mirror of
https://github.com/ROCm/composable_kernel.git
synced 2026-05-12 17:26:00 +00:00
27e0a34e0fa8e0876248a2c2d6ed85b655bf486a
279 Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
Bartłomiej Kocot
|
27e0a34e0f |
[rocm-libraries] ROCm/rocm-libraries#4406 (commit 61f9f90)
[CK] CK Tile grouped convolution direct load ## Motivation CK Tile grouped convolution forward direct load support. ## Technical Details Basic pipeline for direct load and new instances for forward for v1 and v4 pipelines. ## Test Plan test_grouped_convnd_fwd_tile ## Test Result CI pending ## Submission Checklist - [x] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. AICK-130 |
||
|
Enrico Degregori
|
984a3d1828 |
[rocm-libraries] ROCm/rocm-libraries#4372 (commit 738ffd7)
[CK] Workaround blockscale wp test failure ## Motivation Workaround to fix blockscale wp test failure for pipeline v3 ## Technical Details <!-- Explain the changes along with any relevant GitHub links. --> ## Test Plan <!-- Explain any relevant testing done to verify this PR. --> ## Test Result <!-- Briefly summarize test outcomes. --> ## Submission Checklist - [ ] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. |
||
|
Illia Silin
|
569640dc70 |
Revert "Implement device grouped gemm fixed nk multi abd for rdna4 (#3619)" (#3705)
This reverts commit
|
||
|
Zoltán Lakatos
|
301eb5cf08 |
Implement device grouped gemm fixed nk multi abd for rdna4 (#3619)
* device struct implementation * added xdl grouped multi abd fixed nk testing * wmma implementation fixed * avoid unnecessary device mem allocation and code cleanups * cleanup instances definitions * wmma examples added * code cleanups * fix clang format * typo and compilation fixes related to reference gemm * fix compilation error due to std::remove_cvref_t * added missing hip_check_error includes * correction to example instances * review commentes addressed * removed split-k from testing * code formatting --------- Co-authored-by: Zoltán Lakatos <zoltan.lakatos@streamhpc.com> Co-authored-by: illsilin_amdeng <Illia.Silin@amd.com> |
||
|
Kiefer van Teutem
|
2377a62837 |
Adding remaining conv, dynamic_op, and scaleadd_scaleadd_relu flavors for grouped conv fwd (#3529)
* 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> |
||
|
Johannes Graner
|
fabac7e2c3 |
[Conv] Enable bwd weight splitk autodeduction with cap (#3656)
* Enable bwd weight splitk autodeduction with cap * Fix error threshold calculations * Add missing logic to wmma multiple d kernel * Fix threshold calculation * Update test with new applicability |
||
|
Bartłomiej Kocot
|
3d67e6c492 |
[CK TILE] Enable CK TILE Conv Fwd tests in CI and fix check_err (#3624)
* [CK TILE] Enable CK TILE Conv Fwd tests in CI and fix check_err * Update test_grouped_convnd_fwd_tile.cpp * Update test_grouped_convnd_fwd_tile.cpp * Update conv_tuning_params.hpp * clang format fix * Update CMakeLists.txt |
||
|
Johannes Graner
|
c190d8d61f |
[CK tests] Extend conv GPU reference (#3539)
* test_convnd_fwd
* test_convnd_bwd_data
* test_conv_bwd_data_scale
* test_grouped_convnd_fwd_clamp
* test_grouped_convnd_fwd_scale
* multiple A/B tensors and D tensor for fwd GPU ref
* test_grouped_convnd_fwd_scaleadd_ab
* test_grouped_convnd_fwd_bias_clamp
* test_grouped_convnd_fwd_bilinear
* test_grouped_convnd_fwd_gk_bias_clamp
* Extend GPU reference to enable batchnorm epilogue
* test_grouped_convnd_fwd{,_gk}_bias_bnorm_clamp
* test_grouped_conv_bwd_data_bilinear
* test_grouped_convnd_bwd_weight_bilinear
* Add missing template instantiation
* Perform operations in float in reference
* Slightly increase tolerance for batchnorm profiler
* Revert "Slightly increase tolerance for batchnorm profiler"
This reverts commit
|
||
|
Robin Voetter
|
cc75948d1c |
[CK_BUILDER] conv bwd weight testing (#3618)
* ck-builder: restructure testing conv In order to prepare for bwd of conv testing, this commit moves some files and types around so that we can reuse ckt::Args for both forward and backwards convolution. * ck-builder: decouple fwd_ck.hpp and fwd_reference.hpp from fwd.hpp This will allow us to more easily include fwd.hpp from backwards definitions, which is required for initializing bwd values. * ck-builder: fix layout of test_ckb_conv_bwd_weight_xdl_cshuffle_v3 Turns out that the supplied layout isn't actually supported... * ck-builder: ck and reference conv integration for bwd weight * ck-builder: ck bwd weight execution test * ck-builder: ckt::run support for ck-tile bwd weight * ck-builder: ck tile bwd weight execution test * ck-builder: extra debug printing in MatchesReference * ck-builder: make ckt::run return RunResult This type is more convenient than std::tuple, as it will allow us to use google test matchers with this in the future. * ck-builder: RunResult matcher Using EXPECT_THAT(..., SuccessfulRun()) will generate a check and a nice error message about how and why running an algorithm failed. * ck-builder: doc fixes * ck-builder: add missing headers |
||
|
ApoorvaKalyani
|
8daf6ea302 |
Grouped conv_fwd_bias_bnorm_clamp instances and tests (#3525)
* Added bias_bnorm_clamp instances. * fwd_bias_bnorm_clamp comp instances * fwd_bias_bnorm_mem_inter and mem_intra instances * fwd_bias_bnorm_merged_group_instances * fwd_bias_bnorm_clamp_conv3d_bf16 and f16 instances * Device level changes for fwd_bias_bnorm_clamp * Added the test to the regression test list. * Removed the part 2 and 2x instances * Removed the irrelevant checks in wmma * Refactored the instances to adapt to new device implementation * Updated the reference and include files * enabling tests * Added missing profiler * Added missing instance entry , deleted by mistake * Reduce bias bnorm clamp instances to only a single generic one. * Clean up cmakelists file * clang-format * Change bias bnorm clamp tests to use monotone initialization values to avoid tiny off-integer gemm results on RDNA3 from blowing up. * Renaming some instance lists and add functions to be more standardized. * Commented out non default instances. --------- Co-authored-by: kiefer <kiefer.van.teutem@streamhpc.com> |
||
|
Erwin Terpstra
|
d5ae81b292 |
Implement batched gemm add relu gemm add for rdna4 (#3391)
* 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> |
||
|
Wojciech Laskowski
|
b09121f860 |
WMMA support for batched_gemm_reduce (#3332)
Summary: - added new device impl of Batched GEMM Reduce for WMMA - added instance library - added WMMA impl to the Batched GEMM Reduce tests |
||
|
Bartłomiej Kocot
|
0727e85e52 |
[CK_BUILDER] Add grouped conv fwd ck tile profiler (#3518)
* [BULDER] Add grouped conv fwd ck tile profiler * [CK TILE] Fix grouped conv kernels splitk and double lds * Updates * Fixes * Move to ckProfiler * Fixes * fix * fix * Change instances to empty list by default * fix * fix * Update grouped_convolution_signatures.hpp * Update grouped_convolution_forward_tile_algs.hpp * [CK TILE] Add grouped convolution forward tests (#3556) * [CK TILE] Add grouped convolution forward tests * fix jenkins * fixes * comments fixes * unit test * unit test fix * Move instances outside builder * fix includes * clang format fix * readme fix * fix includes * fixes |
