove duplicated code
* Fix output encoding of warp gemm
* Fix wrong shape of WarpGemmAtrributeMfmaIterateK<>
* Remove unused code
* Fix wrong shape of WarpGemmAttributeMfmaImplF16F16F32M4N64K4
* Add type config for bf16_t
* Add 4x64x16 bf16 warp gemm
* Update WarpGemmAtrributeMfmaIterateKAndTransposedCDistribution
* Add 64x4x4 fp16/bf16 warp gemm impl
* Add 64x4x16 fp16/bf16 warp gemm
* Add static assertion for better error diagnostic
* Get Q dram dstr directly form block gemm
* Add missing header: fused_moe.hpp
* Allow specifying different warp-gemm for gemm0 & gemm1
* Store P matrix into LDS before gemm1
* Fix inconsistant kernel name
* Remove constraint on gemm0 & gemm1 block warps
* Remove unsupported vector size from checking list
* Allow using 4x64x16 warp gemm for gemm0
* Finish policy customization
* Finish pipeline modification
F#
* Use block warps in codegen
* Fix wrong rank of m_lds_window origin
* Use better distributed tensor
* Make P-store earlier
* Remove duplicated experssions
* Remove unnecessary tile window
* Create new files for new splitkv pipeline
* Separate old/new pipeline codegen logic
* Sync changes form develop
* Undo gemm kernel/pipeline changes
* Undo gemm example changes
* Remove blank lines
* Fix typo
* Use new warp gemm interface
* Fix link error
* Fix wrong pipeline tag
* Fix more link error
* Avoid unnecessary padding
* Always use vector load for K
* Padding on fastest dimension when necessary
* Force padding Q on hdim_q
* Set high dimension padding flag to false
* Re-format headers
* Use warps=<1, 4, 1> for both gemm0 & gemm1
* Fix complilation errors
* Remove m/l shuffle logics
* Ignore duplicate data when write lse_acc
* Use gemm0 block warps as lds tile width
* Remove hard-coded numbers
* Fix wrong distribution width
* Remove unnecessary code
* Add s_barrier before writing to LDS
* Store Q into LDS before gemm0
* Fix wrong Q tile size
* Use simple Q lds descriptor for debuging
* Use more realistic Q lds descriptor
* Add comment & use better variable name
* Make Q lds space not overlapped with others
* Remove unnecessary block_tile_reduce_sync() call
* Move Q load statements
* Move block_sync_lds() right before use
* Re-order instructions
* Remove necessary lambda expression
* Use 8 threads on kMaxSplits direction while doing reduction
* Tiny correction for using 8 threads on kMaxSplits direction for combine kernel
* Padding num_split direction of o_acc tile window to 4x
* Update splitkv combine pipeline design
* Add kN1 back to splitkv combine pipeline problem
* Fix compilation errors
* Add missing template parameter
* Fix wrong splitkv combine kernel name
* Fix wrong origin
* Fix wrong LDS descriptor shape
* Fix sync & reduction logics
* Remove unnecessary static assertions
* Extract tile size computation logics
* Make sure we can reuse padding flags in combine kernels
* Rename variables
* Use OaccDataType in BlockFmhaSplitKVCombinePipelineTileSizes<>
* Remove unnecessary static assertion
* Fix function name typo
* Add constraint on kN1 template parameter
* Hide K tile loading latency in earlier iteration
* Fix wrong splitkv kernel name
* Use s_shuffling to replace p_shuffling which removes the needs of cross-warp reduction
* Rename pipeline
* Fix wrong pipeline name attribute
* Add GetAlignmentQ() for NWarpSShuffle pipeline
* Separate Q tile into dram tile & register tile concepts
* Remove non-squre warp gemm transpose c type alias
* Fallback tile size changes for fmha fwd splitkv
* Remove redundant change
* Refine naming for the S tile
* Use better naming of the S tile dstr (read from lds)
* Share Q lds with K lds
* Tiny change
* Fix with using static_for for passing CI checking
---------
Co-authored-by: Qianfeng Zhang <Qianfeng.Zhang@amd.com>
[ROCm/composable_kernel commit: 37cdbf4f0e]
* Parser for a vector was added. Additionaly we valid correctnes of numbers
* Remove unnecessary comments
* Review part 1
* Review part 2
* Add const to variadic lambda
* Rename C->K
[ROCm/composable_kernel commit: e758d006a5]
* add ck_tile/smoothquant out stride parameter
* Remove the default stride value
---------
Co-authored-by: so <a.com>
[ROCm/composable_kernel commit: 4e73177684]
* add IsSupportedArgument to gemm_kernel
* add ut and do some refactoring
* switched to ck_tile's integral_constant
[ROCm/composable_kernel commit: feb9a2bd9b]
* Ck-tile, impl. grouped gemm
* Workspace is allocated by user, and is passed to the function
* Prepare test to new api design
* Unify GemTransKernelArgs, removing N0 param
* Add 1 to dim3 in paritioner
* Typo: gem - > gemm
---------
Co-authored-by: Adam Osewski <19374865+aosewski@users.noreply.github.com>
[ROCm/composable_kernel commit: 4cb3d7d7ea]
* add interwave scheduler for gemm mem pipeline
* Fix merge artifacts.
