[CK] Add FP8 KV_BLOCKSCALE support for batch prefill
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Implement per-page K/V quantization for paged attention:
- Add KV_BLOCKSCALE enum to BlockAttentionQuantScaleEnum
- Use exp2 shift trick to eliminate explicit P scaling overhead
- Prefetch physical pages offset for KV cache, overlaps with
computations
## Proposed changes
Please describe the motivation behind the pull request, whether it
enables a new feature or fixes a bug. If there are associated pull
requests or issues, please link them to the pull request.
## Checklist
Please put an `x` into the boxes that apply. You can also fill these out
after creating the PR. If you're not sure, please don't hesitate to ask.
- [ ] I have added tests relevant to the introduced functionality, and
the unit tests are passing locally
- [ ] I have added the test to REGRESSION_TESTS list defined at the top
of CMakeLists.txt in tests/CMakeLists.txt, **IF** the test takes more
than 30 seconds to run.
- [ ] I have added inline documentation which enables the maintainers
with understanding the motivation
- [ ] I have removed the stale documentation which is no longer relevant
after this pull request
- [ ] (If this change is user-facing) I have added release notes which
provide the end users with a brief summary of the improvement from this
pull request
- [ ] I have run `clang-format` on all changed files
- [ ] Any dependent changes have been merged
## Discussion
If this is a relatively large or complex change, feel free to start a
discussion by explaining why you chose the solution you did and what
alternatives you considered
* WIP: add splitk to bquant
* feat: add support for bf8i4 and fp8i4 by calculating correct stride for packed data types
* chore: remove temporary test script
* fix: incorrect tile window length for splitted bq tensor window
* chore: improve comments
* test: add unit tests to cover bquant splitk functionality
* fix: conflict resolution by renaming variables
* [Compiler] Addressing new compiler warnings
Clang enables new lifetime warnings in production and we see build
errors due to this with the staging compiler.
The attributes added in this PR are suggested by the compiler. However,
I'm not very familiar with the code base, so the changes may be
incorrect.
* Update some more instances
* Adds file-level ignores via clang diagnostic pragma
The number of instances was large, so I decided to use file-level scope
to disable the warning via pragma clang diagnostic ignored.
It also showed this warning coming from the gtest dependency. For that,
I did add the respective command line flag to the CMake variables. I
don't know if this is acceptable or not.
* This adds the remaining instances
For a build on gfx90a.
* fix clang format
* Adding couple more instances from gfx1200 build
* Fixed another few instances
---------
Co-authored-by: Illia Silin <98187287+illsilin@users.noreply.github.com>
Co-authored-by: illsilin_amdeng <Illia.Silin@amd.com>
* Adding remaining flavors for grouped conv fwd
As titled. Following variants are added:
- grouped_conv2d_fwd_dynamic_op
- grouped_conv3d_fwd_dynamic_op
- grouped_conv3d_fwd_bilinear
- grouped_conv3d_fwd_convscale
- grouped_conv3d_fwd_convinvscale
- grouped_conv3d_fwd_convscale_add
- grouped_conv3d_fwd_convscale_relu
- grouped_conv3d_fwd_scale
- grouped_conv3d_fwd_combconvscale
- grouped_conv3d_fwd_scaleadd_scaleadd_relu
* Fix incomplete parsing of types from source names in add_instance_library() cmakelists function so we don't build f8 on RDNA3.
* Do not build f8 / bf8 only flavor tests on RDNA3
* Make sure we have proper generic instances for all instance lists related to the post-ces extra flavors, with scalarPerVector = 1. Then disable all but one generic instance per instance list to reduce compile time.
* Post rebase fix: Template parameters for Grouped Conv Fwd Device Impl got tweaked upstream.