||
|
Erwin Terpstra
|
fe40a5d139 |
Implement batched gemm bias permute for RDNA4 (#3534)
* 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 |
||
|
Johannes Graner
|
3f735c127b |
[CK Profiler] Restore CPU tensor initialization when verification is not done on GPU (#3594)
* Fix large case init bounds
* Revert "Fix large case init bounds"
This reverts commit
|
||
|
Johannes Graner
|
f173642087 |
[CK] Refactor GPU verification kernel to gather error stats on GPU (#3551)
* Refactor GPU verification kernel to gather erorr stats on GPU * Check if result is all zero * non-negative error count doesn't need custom Atomics * Remove unnecessary AtomicMaxFloat function * Simpler warp reduction, remove passed flag * Move verification header to include * Fix header path in test * Fix block reduction loop |
||
|
Johannes Graner
|
3ccb15ea02 |
[CK Profiler] Initialize tensors on GPU in CK profiler (#3550)
* Initialize tensors on GPU in CK profiler * Kick CI |
||
|
Erwin Terpstra
|
eb041079a3 |
Implement grouped gemm tile loop for RDNA4 (#3304)
* 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 |
||
|
Johannes Graner
|
18c2ff6019 |
[CK profiler] Perform verification on GPU when using GPU reference (#3482)
* Simple verification kernel for ckProfiler * Verification kernel unit tests * Explicit synchronization * Address review comments |
||
|
Enrico Degregori
|
aad4cf0985 |
Wmma support for gemm_bias_add_reduce (#3316)
* Add tests for gemm_bias_add_reduce * Initial working implementation * Generalize implementation of reduce epilogue * Add tests for all layouts * Add instances * Fix test archs * Fix xdl bug * Remove library/profiler duplications * Fix num_byted error profiler * Fix typos * Fix copyright |
||
|
Erwin Terpstra
|
f9c6ba0403 |
Implement grouped gemm fastgelu for RDNA4 (#3303)
* Implement grouped gemm fastgelu for RDNA4 * chore: some cleanup and minor inconsistencies in grouped gemm profiler * chore: clarified logic and reporting of supported instance warnings |
||
|
Wojciech Laskowski
|
0fd2b2f045 |
Adding support for scale and bilinear ops for WMMA grouped conv fwd (#3450)
* Updated the set of tests for FP16 * Fix typo * Moved f16xi4 test under the correct data layout group * example for gemm_universal_bf16 * Adding examples for gemm_wmma instances * Added the missing parameters * Fixed review comments and added executable to cmakeLists * Fixing clang format * Fixing build erros * Fixed compilation failure. * Modified some code as per gemm_universal_examples * Fixed the gemm specialization error * Fixed the build errors. * Fix strides of a/b_thread_desc The descriptors are larger than needed (even though the compiler don't alloc registers for unused values). * Load in M/NRepeat dims with thread copy's slice instead of a loop * Clone BlockwiseGemmXdlops_pipeline_v1 for WMMA implementation * Implement Intrawave and Interwave variants of pipeline v1 * Add instances for Interwave and Intrawave v1 * Add instances with ABlockLdsExtraM and BBlockLdsExtraN = 0 * Remove instances that are too slow (mostly because of register spilling) * Add a workaround for fp8/bf8->f32 packed conversion issue * Add instances for Interwave and Intrawave v1 * Enable profiling of mixed precision with f8 and int4 on WMMA * Fix segfault in profiler when B is pk_i4_t b_device_buf's size in bytes is larger than b_k_n_permute so b_device_buf.ToDevice reads out-of-bounds. * Remove instances that are too slow (mostly because of register spilling) * Add missing add_device_gemm_wmma_universal_f8_f8_bf16 declarations * Add test case for bf16_i4 * Add missing Regular tests * Add test_gemm_universal_xdl/wmma_fp16 to REGRESSION_TESTS They take more than 30 seconds * Fix a bug that fp16_i4 validation passes only with PermuteB A permutation required by conversion from pk_i4_t to half_t does not depend on PermuteB, they can be used independently. * Use PermuteB with f16_i4 in most instances (as xdl) Some instances use PermuteB = false for checking correctness. See also the previous commit. * Fix cache flushing for pk_i4 * Add mixed precision examples * Disable all tests and instances with f8 on gfx11 Even though f8_f16 and f16_f8 don't require f8 WMMA instructions, gfx11 still lacks hardware instructions for fast f8->f32 conversion. * Add FP16 KM_NK and KM_KN test suites for XDL These tests were added to common .inc for better testing of WMMA instances * Support multiple D in GridwiseGemm_wmma_cshuffle_v3 DeviceGemm_Wmma_CShuffleV3 is changed for new template parameters. * Use ThreadGroupTensorSliceTransfer_v7r3 * Clone for device_gemm_wmma_cshuffle_v3.hpp for future Multiple D support * Clone example/65_gemm_multiply_multiply/gemm_add_add_xdl_fp16.cpp for wmma * Implement DeviceGemmMultipleD_Wmma_CShuffleV3 * Make gemm_add_add_wmma to work with DeviceGemmMultipleD_Wmma_CShuffleV3 * Prepare gemma_add tests for adding wmma * Add gemm_add_fastgelu instances and test * Add a special wrapper to use DeviceGemmMultipleD_Wmma_CShuffleV3 with old API ckProfiler uses DeviceGemmMultipleD (tests also call its functions), the wrapper allows to use DeviceGemmMultipleDSplitK instances there. * removed unnecessary ck parts from compilation * initial gemm_add_multiply instance implementations * fixed profiler help message for gemm_add_multiply * improved multiply_add profiler layout help * fixed template arguments for test instances * added test for gemm_add_multiply * Support multiple D in GridwiseGemm_wmma_cshuffle_v3 DeviceGemm_Wmma_CShuffleV3 is changed for new template parameters. * Use ThreadGroupTensorSliceTransfer_v7r3 * Clone for device_gemm_wmma_cshuffle_v3.hpp for future Multiple D support * Clone example/65_gemm_multiply_multiply/gemm_add_add_xdl_fp16.cpp for wmma * Implement DeviceGemmMultipleD_Wmma_CShuffleV3 * Make gemm_add_add_wmma to work with DeviceGemmMultipleD_Wmma_CShuffleV3 * Prepare gemma_add tests for adding wmma * Add gemm_add_fastgelu instances and test * Add a special wrapper to use DeviceGemmMultipleD_Wmma_CShuffleV3 with old API ckProfiler uses DeviceGemmMultipleD (tests also call its functions), the wrapper allows to use DeviceGemmMultipleDSplitK instances there. * switched to splitK interface * log print added to splitk benchmarks * revert main cmake comments * newline change reverted * added add_fastgelu instances * revert unintended change in xdl add_fastgelu * created gemm_add_add_fastgelu instances * created fastegelu instances * added tests for all splitk fastgelus * Added tests. * multiply_add instances created * updates to add_multiply splitk instances * splitk xdl test fixes * added wmma multiply_multiply instances * fixed ONLY_XDL_AND_WMMA_KERNELS tag * Added gemm_add examples for wmma v1 and v3 * fixed / workarounded i8 instances * Modified the v3 code to added one fp16 bxdl