* Refactor unit tests.
* Switch to interwave scheduler for mem example
---------
Co-authored-by: Adam Osewski <19374865+aosewski@users.noreply.github.com>
Co-authored-by: Adam Osewski <Adam.Osewski@amd.com>
[ROCm/composable_kernel commit: e7b6286441]
* Fix cmake example build
* Support max3 in smoothquant one pass
* support max3 in two pass
* support max3 in add_rmsnorm_rdquant
[ROCm/composable_kernel commit: abae2afc72]
* Block universal gemm.
* Universal block gemm with interwave scheduler - draft.
* Refactoring
* Move a/b_warp_tiles into BlockGemmImpl
* set BlockGemmImpl as a class member
* Change tile size for more suitable to memory bound cases.
* Introduce kKPerThread to WarpGemm
* Add documentation comment.
* Fix Interwave scheduler block gemm.
* Add compute/memory friendly tile configuration.
* Clean
* New tile configurations in gemm mem example.
* Add more static checks and fix loop order in block gemm.
* Add more static checks and use warp gemm mfma dispatcher.
* Add default scheduler block gemm.
* Remove logging in example.
[ROCm/composable_kernel commit: b6bcd76d88]
* Change in fwd-splitkv kernel to support num_splits=1 case
* Update in codegen fwd-splitkv to make num_splits > 1 cases pass
* Specify instance traits in dispatch
* Fix link error for fp8 kernels
---------
Co-authored-by: Po Yen Chen <PoYen.Chen@amd.com>
[ROCm/composable_kernel commit: ce2bdf42a9]
* Generate group mode paged-attn kernel
* Enable paged-kvcache + group mode support
* Add missing header: fused_moe.hpp
* Add comment to explain kernel arg usage
* Make error message more clear
* Add comment for confusing data member names
* Add more comment for confusing variable names
* Fix typo in option description
[ROCm/composable_kernel commit: fb1ccfa9df]
* Finished the feature
* Modified the test file
* Test case update
* addresss comment
* Addressed the review comment
* Fixed the CI error
[ROCm/composable_kernel commit: 2b6458ddf2]
* [CK_TILE] add more stride for layernorm to support un-continuous Tensor
* align CK coding style
* extend strides to layernrom expample
* clang-format...
[ROCm/composable_kernel commit: 8ef8a994e7]
* optimze small N case using vec io and using rcp div
* [Ck_tile] layernorm, add param to control fastdiv; change generate codes and test pass
* [Ck_tile] fix blockSize compute in Generic2dBlockShape
* [Ck_tile]fix kfastfdiv template style
* [Ck_tile] layernorm, fix stype in review
---------
Co-authored-by: dummycoderfe <noplydummmycoder@163.com>
[ROCm/composable_kernel commit: 686a58a912]
Before, generate.py appended the list at the end of the output file.
When running the cmake configuration steps multiple times on the
examples, the blob list (such as fwd_blob_list.txt) would grow at every
configuration.
`library/src/tensor_operation_instance/gpu/mha/CMakeLists.txt` worked around
this issue by removing the output file if it exists.
Now, generate.py overrides the content of the output file.
There is no need for the workaround in the CMakeLists.txt;
and the issue is solved for the example projects too.
[ROCm/composable_kernel commit: 464abd235e]
* more accurate residual
* modify comment
* Fix literal case in README.md
---------
Co-authored-by: Po Yen Chen <PoYen.Chen@amd.com>
[ROCm/composable_kernel commit: cb6c5d39dc]
* add prenorm/postnorm support, refactor using generate.py
* update README
* update README
* fix format
* update some description and fix format
* update format
* format
* use non-raw for loading
* format and update n4096
* dynamic-quant ready
* update readme
* support fused dynamic-quant
* update fused-quant, with smooth
* update README
* update args
* update some based on comment
[ROCm/composable_kernel commit: c3a4800c5f]
* CK-Tile GEMM with memory bound pipeline.
* Memory bound gemm pipeline.
* Fix not closed namespace.
* Block gemm mem pipeline draft.
* Do not use ck_tile:: within ck_tile namespace.
* Refactoring & Move Layout info to pipeline problem.
* Get hot loop and TailNum information before lunching kernel.
* Fixes in pipeline.
* Add comment to load_tile_raw and change variable naming style.
* Few small changes & formatting.
* Do not use macro.
* Add gtests.
* Use AccDataType for Output of MFMA instruction.
* Formatting.
* Refactor gemm examples.
* Switch over to current block gemm.
* Use currently available pipeline policy.
* Refactoring and review comment.s
* Fixes after merge.
* Add missing include.