* adding int8 and fp16 overloads to the elementwise operations
* fixed copilot nits
* Addressing review comments:
- removed unnecessary examples for dynamic op
- removed unnecessary conv specalizations for all the flavors
- removed spurious bilinear and scale source files
* clang-format
* reduced no of tests
---------
Co-authored-by: Wojciech Laskowski <wojciech.laskowski@streamhpc.com>
* chore: split block scale example instances in more separate files to speed up compile times
* wip: fp4 scaffolding for abquant
* feat: add fp4 decoding-while-loading to abquant pipeline
* feat: add support for fp4 CPU verification in abquant
* chore: add time tracking to reference calculation
* feat: add a4w4 test for blockscale gemm
* feat: optimize reference calculation by preconverting values to AccType
* feat: add fp4 to fp8 look-up table
* fix: reference to wrong ComputeDataType field in QuantProblem
* feat: type utilities for determining MFMA compute types
* feat: packed fp4 for abquant weight preshuffle
* feat: add separate tests for a4w4 base case, padding and preshuffleB
* fix: fp4 conversion on gfx950 attempting to use non-supported method
* fix: test case was using quant group sizes which don't work on gfx950 due to larger mfma tile size
* chore: add fp4 preshuffleb mode to block scale example
* chore: sanity check for packed types being 1 byte
* chore: clarify tensor dimension indices with constants
* chore: replace traits check with specialized check for packed types
* style: some minor refactoring and cleanup
* fix: correct conversion table for FNUZ fp8
* chore: add fp4 instances to main abquant instances again
* chore: use same initialization branch for int4 and fp4
* chore: add missing initialization for fp4 in block scale gemm example
---------
Co-authored-by: Thomas Ning <Thomas.Ning@amd.com>
* Add multi AB support to wave transfer
* Improviments to multi ABD examples
* Add instances and use intrawave v1 instead of interwave
* Apply changes to other transfers
* Wave transfer: add support for multiple internal vgpr buffers
* Fix compilation error gfx11
* initial commit
* preshuffleQuant support for ABQuant
* fix mxfp4 to use correct QuantGroupSize
* addressing review comments and seperated Preshufflequant for A and B
* updated grouped gemm example for updated traits definition
* fix for CI failure
* updated grouped_gemm_abquant test for updated traits definition
* updated grouped_gemm_abquant test for updated traits definition
* Add padding support with transpose
Also move check before writing storing is_src_valid during reading
* Add/modify instances to use wave transfer for gemm universal
Condition is changed so now the vectorsize of vmem reading and lds
writing must be equal to 8 in order to use the wave transfer
* Fix clang format
* Modify example
* Fix bwd data
* Add restriction for wave transfer with padding and transpose
Add test case which shows this limitation
* Fix validity checks 8 bit types
* Add validity check gemm_bias_add_reduce
* Add validity check grouped gemm tile loop
* Fix validity checks new flavours
* Minor fixes
* Fix clang format
* WIP: host level interwave pipeline compiles
* WIP: interwave implementation computes correct GEMM result when no aquant
* WIP: quantization works for subset of problem shapes
* WIP: quantization works for subset of problem shapes
* WIP: interwave memory pipeline passes local test
* feat: Add interwave pipeline implementation for memory pipline in aquant
* test: add unit test for aquant memory pipeline
* WIP: host level interwave pipeline compiles
* WIP: interwave implementation computes correct GEMM result when no aquant
* WIP: quantization works for subset of problem shapes
* WIP: quantization works for subset of problem shapes
* WIP: interwave memory pipeline passes local test
* feat: Add interwave pipeline implementation for memory pipline in aquant
* fix: compilation error on gfx950
* chore: remove debug statements from the code
* test: resolve merge conflict
* test: remove non rcr unit tests from test suite
* add block scale parameters to kernel
* add block scale to kernel
* add smoke test
* format
* Revert "format"
This reverts commit 356c3c9706.
* only format my code
* format py
* fix auto not allowd in function prototype
* change instance tttt to ttff
* fix structured binding issue
* change s_acc elementwise op
* async pipeline add block scale
* add quantation P using shift exp2
* precompute (m - shift) once per row
* change blk scale seqstrt ptr name
* fix some name
* fix for deduction guide
* fix some comments
* add P scale to qr_ksvs_pipeline
* add comment to idx_identity
* change the method of calculating descale block index
* unify naming style: use block_scale_ as name prefix
* unify naming style
* update the CHANGELOG.md
* Add FP8 block scale quantization support for FMHA forward kernel
---------
Co-authored-by: Illia Silin <98187287+illsilin@users.noreply.github.com>
Co-authored-by: Po Yen Chen <PoYen.Chen@amd.com>
* wip: test suite for batched gemm multiple d gemm multiple d, working on gridwise implenentation
* wip: many fixes in implementation of batched gemm gemm multiple d
* wip: batched gemm gemm multiple d gridwise op compiling, not working yet
* fix: incorrect d0 grid indexing in batched gemm gemm multipled
* feat: add instances for batched gemm add relu gemm add
* chore: configure instance with low vector transfer size for odd sizes
* chore: add some more validation to device batched gemm gemm multiple d, and removed template parameter that didn't really make sense
* fix: upate device_batched_gemm_gemm_wmma to work with new gridwise changes
* fix: disable odd size tests on XDL archs
* chore: removed temporary logging
* chore: update some references to C tensor to E tensor
* Tentative fix for example template params
* Tentative fix for non-multi-D batched gemm gemm device impl.
