mirror of
https://github.com/ROCm/composable_kernel.git
synced 2026-05-20 12:59:49 +00:00
Navi3 rel (#1176)
* wmma_op + unit test * add arch limitation to wmma test * change arch limitation * Refactor + Add all type unit test(int4 compile failed) * Add f32_16x16x16_bf16 unit test * tempsave * tempsave * tempsave * runtime bug, cannot find symbol * workaround for incorrect HIP warpSize return value * debugging * tempsave * Correctness OK, waiting for optimization * Tidy up + format * temp save * temp save, reproduce the v_bfi_b32 issue * add inline asm for wmmaop test * tidy up * clean some debug purpose code * discard some codes * clang format * clang format * compiler issue fixed + increase tile size * navi3x_multipleD+example * temp save * workable * batchedgemm[OK], groupconv[debug] * groupconv: Sanity check[OK], Performance[Bad] * navi3x_groupconv_need_optimization * create necessary files * save progress * Add Inter-Row thread transfer * save progress * save debugging progress * sanity check pass * fix a host tensor bug and clean up flash-attn code * format * cancel unnecessary change * cancel unnecessary change * cancel unnecessary change * temp save, add asm backend flag to amd_wmma * Mat-A LDS Bypass sanity pass * temp save * gemm sanity fix * Porting new blockwise gemm to flash attention * Example branch provide to compiler team * tempsave * Fix a bug * batched gemm ported * conv A-skip lds ported * Skip B-Lds real gemm * Skip B Lds Gemm + MulD * batched gemm, conv, skip b lds * format * Attn, skip b lds * Change GridwiseOp nam * fix a typo caused bug * Skip A_Lds sanity pass, Skip B_Lds scratch occured * Bug found, intra-row permute off caused * bug found * a fix * disable buffer load due to incorrect 3rd dword * update fmha config, no scratch generated * update 3rd dword * fmha config update * FMHA, add support to gfx1101/gfx1102 * Merge origin dev (#2) * [Navi3x] Fix Gridwise_multiple_d operation (#649) * Add CMake Option "USE_OPT_NAVI3X" * fix bug * standardize docs (#655) * Separate bibtex requirement from rocm-docs-core (#656) * separate bibtex requirement from rocm-docs-core * point requirements to source rocm-docs-core repo * Add CMake Option "USE_OPT_NAVI3X" (#647) * Add CMake Option "USE_OPT_NAVI3X" * remove navi3x opt compile option from cmake script * Conv + quantization + tanh (#645) * Rename file. Prepare to support another activation * Add comment for quantization * Extract out_elementop * Add tanh example * Add conv + bias + tanh quantization instance * Add missing parameter * Refine cmake * Add external api and client example * Extract variable in example * Fix the comment --------- Co-authored-by: zjing14 <zhangjing14@gmail.com> * Add a denorm test fix (#603) * Add type_convert implementations for bf16 * Add the fix for conv_fwd * Add the fix for conv_bwd_data * Add the fix for conv_bwd_weight * Format * Format * Another format * Add a macro to use workaround on MI200 only * Format --------- Co-authored-by: Rosty Geyyer <rosty.geyyer@amd.com> Co-authored-by: zjing14 <zhangjing14@gmail.com> * simplify karg in device/grid of split-k op (#644) * simplify karg in device/grid split-k op * fix mk_kn_mn instances * add more instances * use name from tensor layout * fix 3rd dword of buffer source descriptor (#659) * add fp64 instances (#658) Co-authored-by: root <root@ctr-ubbsmc15.amd.com> * Issue #666: Revert "simplify karg in device/grid of split-k op (#644)" (#665) This reverts commit1108f64591. * Groupnorm + swish external api (#668) * Rename to proper naming * Add example of groupnorm + swish * Extract duplicate code in example * Add