instance. * added bf16 xdl instance. * adding gemm_add wmma_cshuffle and other support (cherry picked from commit ec447e7f564095ea969eddc39ec77b843aa52976) Co-authored-by: Cenxuan <cenxuan@streamhpc.com> * add instances into camkelists (cherry picked from commit 23bf2d2771c939ea3ca7f493433c55255bffd08e) Co-authored-by: Cenxuan <cenxuan@streamhpc.com> * This is work in progress, edited the template parameters in order to build (cherry picked from commit b4fde8a3314cb44659c4bbda35f1a0133c63dc41) Co-authored-by: Cenxuan <cenxuan@streamhpc.com> * temp work saved, changed the BDataType to f16 or bf16 since wmma currently not support non-equal A and B datatype (cherry picked from commit 22fbd68f1db458ab50780a394ee2544c7a1484d1) Co-authored-by: Cenxuan <cenxuan@streamhpc.com> * added datatype and use clang-format-12 (cherry picked from commit ae4e853682ef1bb27784b2f965b4a66b3751ceec) Co-authored-by: Cenxuan <cenxuan@streamhpc.com> * Fixing build errors * Added instances for v3 * Adding instances and executables * Code update of template parameters modified. * Renamed file. * Added tests. * resolved error tests. * Fixing build errors * Updated comments * removed the changes as per the MR review comment. * Updated tests. * fp8 instances - not tested * Restored the Cmake file that was reverted by mistake during rebase. * fixed wmma_op test * Updated comments. * Updated the template parameter description * fixed rdna4 instances * fixed back compatibility on gfx11 * cleanups * fix ckProfiler * one more cmake fix * added fp8 instances * Updated tests to ad BF16 instances as per review comment * Added include file and cleaned up(as per review comment) * Updated and optimized the example code for all types. * Fixed clang format * Resolve "Implement `device_gemm_bilinear` for RDNA4" * test generalization to handle FP16 shuffle better * added missing changes * Added bf16 wmma instance for add_relu * Added f16 wmma instance and corrected bf16 instance errors. * Added instances to Cmake * Modified the template parameters to make the instances work. * Fixed typo in profiler * Added v3 instances for gemm_add_relu * addressed core review comments * Added test for gemm_add_relu wmma instance * Cleaned up the code. * Added examples for gemm_add_relu * Fixing typo to resolve build errors. * Fixes applied to fix the precision loss. * fix billinear test after merge * Removed the old wmma instances. * Added wrapper and renamed the wmma_v3 instances * Updated copyrights and added wrappers. * Fixes applied according to review comments * Apply 1 suggestion(s) to 1 file(s) Co-authored-by: Robin Voetter <robin@streamhpc.com> * Removed the old wmma instances. * Updated wrapper for the v3 instances * removed the old wmma examples * Renamed the v3 instances * Deleted the gtest file added by mistake. * Updated thge profiler with wrapper * Fixed test errors. * Fixed the review comments * Fixed the if condition MACROS. * REVERTED THE PROFILER CHANGES * Revert "REVERTED THE PROFILER CHANGES" This reverts commit |
||
|
Kiefer van Teutem
|
2ea710e88b |
Grouped convolution forward device implementation and base flavors for RDNA3/4 (#2964)
* Fixed typos for padded instances * Added tests for fp16, KM_KN and KM_NK * Padding not supported for when BDataType is pk_i4_t. Added fix for correct check and removed padding instances. * Fixed typos * Updated the set of tests for FP16 * Updated the set of tests for FP16 * Fix typo * Moved f16xi4 test under the correct data layout group * example for gemm_universal_bf16 * Adding examples for gemm_wmma instances * Added the missing parameters * Fixed review comments and added executable to cmakeLists * Fixing clang format * Fixing build erros * Fixed compilation failure. * Modified some code as per gemm_universal_examples * Fixed the gemm specialization error * Fixed the build errors. * Fix strides of a/b_thread_desc The descriptors are larger than needed (even though the compiler don't alloc registers for unused values). * Load in M/NRepeat dims with thread copy's slice instead of a loop * Clone BlockwiseGemmXdlops_pipeline_v1 for WMMA implementation * Implement Intrawave and Interwave variants of pipeline v1 * Add instances for Interwave and Intrawave v1 * Add instances with ABlockLdsExtraM and BBlockLdsExtraN = 0 * Remove instances that are too slow (mostly because of register spilling) * Add a workaround for fp8/bf8->f32 packed conversion issue * Add instances for Interwave and Intrawave v1 * Enable profiling of mixed precision with f8 and int4 on WMMA * Fix segfault in profiler when B is pk_i4_t b_device_buf's size in bytes is larger than b_k_n_permute so b_device_buf.ToDevice reads out-of-bounds. * Remove instances that are too slow (mostly because of register spilling) * Add missing add_device_gemm_wmma_universal_f8_f8_bf16 declarations * Add test case for bf16_i4 * Add missing Regular tests * Add test_gemm_universal_xdl/wmma_fp16 to REGRESSION_TESTS They take more than 30 seconds * Fix a bug that fp16_i4 validation passes only with PermuteB A permutation required by conversion from pk_i4_t to half_t does not depend on PermuteB, they can be used independently. * Use PermuteB with f16_i4 in most instances (as xdl) Some instances use PermuteB = false for checking correctness. See also the previous commit. * Fix cache flushing for pk_i4 * Add mixed precision examples * Disable all tests and instances with f8 on gfx11 Even though f8_f16 and f16_f8 don't require f8 WMMA instructions, gfx11 still lacks hardware instructions for fast f8->f32 conversion. * Add FP16 KM_NK and KM_KN test suites for XDL These tests were added to common .inc for better testing of WMMA instances * Support multiple D in GridwiseGemm_wmma_cshuffle_v3 DeviceGemm_Wmma_CShuffleV3 is changed for new template parameters. * Use ThreadGroupTensorSliceTransfer_v7r3 * Clone for device_gemm_wmma_cshuffle_v3.hpp for future Multiple D support * Clone example/65_gemm_multiply_multiply/gemm_add_add_xdl_fp16.cpp for wmma * Implement DeviceGemmMultipleD_Wmma_CShuffleV3 * Make gemm_add_add_wmma to work with DeviceGemmMultipleD_Wmma_CShuffleV3 * Prepare gemma_add tests for adding wmma * Add gemm_add_fastgelu instances and test * Add a special wrapper to use DeviceGemmMultipleD_Wmma_CShuffleV3 with old API ckProfiler uses DeviceGemmMultipleD (tests also call its functions), the wrapper allows to use DeviceGemmMultipleDSplitK instances there. * removed unnecessary ck parts from compilation * initial gemm_add_multiply instance implementations * fixed profiler help message for gemm_add_multiply * improved multiply_add profiler layout help * fixed template arguments for test instances * added test for gemm_add_multiply * Support multiple D in GridwiseGemm_wmma_cshuffle_v3 DeviceGemm_Wmma_CShuffleV3 is changed for new template parameters. * Use ThreadGroupTensorSliceTransfer_v7r3 * Clone for device_gemm_wmma_cshuffle_v3.hpp for future Multiple D support * Clone example/65_gemm_multiply_multiply/gemm_add_add_xdl_fp16.cpp