* Add load tile overload which accepts output tensor as parameter.
* This give 8% perf boost at the cost of using more registers.
* Rename example.
* Small changes.
* Fix compilation err and lower K.
* Support different layouts for A/B
* Fix vector size for different layouts.
* Rename Alignment into VectorSize
* Unblock tests.
[ROCm/composable_kernel commit: 24d996aae1]
* Add reduce2d new api
* Prevent user use cross warp reduction
* Fix bug of std caculation
* Add rmsnorm2d
* Add rmsnorm small example
* Remove static assert to prevent compile fail
* Add script to test performance and correctness
* Add missing cmake change
* refine naming
* refine example of rmsnorm
* Fix bug of rmsnorm
* Refine naming
* Fix cmake
* clang format
* Refine pipeline name
* Add add_rmsnorm2d_rdquant kernel
* Add reduce op
* host verification
* Fix bug of one pass pipeline
* Refine tile size
* Add two pass pipeline
* Rename two pass to three pass
* Fix bug of kSaveX == false
* Add instance library
* Add test script
* Fix bug of x verification
* Add save_x to trait
* Add README
* Move reduce2d into reduce folder
* Fix bug of welford when number of m warp > 1
* remove reduncant comment
* 1. move 06_rmsnorm2d to 10_rmsnorm2d
2. move 07_add_rmsnorm2d_rdquant to 11_add_rmsnorm2d_rdquant
* clang format and add missing header
* Add host validation of add + layernorm2d + rsquant
* Revert "Add host validation of add + layernorm2d + rsquant"
This reverts commit 936cb45797.
* Remove deprecated flag
[ROCm/composable_kernel commit: 3d60953477]
* Add ceil_to_qualified_tile_length()
* Rename kK0BlockLength to kQKHeaddim
* Add kSubQKHeaddim concept to support headdim96
* Fix in math.hpp to avoid using __half interfaces
* Add LdsBufferSequence instance for headdim96
* Update in fmha_fwd/fmha_fwd_splitkv codegen to support hd96 testing
* Disable hd96 instance generation in codegen fmha_fwd and fmha_fwd_splitkv to save compiling time
* Reformat one file
* Fix text alignment in fmha_fwd_splitkv.py
---------
Co-authored-by: Po Yen Chen <PoYen.Chen@amd.com>
[ROCm/composable_kernel commit: 8632221814]
* topk_softmax
* remove some file
* fix atomix linear_offset
* address various comment, and change sfc get_index api to static(tuple)
[ROCm/composable_kernel commit: b098b71b05]
* Use pre-defined constants for readability
* Use vector write for o_acc tensor
* Remove no-longer used policy method
* Deprecate no-longer used policy/pipeline
* Specify gemm0/gemm1 block warps separately in codegen
* Fix wrong ps_idx creation logic
* Add single-warp block gemm
* Supoprt single-warp gemm0
* Make MakeCBlockTile() as static method
* Use MakeCBlockTile() to get underlying tile distribution
* Use kNumGemm1Warps to compute # threads for gemm1
* Put normal case in the if clause
* Refine fmha splitkv block mapping
* Refine & fix the lse_acc/o_acc layout
* Fix wrong LDS size for K tile
* Use kK0=64 for hdim=128,256 fmha splitkv kernels
* Use kK1=64 for hdim=32,64,128 fmha splitkv kernels
* Undo kK0/kK1 changes
* Use more reasonable GetAlignmentV() computation
* Using store_tile() in fmha splitkv kernel epilogue
[ROCm/composable_kernel commit: 54f0e6f4bb]
* port layernorm
* change warp_welford.hpp
* Update warpshuffle
* 1. Add save mean and save std back
2. Move construction of tensor_view and tile_window to operator()
* refine welford max count calculation
* unify layernorm api
* Rename file
* Remove save mean and inv std
* Revert "refine welford max count calculation"
This reverts commit 022365802b.
* Fix order of parameter
* refine welford max count calculation again
* Remove fp32 instances
* Fix bug of padding
* refactor api
* Support bf16
* Extract common function
* Refine arg of operator()
* Add kMThreadPerBlock to template parameter
* clang format
* Refine variable name
* Refine file name
* remove redundant line
* refactor layernorm2d pipeline and add block-per-block utility
* fix name
* rename more
* add more block-per-tile instance
* remove duplicated define
* update instance for 2048, 1024 case
* support up to 2048 now
* opt loading
* add n1536
* Add two pass pipeline
* format
* Fix incorrect type
* parallel compilation
* Use smaller N
* fix 2p pass
* Support Repeat_M in distribution
* Refine nameing
* Add reduce example
---------
Co-authored-by: letaoqin <letaoqin@amd.com>
Co-authored-by: aska-0096 <haocwang@amd.com>
Co-authored-by: rocking <ChunYu.Lai@amd.com>
Co-authored-by: carlushuang <carlus.huang@amd.com>
[ROCm/composable_kernel commit: 0394f8a713]