groupnorm + swish instances * Ractor instance generation, split into multiple cpp file * Add external api and client example * Refine profiler message * Use ck math version of exp * Refine problem size in example * Add host version of exp * add a marco to turn on/off denorm fix (off by default) (#673) * add a marco to turn off denorm fix by default * expose the marco --------- Co-authored-by: root <root@ctr-ubbsmc15.amd.com> * fixed quant example (#672) Co-authored-by: root <root@ctr-ubbsmc15.amd.com> * Add dependabot config and pin rocm-docs-core (#663) * [gtest] suppress unsafe buffer warn (#670) ref: https://github.com/ROCmSoftwarePlatform/MIOpen/pull/1912 * Add memory index guard in wmma device ops (#667) * Add more macros to turn on/off denorm fix (#678) Co-authored-by: Rosty Geyyer <rosty.geyyer@amd.com> * Fix a typo (#676) * Add (#677) * Allow using ROCm release candidate compilers. (#679) * enable use of rocm5.5 release candidate 4 * upgrade to ROCM5.5 RC5 * try fix the PUB_KEY error, remove the cmake-data package * upgrade to latest cmake version * use private dockerhub repo for rocm5.5 rc5 * add missing bracket * add vector load check * solve conflicts --------- Co-authored-by: Sam Wu <sjwu@ualberta.ca> Co-authored-by: Sam Wu <sam.wu2@amd.com> Co-authored-by: rocking5566 <ChunYu.Lai@amd.com> Co-authored-by: zjing14 <zhangjing14@gmail.com> Co-authored-by: Rostyslav Geyyer <46627076+geyyer@users.noreply.github.com> Co-authored-by: Rosty Geyyer <rosty.geyyer@amd.com> Co-authored-by: carlushuang <carlus.huang@amd.com> Co-authored-by: root <root@ctr-ubbsmc15.amd.com> Co-authored-by: Jun Liu <Liu.Jun@amd.com> Co-authored-by: Illia Silin <98187287+illsilin@users.noreply.github.com> * Disable SkipLDS & Align AIT api (#3) * fix layernorm, reduction Ops (#4) * [Navi3x] Fix Gridwise_multiple_d operation (#649) * Add CMake Option "USE_OPT_NAVI3X" * fix bug * standardize docs (#655) * Separate bibtex requirement from rocm-docs-core (#656) * separate bibtex requirement from rocm-docs-core * point requirements to source rocm-docs-core repo * Add CMake Option "USE_OPT_NAVI3X" (#647) * Add CMake Option "USE_OPT_NAVI3X" * remove navi3x opt compile option from cmake script * Conv + quantization + tanh (#645) * Rename file. Prepare to support another activation * Add comment for quantization * Extract out_elementop * Add tanh example * Add conv + bias + tanh quantization instance * Add missing parameter * Refine cmake * Add external api and client example * Extract variable in example * Fix the comment --------- Co-authored-by: zjing14 <zhangjing14@gmail.com> * Add a denorm test fix (#603) * Add type_convert implementations for bf16 * Add the fix for conv_fwd * Add the fix for conv_bwd_data * Add the fix for conv_bwd_weight * Format * Format * Another format * Add a macro to use workaround on MI200 only * Format --------- Co-authored-by: Rosty Geyyer <rosty.geyyer@amd.com> Co-authored-by: zjing14 <zhangjing14@gmail.com> * simplify karg in device/grid of split-k op (#644) * simplify karg in device/grid split-k op * fix mk_kn_mn instances * add more instances * use name from tensor layout * fix 3rd dword of buffer source descriptor (#659) * add fp64 instances (#658) Co-authored-by: root <root@ctr-ubbsmc15.amd.com> * Issue #666: Revert "simplify karg in device/grid of split-k op (#644)" (#665) This reverts commit1108f64591. * Groupnorm + swish external api (#668) * Rename to proper naming * Add example of groupnorm + swish * Extract duplicate code in example * Add groupnorm + swish instances * Ractor instance generation, split into multiple cpp file * Add external api