for wmma * Implement DeviceGemmMultipleD_Wmma_CShuffleV3 * Make gemm_add_add_wmma to work with DeviceGemmMultipleD_Wmma_CShuffleV3 * Prepare gemma_add tests for adding wmma * Add gemm_add_fastgelu instances and test * Add a special wrapper to use DeviceGemmMultipleD_Wmma_CShuffleV3 with old API ckProfiler uses DeviceGemmMultipleD (tests also call its functions), the wrapper allows to use DeviceGemmMultipleDSplitK instances there. * switched to splitK interface * log print added to splitk benchmarks * revert main cmake comments * newline change reverted * added add_fastgelu instances * revert unintended change in xdl add_fastgelu * created gemm_add_add_fastgelu instances * created fastegelu instances * added tests for all splitk fastgelus * Added tests. * multiply_add instances created * updates to add_multiply splitk instances * splitk xdl test fixes * added wmma multiply_multiply instances * fixed ONLY_XDL_AND_WMMA_KERNELS tag * Added gemm_add examples for wmma v1 and v3 * fixed / workarounded i8 instances * Modified the v3 code to added one fp16 bxdl instance. * added bf16 xdl instance. * adding gemm_add wmma_cshuffle and other support (cherry picked from commit ec447e7f564095ea969eddc39ec77b843aa52976) Co-authored-by: Cenxuan <cenxuan@streamhpc.com> * add instances into camkelists (cherry picked from commit 23bf2d2771c939ea3ca7f493433c55255bffd08e) Co-authored-by: Cenxuan <cenxuan@streamhpc.com> * This is work in progress, edited the template parameters in order to build (cherry picked from commit b4fde8a3314cb44659c4bbda35f1a0133c63dc41) Co-authored-by: Cenxuan <cenxuan@streamhpc.com> * temp work saved, changed the BDataType to f16 or bf16 since wmma currently not support non-equal A and B datatype (cherry picked from commit 22fbd68f1db458ab50780a394ee2544c7a1484d1) Co-authored-by: Cenxuan <cenxuan@streamhpc.com> * added datatype and use clang-format-12 (cherry picked from commit ae4e853682ef1bb27784b2f965b4a66b3751ceec) Co-authored-by: Cenxuan <cenxuan@streamhpc.com> * Fixing build errors * Added instances for v3 * Adding instances and executables * Code update of template parameters modified. * Renamed file. * Added tests. * resolved error tests. * Fixing build errors * Updated comments * removed the changes as per the MR review comment. * Updated tests. * fp8 instances - not tested * Restored the Cmake file that was reverted by mistake during rebase. * fixed wmma_op test * Updated comments. * Updated the template parameter description * fixed rdna4 instances * fixed back compatibility on gfx11 * cleanups * fix ckProfiler * one more cmake fix * added fp8 instances * Updated tests to ad BF16 instances as per review comment * Added include file and cleaned up(as per review comment) * Updated and optimized the example code for all types. * Fixed clang format * Resolve "Implement `device_gemm_bilinear` for RDNA4" * test generalization to handle FP16 shuffle better * added missing changes * Added bf16 wmma instance for add_relu * Added f16 wmma instance and corrected bf16 instance errors. * Added instances to Cmake * Modified the template parameters to make the instances work. * Fixed typo in profiler * Added v3 instances for gemm_add_relu * addressed core review comments * Added test for gemm_add_relu wmma instance * Cleaned up the code. * Added examples for gemm_add_relu * Fixing typo to resolve build errors. * Fixes applied to fix the precision loss. * fix billinear test after merge * Removed the old wmma instances. * Added wrapper and renamed the wmma_v3 instances * Updated copyrights and added wrappers. * Fixes applied according to review comments * Apply 1 suggestion(s) to 1 file(s) Co-authored-by: Robin Voetter <robin@streamhpc.com> * Removed the old wmma instances. * Updated wrapper for the v3 instances * removed the old wmma examples * Renamed the v3 instances * Deleted the gtest file added by mistake. * Updated thge profiler with wrapper * Fixed test errors. * Fixed the review comments * Fixed the if condition MACROS. * REVERTED THE PROFILER CHANGES * Revert "REVERTED THE PROFILER CHANGES" This reverts commit |
||
|
Johannes Graner
|
bb8445dca8 |
[CK] Integrate GPU reference into ckProfiler for convolutions (#3379)
Refactor and integrate CK GPU references into ckProfiler. - All convolution layouts and groupings supported for all three directions - Unit tests verifying GPU and CPU reference is the same - Support added to profiler (do_verification = 2 enables GPU reference) - One profiler-based test per direction changed to GPU reference to demonstrate usag Closes AICK-427 |
||
|
Enrico Degregori
|
87dd073887 |
Wmma support for grouped convolution bwd weight (#2947)
* Convolution bwd weight device implementation
* Merge branch 'grouped_conv_bwd_weight_device_impl_wmma' into 'feature/conv_bwd_weight_wmma'
Convolution bwd weight device implementation
See merge request amd/ai/composable_kernel!38
* Fix bug and disable splitK=-1 tests for wmma
* Add generic instances for bf16 f32 bf16
* check gridwise level validity in device impl for 1 stage D0
* Fix bugs in device implementation:
- rdna3 compilation error
- gridwise layouts (need to be correct to ensure that CheckValidaity()
works correctly)
* Add padding in conv to gemm transformers for 1x1Stride1Pad0 specialization
* Remove workaround for 1x1Stride1Pad0 conv specialization
* Add instances for xdl parity (for pipeline v1)
* Add two stage instances (xdl parity)
* Add multiple Ds instances
* Add examples
* Uncomment scale instances
* Fix copyright
* Fix examples compilation
* Add atomic add float4
* Fix compilation error
* Fix instances
* Compute tolerances in examples instead of using default ones
* Compute tolerances instead of using default ones in bilinear and scale tests
* Merge branch 'grouped_conv_bwd_weight_instances_examples' into 'feature/conv_bwd_weight_wmma'
Grouped conv: Instances and example bwd weight
See merge request amd/ai/composable_kernel!47
* Device implementation of explicit gemm for grouped conv bwd weight
Based on batched gemm multiple D
* Add instances for pipeline v1 and v3
* Add support for occupancy-based splitk
* Fix ckProfiler dependencies
* Review fixes
* Merge branch 'explicit_bwd_weight' into 'feature/conv_bwd_weight_wmma'
Device implementation of explicit gemm for grouped conv bwd weight
See merge request amd/ai/composable_kernel!52
* Fix cmake file for tests
* fix clang format
* fix instance factory error
* Adapt all grouped conv bwd weight vanilla Xdl instances to 16x16. MRepeat doubled for all but 12 of them (some static assert failure). Also added custom reduced profiler target for building grouped conv bwd weight vanilla only profiler. Verified with gtest test.
* Revert "Adapt all grouped conv bwd weight vanilla Xdl instances to 16x16. MRepeat doubled for all but 12 of them (some static assert failure). Also added custom reduced profiler target for building grouped conv bwd weight vanilla only profiler. Verified with gtest test."