* Tentative fix for xdl example template params
* Tentative fix for profiler build on gfx90a
* chore: improve device batched gemm gemm multi D comment to include all ops and dimensions
* chore: explicitly call ck::make_tuple to prevent issues when std::make_tuple would apply
* fix: make the gemm1 data types match what happens in the device op
* feat: add d0s/d1s datatypes and layouts to the device op type string
* chore: change element-wise op so addition happens in fp32
* chore: add static asserts for gemm0/gemm1 calculated wave sizes
* chore: also updated other element-wise ops to use fp32 calculations
* chore: log number of supported instances
* chore: update instance comment
* chore: disable kernel timing in example by default
* fix: gemm1 wave size calculation
* fix: make sure batched gemm multiple d gemm multiple d profiler performs correct type conversions
* chore: remove increased tolerance in batched gemm gemm multiple d example
* chore: add comment explaining that verification fails for certain input values
* chore: clarify instance comment
---------
Co-authored-by: kiefer <kiefer.van.teutem@streamhpc.com>
Add signal-based synchronization for persistent GEMM kernels where
input data becomes available incrementally. Uses modulo wraparound
(like PyTorch's AsyncMM) for chunk index calculation:
chunk_idx = ((tile_idx + tile_idx_pivot) / tiles_per_chunk) % num_chunks
Key components:
- PersistentAsyncInputScheduler struct with tiles_per_chunk_m,
chunk_signals, tile_idx_pivot_m, and num_chunks fields
- wait_eq_wave method using __builtin_amdgcn_s_sleep for power efficiency
- IsSupportedArgument validation for scheduler parameters
- Example demonstrating async input scheduling with simulated producer
- GTest unit tests covering all layout combinations
* Add support to fp16 + compute fp16 and bf16 + compute bf16 contractions
Enables hipTensor to access the WMMA HW functionalities
for these combinations of datatype on gfx11 and gfx12.
* Fix change to contraction scale tests
* Fix clang-format
* feat: test setup for batched contraction (aka batched gemm multiple d e permute)
* wip: device struct for WMMA batched contraction multiple d based on new gridwise op
* feat: working batched contraction on RDNA, non-naive tensor descriptors for gridwise_gemm_wmma_cshuffle_v3, test setup for odd cases
* fix: failure to resolve template parameters when calling new function overload
* fix: passing reference type as parameter instead of underlying types
* fix: merge error caused duplicate definitions
* fix: make sure constness of template and parameters types match
* fix: don't compile batched contraction test on unsupported architectures
* feat: add example for new wmma implementation, and consolidate example code between platforms
* style: return inline instead of with branch
* chore: add extra assert on vector memory access sizes
* chore: clean up some unused variables
* fix: correct tail number calculation, added small cases and extra instances to the test
* fix: properly support wave transfer by generating correct grid descriptors dependent on the transfer method
* formatted
* formatted
* formatting
* formatting
* formatting
* [CK TILE GEMM] Refactor block_scale_gemm examples
- Split cpp file to reduce building time
- Support multiple GemmConfig
* [CK TILE GEMM] Refactor block_scale_gemm examples
- Update Readme
* enable prefill shapes
* [CK TILE GEMM] Refactor block_scale_gemm examples
- Add support for rowcol and tensor GEMM operations
* [CK TILE GEMM] Refactor block_scale_gemm examples
- Update README
* adding preshuffle quant as new parameter and its associated new files
* remove debugging statements
* adding test
* enable preshuffle quant with permuteN
* updating readme and correcponding gemmconfigs
* updating cmake file
* fixing CI failures for grouped quant gemm
* debugging permuteN
* debugging
* debugging PermuteN
* initial commit
* resolving merge conflicts
* adding test cases
* initial commit with prints
* debugging
* fine-grained working
* debugging medium grained
* fixing the tile window
* formatting
* enabling prefill shapes
* working prefill shapes
* formatted
* clean up
* code cleanup
* bug fix after merging with develop
* G128 working for both prefill and decode shapes for preshufflequant
* clean up after merging with develop
* fixing group 64 for decode shapes
* non preshufflequant working for group size 128
* enable preshuffleb and preshufflequant with variour group sizes
* reduce build time by splitting example into diff datatype files
* Adding tests for preshuffleQuant
* address review comment
* fix for gfx1201
* compile time fix for gfx1201
* clang formatted
---------
Co-authored-by: Cong Ma <congma13@amd.com>
Co-authored-by: Thomas Ning <Thomas.Ning@amd.com>
Co-authored-by: Agarwal <khuagarw@ctr2-alola-login-03.amd.com>
* CMakeLists.txt hack for Windows.