and client example * Refine profiler message * Use ck math version of exp * Refine problem size in example * Add host version of exp * add a marco to turn on/off denorm fix (off by default) (#673) * add a marco to turn off denorm fix by default * expose the marco --------- Co-authored-by: root <root@ctr-ubbsmc15.amd.com> * fixed quant example (#672) Co-authored-by: root <root@ctr-ubbsmc15.amd.com> * Add dependabot config and pin rocm-docs-core (#663) * [gtest] suppress unsafe buffer warn (#670) ref: https://github.com/ROCmSoftwarePlatform/MIOpen/pull/1912 * Add memory index guard in wmma device ops (#667) * Add more macros to turn on/off denorm fix (#678) Co-authored-by: Rosty Geyyer <rosty.geyyer@amd.com> * Fix a typo (#676) * Add (#677) * Allow using ROCm release candidate compilers. (#679) * enable use of rocm5.5 release candidate 4 * upgrade to ROCM5.5 RC5 * try fix the PUB_KEY error, remove the cmake-data package * upgrade to latest cmake version * use private dockerhub repo for rocm5.5 rc5 * add missing bracket * Disable SkipLDS & Align AIT api * Update dependabot config (#682) Co-authored-by: samjwu <samjwu@users.noreply.github.com> * update attn api * solve type_convert bug + enable --------- Co-authored-by: Sam Wu <sjwu@ualberta.ca> Co-authored-by: Sam Wu <sam.wu2@amd.com> Co-authored-by: rocking5566 <ChunYu.Lai@amd.com> Co-authored-by: zjing14 <zhangjing14@gmail.com> Co-authored-by: Rostyslav Geyyer <46627076+geyyer@users.noreply.github.com> Co-authored-by: Rosty Geyyer <rosty.geyyer@amd.com> Co-authored-by: carlushuang <carlus.huang@amd.com> Co-authored-by: root <root@ctr-ubbsmc15.amd.com> Co-authored-by: Jun Liu <Liu.Jun@amd.com> Co-authored-by: Illia Silin <98187287+illsilin@users.noreply.github.com> Co-authored-by: samjwu <samjwu@users.noreply.github.com> Co-authored-by: haocwang <Haocong.WANG@amd.com> * fix typo * Fix attention with causal mask * multiple fix, try ait compile * Add A/B not use LDS pipeline * Clang format, Add gfx1101, gfx1102 support of FMHA example * cancel change of format script * 1. Enable 2-stage global Prefetch ( May cause VGPR spilling) 2. Enable FP16 accumulator blockwise_gemm * clang-format * 1. change blockwise gemm loopover direction from kmn to mnk ( ~1% improvement) 2. change kernel timing mode to 50 warmup + 50 timed repeat * Update low level abstration of blockwise gemm wmma * (2/5) bilinear gemm pass, perf bug: skip a lds has lower performance than skip b lds * (3/5) batched gemm pass, perf bug: skip a lds has lower performance than skip b lds * (4/5) grouped conv pass * (5/5) attention pass, todo: debug lds perf bug * AIT Attention API refactor (#8) * sanity pass * sanity pass 2 * confirm significant performance regression. * turn on all instances * turn off instance format * Fix bug & tunning & format * DML meta, self_attn+cross_attn * sanity pass * remove useless flag * update tile and problem size used in AIT attention * bug fix in grouped conv supporting check * deprecate inline asm wmma * Bug fix: double lds skip * clang-format * Fix errors in 1. example, fmha 2. gridwise pipeline 3. deviceop, fmha, change some containers from vector to array * part2 of previous commit * clang format * API fix of gridwisegemmpipeline * separate array base and vector base attention tensor transformation * fix gemm * clang format * add gemm fp16 instances * Temp save * fpAintB kernel compile pass * Sanity pass. * Temp save * debug code enabled * Fp16AInt8B_GEMM sanity * MQA implementation * GQA-4 example * tempsave * Compile pass * New implementation of fp16Aint8B Gemm, Acheieve similar math throughput with native fp16 