This reverts commit
|
||
|
Enrico Degregori
|
ce99cab605 |
Wmma support for gemm_ab_scale (#3314)
* Support gemm_ab_scale: - Add tests - Integrate scaling implementation in multiple D - Generalize existing b_scale for ab_scale - Add instances - Generalize implementation for ScaleBlockM, ScaleBlockN, ScaleBlockK - Add support for all layouts supported by xdl - Fix splitk xdl * Fix copyright * Wmma support for gemm_blockscale_wp (#3315) * Support for preshuffle with ab scale - add support for b preshuffle in GridwiseGemm_wmma_cshuffle_v3_ab_scale - add support for AScaleLayout amnd BScaleLayout (can be different from ALayout and BLayout, respectively) - add Run method in v1 pipeline to support preshuffle + scaling - add support for preshuffle gemms in common invoker - Add splitk support * Fix copyright header |
||
|
Bartłomiej Kocot
|
82f796a1f0 | Profile resnet layout fixes (#3360) | ||
|
Enrico Degregori
|
161835533b |
Wmma support for gemm_multiply_multiply_wp (#3278)
* Initial implementation with splitK support * Add gfx11 support * Fix compilation error * Add instances * Add irregular instances * Fix GetBuffer arguments * Minor changes * Address review comments * Fix compilation errors * Fix copyright header |
||
|
Erwin Terpstra
|
46f1d740f0 |
Add grouped gemm instances for RDNA4 (#3237)
* wip: grouped_gemm implementation based on wmma kernel + example for fp16 * chore: clean up grouped_gem_wmma_splitk_fp16 example * chore: add cmake options to fully disable XDL or WMMA kernels * feat: add tests for grouped gemma wmma instances for f16 and bf16 (all layouts) * chore: add grouped gemm wmma bf16 example * refactor: reuse more code between instance factory functions * chore: turn test failure if not all batch sizes are supported into a warning * chore: made failing of test on unsupported instances conditional to not break old tests * chore: add log message to failure case where AK1/BK1/KBatch is too high for K value * fix: issue with new overloads of GridwiseGemm_wmma_cshuffle_v3::Run() * fix: stray comma after parameter list * fix: compilation issues on RDNA3 and tests failing due to unsupported problems still being ran * chore: update copyright in header comments * nit: minor feebdack * refactor: unified XDL / wma tests * fix: properly disable FP8 instances when ONLY targeting gfx11 * refactor: add v3 suffix to grouped_gemm device struct name * fix: small typos in example code * fix: fully exclude xdl/wmma instances when using the corresponding cmake flags * chore: remove unused destructor and added pipeline support checks to remove unnecessary paths * fix: make sure to not add instance library to group if library was skipped * fix: make sure xdl grouped gemm doesnt fail the new test * fix: explicitly exclude test if no xdl/wmma support, as pattern matching fails in this case * fix: examples not working since dependent types and functions were moved to ck namespace in develop * fix: tests failing when compiling for just gfx11 due to trying to run unsupported instances * chore: replace/add copyright headers with new format |
||
|
Aviral Goel
|
0aadb4b2c4 |
chore(copyright): update copyright header for profiler directory (#3205)
* chore(copyright): update copyright header for tile_engine directory * chore(copyright): update copyright header for script directory * chore(copyright): update copyright header for test_data directory * chore(copyright): update copyright header for python directory * chore(copyright): update copyright header for profiler directory |
||
|
jefyang1
|
d30babbd00 | Add new gemm multiply multiply instances on gfx950 (#3213) | ||
|
Enrico Degregori
|
7414a0f4d4 |
Wmma support for gemm_reduce (#3145)
* Initial implementation GEMM+Reduce: - device struct - epilogue struct * Fix tests, improve profiler and add initial instances * Add instances * Fix compilation error * Address review comments * Fix logging --------- Co-authored-by: Illia Silin <98187287+illsilin@users.noreply.github.com> |
||
|
Enrico Degregori
|
507d81c3af |
Fix splitk preshuffle (#3137)
* Fix splitK multiply_multiply_wp * Add tests for gemm_multiply_multiply_wp * Add tests for gemm_universal_preshuffle (KBatch = 1) * Add tests gemm_blockscale_wp * Fix splitk gemm universal preshuffle * Run new tests on arch supporting fp8 * Restore example * Fix strides profiler * Fix tests * Fix clang format * Finalize profiler preshuffle with tolerances * Minor improvements to splitk related changes * Address review comments: clang format and ckProfiler typo * Remove b_k_split_offset from SplitKBatchOffset struct |
||
|
Enrico Degregori
|
4ebc48a3cd |
WMMA gemm_add_relu_add_layernorm (#2989)
* Summary:
- Refactor epilogue (with CShuffle) to support fused operations:
- EpilogueCShuffleBase holds common parts
- EpilogueCShuffle: runs CShuffle and write out
- EpilogueWelfordCShuffle: holds Welford specific arguments, runs CShuffle, write out, Welford first part and Welford write out
- Extend thread transfer v7r3:
- Support for intermediate data type different from src and dst type
- New functionality to write to dst buffer and keep data (to be able to use them for additional operations)
* Adress review comments
|
||
|
Enrico Degregori
|
06973b1cf4 | Fix multi-abd tests bug (#3099) | ||
|
kabrahamAMD
|
c4b2da9cbd |
implement device batched gemm b scale for wmma (#2825)
* rebased on top of develop * fixed missing shuffeling and wrong indexing * added tests for batched_b_scale * added missing files * fixed wrong stride computation and removed k batching (for now) due to precision issues * reinstated k-batching with PRNG constrained to -1..1 * added specialization of GeneratorTensor_3 for int4 and fixed internal overflow * added k-batching to reference and increased tolerances for test * changed gemm_b_scale and gemm_universal tests to use correct parameters * adressed review commentsd * ported fixes back to non-batched version of b_scale * adressed review comments * run clang-format on older commits * add type-conversion to AccDataType and then to CDataType to exactly mimic GPU's behavior * added newline at end of file * reflected changes from muitl-abd branch in batched b_scale * fixed gfx11 issue * changed range for pki4 to -1...1 (-0.5...0.5 never really made sense for i4 anyway and always should have caused compiler errors, but since there was no int4 specialization of GeneratorTensor3 until now, this passed * run clang format * set range of i4 generation to 0...1 for upstream tests to pass. This replicated previous behavior, which however means that it is NOT properly tested. * reduced range for pk_i4 even further to 0..0 * removed failing xld instances. Failure now uncovered now that tests were fixed * removed generation of int4 values entierly * divide B buffer by BPackedSize --------- Co-authored-by: Kevin Abraham <kevin.abraham@streamhpc.com> |