* Add Windows build instructions.
* Fix type issue with variadic min function.
* Use std::common_type to fix the variadic min/max functions.
* Enable CPU guard compilation on Windows.
* Suppress warnings related to std::getenv on Windows platform.
* Git ignore the output directory on Windows platform.
* Powershell script for running tests and generating reports.
* Improve test logging.
* Disable non-conv tests.
* Fix Debug build on Windows.
* More debug build changes.
* Update Windows build instructions.
* Enable all tests.
* Test fixes.
* Suppress not found linker options warning.
* Update unsigned long literals and format specifiers to work correctly in Windows
* Fix conv 3D bwd weight bilinear tests on Windows.
* Revert changes on .gitignore.
* Clean-up CMake project file for Windows builds.
* clang-format
* Fix definition of CMAKE_PREFIX_PATH on both Linux and Windows platforms.
* Fix building examples on Windows.
* Update Readme.
* Remove the suppression of the deprecated warnings.
* Remove Windows specific min/max implementations from CK Tile math core.
* Remove unnecessary no-op on Windows.
---------
Co-authored-by: User <user@example.com>
Co-authored-by: Ville Pietilä <none>
Co-authored-by: John Afaganis <john.afaganis@amd.com>
Co-authored-by: Ville Pietilä <>
* solve compiler issue
* solve the gfx950 mfma shuffle regression
* refactor jenkinsfile to handle arch name better
* [CK TILE] set divisor to count of thread along k dimension
* fix the compiler error
* solve degradation
* Finish the multiplies fix
* fix the scales
* solve compilation error
* solve the composes
* solve the error of tile sweeper
* fix the test and example
* fix for gfx950
---------
Co-authored-by: Max Podkorytov <4273004+tenpercent@users.noreply.github.com>
Co-authored-by: illsilin_amdeng <Illia.Silin@amd.com>
Co-authored-by: Cong Ma <congma13@amd.com>
* feat: grouped gemm tile loop support for RDNA4
* fix: removed extra parameter from grouped gemm example instance
* fix: FP8 check incorrectly enabling FP8 on RDNA3
- Enable page_size=1 support in batch prefill codegen (linear layout only).
- Implement per-token page lookup in `kv_offset_array_transform` for page_size=1 to handle 3D input tensors correctly.
- Relax `kPageBlockSize` alignment assertion for the page_size=1 case.
* nt on fp8 blockscale
* some improve and tests needs to be fixed
* update
* fix format
* revert useless change
* revert any change in amd_buffer_coherence
* WIP
* Add Unit tests for the Multi Reduction Kernel
* clang format
* Rename multiblock to threadwise
* Multiblock WIP
* Fix multi reduce multi block unit tests
* Multi Reduce Tile Engine: WIP
* refactoring + try addressing precision error
* Fix multiops examples
* Cleanup
* Clean up tile engine's reduce op
* Update changelog
* Fix remod/clang
* Fix dates
* Fix documentation & missing file
* Fix comments
* Use the update_tile api in the multi-block kernel
* Unify threadwise/multiblock into a single kernel + default multiblock output to float in tests
* Add TileParitioner
* Cleanup
* Add warning when no data to process, in the example
* Refactoring Reduce kernel Tile Partioner + cleanup
* Move the tile partioner to its own file
* Add missing includes
* Fix copyright header with update_amd_copyright_headers.py
* Fix change of interface in Reduce2dProblem
---------
Co-authored-by: Damien Lejeune <damien.lejeune@amd.com>
Co-authored-by: Adam Osewski <19374865+aosewski@users.noreply.github.com>