Gemm * format * Todo: fix gemm_bilinear_wmma instances compilation bug * Solve a bug when K1=16 * remove unnecessary changes * Remove tensor layout limitation to LDS usage in tesnor contraction * update self-attention and cross-attention * fix a typo of name * Add arch limiter for fp8 gemm * enable fp8 gemm_xdl for all gfx9 targets * temporarily disable gemm_xdl_fp16_fp8 on MI100/200 * fix the cmake logic for gemm_xdl_fp16_fp8 * re-enable the gemm_xdl_fp16_fp8 on MI100/200 --------- Co-authored-by: aska-0096 <haocwang@amd.com> Co-authored-by: Sam Wu <sjwu@ualberta.ca> Co-authored-by: Sam Wu <sam.wu2@amd.com> Co-authored-by: rocking5566 <ChunYu.Lai@amd.com> Co-authored-by: Rostyslav Geyyer <46627076+geyyer@users.noreply.github.com> Co-authored-by: Rosty Geyyer <rosty.geyyer@amd.com> Co-authored-by: carlushuang <carlus.huang@amd.com> Co-authored-by: root <root@ctr-ubbsmc15.amd.com> Co-authored-by: Jun Liu <Liu.Jun@amd.com> Co-authored-by: Illia Silin <98187287+illsilin@users.noreply.github.com> Co-authored-by: samjwu <samjwu@users.noreply.github.com> Co-authored-by: haocwang <Haocong.WANG@amd.com> Co-authored-by: illsilin <Illia.Silin@amd.com> [ROCm/composable_kernel commit:1837040a9c]
This commit is contained in:
zjing14
@@ -111,6 +111,12 @@ list(APPEND GEMM_INSTANCES
|
||||
device_gemm_xdl_c_shuffle_fp8_fp8_fp8_km_kn_mn_instance.cpp
|
||||
device_gemm_xdl_c_shuffle_fp8_fp8_fp8_km_nk_mn_instance.cpp)
|
||||
|
||||
list(APPEND GEMM_INSTANCES
|
||||
device_gemm_wmma_f16_f16_f16_mk_kn_mn_instance.cpp
|
||||
device_gemm_wmma_f16_f16_f16_mk_nk_mn_instance.cpp
|
||||
device_gemm_wmma_f16_f16_f16_km_kn_mn_instance.cpp
|
||||
device_gemm_wmma_f16_f16_f16_km_nk_mn_instance.cpp)
|
||||
|
||||
add_instance_library(device_gemm_instance ${GEMM_INSTANCES})
|
||||
|
||||
set(ENABLE_PIPELINE_V2_OPT)
|
||||
|
||||
@@ -0,0 +1,78 @@
|
||||
// SPDX-License-Identifier: MIT
|
||||
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
|
||||
|
||||
#include <cstdlib>
|
||||
|
||||
#include "ck/ck.hpp"
|
||||
#include "ck/tensor_operation/gpu/device/tensor_layout.hpp"
|
||||
#include "ck/tensor_operation/gpu/device/gemm_specialization.hpp"
|
||||
#include "ck/tensor_operation/gpu/device/impl/device_gemm_wmma.hpp"
|
||||
#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
|
||||
|
||||
namespace ck {
|
||||
namespace tensor_operation {
|
||||
namespace device {
|
||||
namespace instance {
|
||||
|
||||
using F16 = ck::half_t;
|
||||
using F32 = float;
|
||||
|
||||
using Row = ck::tensor_layout::gemm::RowMajor;
|
||||
using Col = ck::tensor_layout::gemm::ColumnMajor;
|
||||
|
||||
template <ck::index_t... Is>
|
||||
using S = ck::Sequence<Is...>;
|
||||
|
||||
using PassThrough = ck::tensor_operation::element_wise::PassThrough;
|
||||
|
||||
// static constexpr auto GemmDefault = ck::tensor_operation::device::GemmSpecialization::Default;
|
||||
static constexpr auto GemmMNKPadding = ck::tensor_operation::device::GemmSpecialization::MNKPadding;
|
||||
|
||||
// Compilation parameters for a[k, m] * b[k, n] = c[m, n]
|
||||
using device_gemm_wmma_f16_f16_f16_km_kn_mn_instances = std::tuple<
|
||||
// clang-format off
|
||||
//######################| ALayout| BLayout| CLayout| AData| BData| CData| AccData| CShuffle| A| B| C| GEMM| NumPrefetch| Block| MPer| NPer| KPer| K1| MPer| NPer| M| N| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds| BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CShuffle| CShuffle| CShuffleBlockTransfer| CShuffleBlockTransfer|
|
||||