||
|
Illia Silin
|
3348f01e6f |
re-enable clang-format by default (#3030)
* re-enable clang-format by default * fix clang format |
||
|
yinglu
|
fada1a3cae |
Conv:TF32: add more instances - 2 (#2879)
* add instances of device_grouped_conv_fwd_xdl_f32_comp_instances * add instances of device_grouped_conv_fwd_xdl_f32_tf32_mem_instances * add instances of device_grouped_conv_fwd_xdl_large_tensor_f32_tf32_instances * tf32:conv:add instances for base class DeviceConvFwd * tf32:conv:add instances for base class DeviceGroupedConvBwdDataMultipleD * tf32:conv:add instances for base class DeviceGroupedConvBwdWeight * add tf32 in profiler * remove gnhwc/ngchw/ngcdhw instances * remove non-ndhwgc/nhwgc/nhwc instances * add check in IsSupportedArgument() |
||
|
Sami Remes
|
9d4bfe3932 |
Add KBatch support for gemm_ab_scale (#2740)
* Add KBatch support for gemm_ab_scale
* Revert kernel parameters change
* Remove printing
* fix formatting
* fix check
* Use {} in if
---------
Co-authored-by: Adam Osewski <19374865+aosewski@users.noreply.github.com>
|
||
|
linqunAMD
|
e78a897ec0 |
[CK] Add command option instance_index and param_mask to run partial ck test (#2889)
* [CK] Add command option instance_index and param_mask to run partial ck test Many CK test are instance test. it will loop all instance in the instance library. It causes test often out-of-time if we run test on simulator/emulator. This PR add option instance_index and param_mask to reduce the workload of instance test instance_index: only run test 1 available instance with specified index. param_mask: filter the embedded parameter with specified mask * fix CI error * fix clang format --------- Co-authored-by: illsilin_amdeng <Illia.Silin@amd.com> |
||
|
Bartłomiej Kocot
|
5477811670 |
Grouped Conv Bwd Data out index calculation optimizations (#2917)
* Grouped Conv Bwd Data index calculation optimizations * fixes * refactor instances * gfx12 fixes * temporary disable splitK for gfx12 |
||
|
emezh
|
2aa06fbd45 | fix copy-paste bug in get_matrix_b; re-enable all tests in multi_abd (#2939) | ||
|
emezh
|
db2524be2d |
Verify HostTensorDescriptor when it is created (#2829)
* add proper GEMM layout verification * Handle "auto" strides. CalculateStrides only called when tensor's strides are empty or all of them are <=0 (auto strides). CalculateStrides now supports GEMM::ColumnsMajor order. The assumption is still that it applies only to the inner two dims. ValidateStrides throws if any of the tensor's strides is <=0. profile_gemm_multiply_add updated to support "auto" strides for tensors. Manual tests for profile_gemm_multiply_add (matrix B in Row and Col modes) auto-strides bin/ckProfiler gemm_multiply_add 0 0 1 1 0 1 128 128 128 0 0 0 0 0 bin/ckProfiler gemm_multiply_add 0 1 1 1 0 1 128 128 128 0 0 0 0 0 bin/ckProfiler gemm_multiply_add 0 0 1 1 0 1 128 128 128 -1 -1 -1 -1 -1 Note, -1 should be deprecated (use 0 instead) explicit strides (same as auto) bin/ckProfiler gemm_multiply_add 0 0 1 1 0 1 128 128 128 128 128 128 128 128 bin/ckProfiler gemm_multiply_add 0 1 1 1 0 1 128 128 128 128 128 128 128 128 explicit strides (not the same as auto) bin/ckProfiler gemm_multiply_add 0 0 1 1 0 1 128 128 128 130 132 134 136 138 bin/ckProfiler gemm_multiply_add 0 1 1 1 0 1 128 128 128 130 132 134 136 138 mix of explicit and auto strides bin/ckProfiler gemm_multiply_add 0 0 1 1 0 1 128 128 128 128 128 128 128 0 invalid stride bin/ckProfiler gemm_multiply_add 0 0 1 1 0 1 128 128 128 0 0 0 0 64 terminate called after throwing an instance of 'std::runtime_error' what(): Invalid strides for RowMajor: mLens: 128 128 , mStrides: 64 1 Aborted (core dumped) * - add more names to ck::tensor_layout for easier namespace hierarchy checking - updated convolutional layouts to use explicit ones or BaseConvolutionalLayout where it is not clear which layout to use (TBD) - see include/ck/library/utility/convolution_host_tensor_descriptor_helper.hpp * added handling of partially initialized strides for GEMM. fixed more tests. * clang-format and more fixes * replace long dash by a simple hyphen - causes build failure in CK codegen. * increase sizeof input, otherwise output size becomes zero or negative with large filter size * select stride based on layout * specify layout explicitly to avoid errors in HostTensorDescriptor creation * add validation for higher GEMM tensor dimensions.; Add docstring to `HostTensorDescriptor` * Not clear why permute test in test/permute_scale/test_permute_scale.cpp uses a lot of invalid strides. Setting layout to BypassLayoutVerification to avoid a lot of errors * fix test (incl removing invalid config) * fix moe examples: - (in .cpp) add layout argument to non-2D tensors - (in .hpp) fix asserts/failures that show up in Debug mode, specifically addressing 2D tensor by a single index (and 3D tensor by 2d index) * fix moe_gemm2 example. * fix profile and wmma examples * clean-up early mods for ckprofile. verified with: ``` ckProfiler gemm_multiply_add 0 0 1 1 0 1 128 128 128 0 0 0 0 0 ckProfiler gemm_multiply_add 0 1 1 1 0 1 128 128 128 0 0 0 0 0 ckProfiler gemm_multiply_add 0 0 1 1 0 1 128 128 128 130 132 134 136 138 ckProfiler gemm_multiply_add 0 1 1 1 0 1 128 128 128 130 132 134 136 138 # ckProfiler gemm_fastgelu 1 0 1 2 0 1 128 128 128 0 0 0 ckProfiler gemm_fastgelu 1 1 1 2 0 1 128 128 128 0 0 0 ckProfiler gemm_fastgelu 1 2 1 2 0 1 128 128 128 0 0 0 ckProfiler gemm_fastgelu 1 3 1 2 0 1 128 128 128 0 0 0 ckProfiler gemm_fastgelu 1 0 1 2 0 1 128 128 128 128 128 128 # ckProfiler gemm_add_relu 0 0 1 1 0 1 128 128 128 0 0 0 0 # ckProfiler gemm_add_relu 0 1 1 1 0 1 128 128 128 0 0 0 0 # not implemented # ckProfiler gemm_add_relu 0 2 1 1 0 1 128 128 128 0 0 0 0 # not implemented # ckProfiler gemm_add_relu 0 3 1 1 0 1 128 128 128 0 0 0 0 # not implemented ckProfiler gemm_add_relu 0 0 1 1 0 1 128 128 128 128 128 128 128 # ckProfiler gemm_add_relu_add_layernorm 1 0 1 1 0 0 128 128 128 0 0 0 0 0 ckProfiler gemm_add_relu_add_layernorm 1 1 1 1 0 0 128 128 128 0 0 0 0 0 ckProfiler gemm_add_relu_add_layernorm 1 2 1 1 0 0 128 128 128 0 0 0 0 0 ckProfiler gemm_add_relu_add_layernorm 1 3 1 1 0 0 128 128 128 0 0 0 0 0 ckProfiler gemm_add_relu_add_layernorm 1 0 1 1 0 0 128 128 128 130 132 134 136 138 # example_gemm_add_multiply_dl_fp16 example_gemm_add_multiply_xdl_fp16 # ckProfiler gemm_blockscale_wp 7 1 1 1 1 0 1 128 128 128 0 0 0 ckProfiler gemm_blockscale_wp 7 1 1 1 1 0 1 128 128 128 128 128 128 ``` * temporary skip first 8 test configs - they throw error * temporary skip first 8 test configs in wmma too - they throw error --------- Co-authored-by: Illia Silin <98187287+illsilin@users.noreply.github.com> |