//######################| | | | Type| Type| Type| Type| DataType| Elementwise| Elementwise| Elementwise|Specialization| | Size| Block| Block| Block| | WMMA| WMMA| Repeat| Repeat| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraM| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraN| MRepeat| MRepeat| ClusterLengths| ScalarPerVector|
|
||||
//######################| | | | | | | | | Operation| Operation| Operation| | | | | | | | | | | | Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| | Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerStore| PerStore| MBlock_MPerBlock| |
|
||||
//######################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | NBlock_NPerBlock| |
|
||||
/* Prefetch 2, consume enormous vgpr resource*/
|
||||
// 8 Waves
|
||||
DeviceGemmWmma_CShuffle< Col, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 2, 256, 128, 128, 32, 8, 16, 16, 4, 2, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, 1, 1, S<1, 32, 1, 8>, 8>,
|
||||
// 4 Waves
|
||||
DeviceGemmWmma_CShuffle< Col, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 2, 128, 128, 64, 64, 8, 16, 16, 4, 2, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, 1, 1, S<1, 32, 1, 4>, 8>,
|
||||
// 2 Waves
|
||||
DeviceGemmWmma_CShuffle< Col, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 2, 64, 64, 32, 32, 8, 16, 16, 4, 1, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, 1, 1, S<1, 16, 1, 4>, 8>,
|
||||
// 1 Wave
|
||||
DeviceGemmWmma_CShuffle< Col, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 2, 32, 16, 16, 32, 8, 16, 16, 1, 1, S<2, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, S<2, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
/* Prefetch 1, prefer larger KPerBlock value for better latency hiding*/
|
||||
// 8 Waves
|
||||
DeviceGemmWmma_CShuffle< Col, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 256, 128, 256, 64, 8, 16, 16, 4, 4, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, 1, 1, S<1, 32, 1, 8>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Col, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 256, 128, 128, 64, 8, 16, 16, 4, 2, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, 1, 1, S<1, 32, 1, 8>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Col, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 256, 128, 160, 64, 8, 16, 16, 2, 5, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, S<8, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, 1, 1, S<1, 64, 1, 4>, 8>,
|
||||
// 4 Waves
|
||||
DeviceGemmWmma_CShuffle< Col, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 128, 128, 128, 32, 8, 16, 16, 4, 4, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, 1, 1, S<1, 32, 1, 4>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Col, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 128, 256, 64, 64, 8, 16, 16, 8, 2, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, 1, 1, S<1, 32, 1, 4>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Col, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 128, 64, 256, 64, 8, 16, 16, 2, 8, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, 1, 1, S<1, 32, 1, 4>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Col, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 128, 64, 80, 64, 8, 16, 16, 1, 5, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, S<8, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, 1, 1, S<1, 64, 1, 2>, 8>,
|
||||
// 2 Waves
|
||||