||
|
Enrico Degregori
|
3d29bff2f0 |
Wmma support for multiple ABD GEMM (#2803)
* multi_abd wmma support: - Add multiple A and B support to multiple D implementation (gridwise level) - Add multi_abd GEMM (device level) - Add instances (xdl parity) - Add tests (both xdl and wmma) - Add examples - Add ckProfiler support (both xdl and wmma) * Fix bug in device print function * Fix unused template parameter * Fix batched gemm for multiABD gridwise implementation * Fix gemm_universal_reduce with multiABDs gridwise implementation --------- Co-authored-by: Illia Silin <98187287+illsilin@users.noreply.github.com> |
||
|
Wojciech Laskowski
|
f97b2a3f5d |
Added wmma support for gemm quantization: (#2841)
- profiler for gemm quantization for DL/XDL - tests for gemm quantization for DL/XDL - implementation for gemm quantization for WMMA - profiler/tests for gemm qunatization for WMMA Co-authored-by: Illia Silin <98187287+illsilin@users.noreply.github.com> |
||
|
Wojciech Laskowski
|
b25d4d684a |
WMMA support for GEMM reduce (#2823)
Added gemm + reduce instance library for RDNA4. This includes: - New device implementation running GEMM and reduction kernel - instances for wmma (xdl parity) - examples for wmma (xdl parity) - tests for existing xdl and wmma |
||
|
Enrico Degregori
|
b740380906 |
Wmma support for multiple Ds based GEMMs (#2613)
* Fixed cmake errors related to gemm_bilinear. Previously, if the above flags are set, cmake build fails: GPU_TARGETS="gfx1100;gfx1201" -D DTYPES="fp16;bf16;fp8" * Fixed cmake build errors related to test_fp8 * Updates to support mixed precision (cherry picked from commit e65d71180393e7b66169c56565a6bac740427de6) Co-authored-by: Anca Hamuraru <anca@streamhpc.com> * Adding support for RRR, F8xF16xF16 gemm_universal_wmma - wip (cherry picked from commit f8c06322df0abcbd5945a56cdf5bffe56480f9f0) Co-authored-by: Anca Hamuraru <anca@streamhpc.com> * Added support for F8xF16xF16 to gemm_wmma_universal (cherry picked from commit 15c851de6daa513a12c2e3af299bab0176175fb5) Co-authored-by: Anca Hamuraru <anca@streamhpc.com> * Added support for F16xF8xF16 to gemm_wmma_universal * Added support for BF16xI4xBF16 to gemm_wmma_universal (cherry picked from commit c6a4a69d2d43d59bae8bdabfae80d648646f217e) Co-authored-by: Anca Hamuraru <anca@streamhpc.com> * Added support for F16xI4xF16 to gemm_wmma_universal * Fixed IsSupportedArgument to check ComputeTypeA, ComputeTypeB instead of ADataType, BDataType * Added missing test class for FP16_KM_NK * Pre-commit hooks fixes * Added padding instances for f16xf16xf16 * Fixed cmake errors related to gemm_bilinear. Previously, if the above flags are set, cmake build fails: GPU_TARGETS="gfx1100;gfx1201" -D DTYPES="fp16;bf16;fp8" (cherry picked from commit |
||
|
Kiefer van Teutem
|
7330ec37ee |
Implement batched gemm gemm for RDNA (3 and 4) (#2612)
* Create new copies of existing device struct and gridwise struct for batched_gemm_softmax_gemm and disable the softmax part. Still based on old wmma pipelines. Also copy the example and remove the softmax part from the reference calculation. Works and results match reference except for tiny float errors in problem 2. * Turn DeviceBatchedGemmGemm_Wmma_CShuffleV3 into a proper DeviceBatchedGemmGemm derived class, with the right argument and invoker functions. Update example to use new definitions. * Remove unused cross-attention and self-attention kernels, arguments, and invokers. Also remove other unused Argument types. * Remove masking related code, test unusual sizes in example. * Remove remaining softmax related code from GridwiseBatchedGemmGemm_wmma_cshuffle_v3 and example. * Remove code related to numDims, bias, and TensorSpec from Device struct and example. * Add layout template parameters to device struct * Move (NPerBlock, LTilePerBlock) device struct template arguments up by two places to match XDL template argument ordering. * Merge accumulation data types into one type to match XDL device struct. * Remove NPerWmma template parameter from device struct and just set it equal to LPerWmma. Now device struct template params exactly match those for XDL batched gemm gemm. * Add support for RCCR layout and test this in example * Add batched_gemm_gemm_wmma to instance library + profiler, and add gtest just like for xdl. * Add RCCR instance and additional RCRR instance to library. * Remove unused permute and alpha related code. Time all tests. Fix B1 strides in argument verification. * Remove references to G0, G1 in favor of batch, reduce dimensionality of length and stride arrays. * Managed to replace old wmma gridwise pipeline and blockwise struct with new wmma blockwise pipeline. Some cleanup required but all tests pass. * Make TransposeC a proper template parameter that gets passed all the way from BlockGemmPipeline_Selector to WmmaGemm so we can use the correct settings for bacthed gemm gemm as well as regular gemm. Gemm universal tests now pass again. * Replace old LoopSched and PipelineVer params with BlockwiseGemm pipeline equivalents, and use these in instance factory. The v3 pipeline does not work yet, but v1 works for intrawave and interwave. * Adapt the A wave descriptor to deal with RDNA4 wmma. This fixes batched gemm gemm functionality on RDNA4. * Fixed two aspects of the v3 pipeline that were incorrect: First of all the blockwise copy operator was invoked once too many in all cases (RunRead and move window), which broke batched gemm gemm when the blockwise pipeline was used multiple times. Furthermore we should be using the mainloop (hotloop) for num_k_loop >=2 instead of num_k_loop >=3. Now we can use support any K dimension. * Remove num prefetch parameter from gridwise struct since we don't use it and it doesn't do anything, * Remove unused non-lds paths. * Test and update the IsSupportedArgument() and CheckValidity() functions for all layouts + padding modes and various problem sizes. * Add a lot of instances to the profiler with various blocksizes and pipelines, all verified. * Add support for BF16: instance library, tests, and examples. * Add examples for int8 and fp8, had to add type_convert_sp template specializations for the latter. * Template the library instance lists and add default padding instances. * Move memory calculations from the kernel to the Argument contructor. Also actually parse and use the user-provided batch strides. * Actually parse and use user-provided regular strides. * More refactor: remove references to multiple dims per dims, and g0 / g1. Also move xdl specific test utils out of generic test util header. * Small post-rebase-on-develop fix due to bscale-related pipeline changes. All tests rerun + tested bscale and regular gemm. * Introduce the correct GetCThreadDescriptor function in the blockwise gemm pipelines for the TransposeC=true case. It turns out to be identical for our batched gemm gemm (gemm0) usecases, but could theoretically be different for wmma_gemm instances with smaller-than-4-byte output data size. * Remove unused NumPrefetch template parameter, we don't need to match the XDL template params one-to-one. * Implement proper TailNum and HasMainLoop template parameters for the v3 pipeline. Now the Run() function knows at compile time whether there are 1, 2, or more loops in total, and adds or removes sections accordingly. It still uses the blockwise copy operators the correct amount of times. * Add print lambda with env check and file and func to device and gridwise level compatibility error messages. Also respect compatibility in example script. * RDNA3 does not support fp8 |
||
|
Yi DING
|
bab747b017 | Fix typo in profiler/include/profiler/profile_gemm_blockscale_wp_impl.hpp (#2767) | ||
|
mirchen-amd
|
60320e90c1 |
Mirchen/gemm blockscale wp segfault fix (#2638)
* Add stride validation to prevent segfault in blockscale GEMM
* run clang-format
* Update profiler/include/profiler/profile_gemm_blockscale_wp_impl.hpp
Co-authored-by: rahjain-amd <Rahul.Jain@amd.com>
* added stride length checking to more gemm examples in ckprofiler
* ran clang format
* added validation header and implement in core gemm operations
* remove ck_tile transpose and gemm stages from CI (#2646)
* update CK build instruction step 4 (#2563)
Co-authored-by: Aviral Goel <aviral.goel@amd.com>
* Fixes to "General 2D Reduction Kernel" (#2535) (#2656)
* fix reduce2d
- revret the combine_partial_results() chnages
- remove auto from function def
* clang-format
* enable aiter test_mha in daily CI (#2659)
* feat(copy_kernel): add basic copy kernel example with beginner friendly documentation (#2582)
* feat(copy_kernel): add basic copy kernel example with documentation
* docs(CHANGELOG): Updated changelog
* chore: performed clang format
* Update example/ck_tile/39_copy/copy_basic.cpp
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
* Update example/ck_tile/39_copy/README.md
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
* Update example/ck_tile/39_copy/README.md
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
* Update example/ck_tile/39_copy/README.md
Co-authored-by: spolifroni-amd <Sandra.Polifroni@amd.com>
* Update example/ck_tile/39_copy/README.md
Co-authored-by: spolifroni-amd <Sandra.Polifroni@amd.com>
* Update example/ck_tile/39_copy/README.md
Co-authored-by: spolifroni-amd <Sandra.Polifroni@amd.com>
* fix(terminology): follow amd terms
* extract elementwise copy to a new kernel
* fix(copy_kernel): bug in verification
* add comments about vgpr usage
* lint and nits
* add notes and comments
* print hostTensor via stream
* print hostTensor via stream
---------
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Co-authored-by: spolifroni-amd <Sandra.Polifroni@amd.com>
* [CK_TILE] FMHA BWD Optimization For GFX950 (#2628)
* simplify fmha_bwd_kernel MakeKargs & dq_dram_window
* simply duplicate
* trload pipeline
* Try two-stage
* add prefetch
* optimize & iglp
* Fix num_byte calculations to use nhead_k for K & V size (#2653)
Simple fix just to calculate the number of bytes correctly for what's reported in the output. I was getting 6200 GB/s which is past the SoL of MI300.
Before:
```
./bin/tile_example_fmha_fwd -prec=bf16 -b=2 -s=1 -s_k=32768 -h=32 -h_k=8 -d=128 -page_block_size=128 -num_splits=8 -iperm=0 -operm=0 -v=0 -kname=1
[bf16|batch|bshd] b:2, h:32/8, s:1/32768, d:128/128, scale_s:0.0883883, bias:n, p_drop:0, lse:0, squant:0, mask:n, v:r, num_splits:8, page_block_size:128, fmha_fwd_splitkv_d128_bf16_batch_b16x64x64x128x64x128_r1x4x1_r1x4x1_w16x16x16_w16x16x16_qr_nwarp_sshuffle_vr_ps_nlogits_nbias_nmask_lse_nsquant_pagedkv, fmha_fwd_splitkv_combine_d128_bf16_batch_b32_unused_ps_nlse_nsquant, 0.173 ms, 6.20 TFlops, 6202.95 GB/s
```
After:
```
./bin/tile_example_fmha_fwd -prec=bf16 -b=2 -s=1 -s_k=32768 -h=32 -h_k=8 -d=128 -page_block_size=128 -num_splits=8 -iperm=0 -operm=0 -v=0 -kname=1
[bf16|batch|bshd] b:2, h:32/8, s:1/32768, d:128/128, scale_s:0.0883883, bias:n, p_drop:0, lse:0, squant:0, mask:n, v:r, num_splits:8, page_block_size:128, fmha_fwd_splitkv_d128_bf16_batch_b16x64x64x128x64x128_r1x4x1_r1x4x1_w16x16x16_w16x16x16_qr_nwarp_sshuffle_vr_ps_nlogits_nbias_nmask_lse_nsquant_pagedkv, fmha_fwd_splitkv_combine_d128_bf16_batch_b32_unused_ps_nlse_nsquant, 0.163 ms, 6.58 TFlops, 1644.53 GB/s
```
* [CK_TILE] FMHA BWD Decode Pipeline (#2643)
* Fix distr
* Duplicate block_fmha_bwd_dq_dk_dv_pipeline_trload_kr_ktr_vr
* decode 16x16 o2
* fix (#2668)
* Optimize fmha fwd decode & prefill for gfx950 (#2641)
* Fix for fwd/bwd kernel build filter
* fix bwd code
* save an example for __bf16 type
* temp save, waiting for debug
* tempsave, fmha_decode
* temp save, change all instance to 1wave
* fix async copytest bug
* Add block_sync_lds_direct_load utility
* fix the s_waitcnt_imm calculation
* Improve s_waitcnt_imm calculation
* fix vmcnt shift
* add input validation and bug fix
* remove unnecessary output
* move test_copy into test
* temp save
* tempsave
* compile pass
* tempsave, trload+asyncload done
* tempsave. asynccopy+trload sanity checked
* remove unnecessary features
* fix the lds alignment caused performance regression
* enable prefill overload operator().
* remove all lds bankconflict with xor layouts
* enable larger tile size; upgrade xor pattern
* upgrade prefill pipeline; simple iglp; consistent data produce and consume order
* small refactor
* Load Q through lds, implement xor;
* add vmcnt guard before load ktile
* Add v_permlaneb32 for block_reduce. Disable it as it will cause un-coexecutable packed math in FA
* Add XOR fold strategy for hdim<128, but perf dropped; disable it by default; wait further perf debug
* add __restrict__ to tr load
* merge fa_decode pipeline into fmha_fwd api
* remove unnecessary files; rename some files
* Remove unnecessary changes
* bug fix, clang format;
* remove non-necessary change
* fix clangformat with 18.1.3
* fix bugs
* fix bug
* fix bug on non-gfx950
* fix bugs in gemm
* fix bug in pki4
* tempsave, update the blocksync functions
* change the warp setting for hdim32 fmha fwd
* clang format
* fix conflict. disable all v-col instance for fmha fwd
* Fix the bug
* clang format
---------
Co-authored-by: Max Podkorytov <4273004+tenpercent@users.noreply.github.com>
* Revert "Optimize fmha fwd decode & prefill for gfx950 (#2641)" (#2670)
This reverts commit
|