DeviceGemmWmma_CShuffle< Col, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 64, 16, 64, 64, 8, 16, 16, 1, 2, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, 1, 1, S<1, 16, 1, 4>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Col, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 64, 64, 32, 64, 8, 16, 16, 4, 1, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, 1, 1, S<1, 16, 1, 4>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Col, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 64, 32, 64, 64, 8, 16, 16, 2, 2, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, 1, 1, S<1, 16, 1, 4>, 8>,
|
||||
// 1 Wave
|
||||
DeviceGemmWmma_CShuffle< Col, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 32, 16, 32, 64, 8, 16, 16, 1, 2, S<2, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, S<2, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Col, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 32, 16, 16, 64, 8, 16, 16, 1, 1, S<2, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, S<2, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, 1, 1, S<1, 16, 1, 2>, 8>
|
||||
// clang-format on
|
||||
>;
|
||||
|
||||
void add_device_gemm_wmma_f16_f16_f16_km_kn_mn_instances(
|
||||
std::vector<std::unique_ptr<
|
||||
DeviceGemm<Col, Row, Row, F16, F16, F16, PassThrough, PassThrough, PassThrough>>>&
|
||||
instances)
|
||||
{
|
||||
add_device_operation_instances(instances, device_gemm_wmma_f16_f16_f16_km_kn_mn_instances{});
|
||||
}
|
||||
|
||||
} // namespace instance
|
||||
} // namespace device
|
||||
} // namespace tensor_operation
|
||||
} // namespace ck
|
||||
@@ -0,0 +1,78 @@
|
||||
// SPDX-License-Identifier: MIT
|
||||
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
|
||||
|
||||
#include <cstdlib>
|
||||
|
||||
#include "ck/ck.hpp"
|
||||
#include "ck/tensor_operation/gpu/device/tensor_layout.hpp"
|
||||
#include "ck/tensor_operation/gpu/device/gemm_specialization.hpp"
|
||||
#include "ck/tensor_operation/gpu/device/impl/device_gemm_wmma.hpp"
|
||||
#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
|
||||
|
||||
namespace ck {
|
||||
namespace tensor_operation {
|
||||
namespace device {
|
||||
namespace instance {
|
||||
|
||||
using F16 = ck::half_t;
|
||||
using F32 = float;
|
||||
|
||||
using Row = ck::tensor_layout::gemm::RowMajor;
|
||||
using Col = ck::tensor_layout::gemm::ColumnMajor;
|
||||
|
||||
template <ck::index_t... Is>
|
||||
using S = ck::Sequence<Is...>;
|
||||
|
||||
using PassThrough = ck::tensor_operation::element_wise::PassThrough;
|
||||
|
||||
// static constexpr auto GemmDefault = ck::tensor_operation::device::GemmSpecialization::Default;
|
||||
static constexpr auto GemmMNKPadding = ck::tensor_operation::device::GemmSpecialization::MNKPadding;
|
||||
|
||||
// Compilation parameters for a[k, m] * b[n, k] = c[m, n]
|
||||
using device_gemm_wmma_f16_f16_f16_km_nk_mn_instances = std::tuple<
|
||||
// clang-format off
|
||||
//######################| ALayout| BLayout| CLayout| AData| BData| CData| AccData| CShuffle| A| B| C| GEMM| NumPrefetch| Block| MPer| NPer| KPer| K1| MPer| NPer| M| N| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds| BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CShuffle| CShuffle| CShuffleBlockTransfer| CShuffleBlockTransfer|
|
||||
//######################| | | | Type| Type| Type| Type| DataType| Elementwise| Elementwise| Elementwise|Specialization| | Size| Block| Block| Block| | WMMA| WMMA| Repeat| Repeat| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraM| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraN| MRepeat| MRepeat| ClusterLengths| ScalarPerVector|
|
||||
//######################| | | | | | | | | Operation| Operation| Operation| | | | | | | | | | | | Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| | Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerStore| PerStore| MBlock_MPerBlock| |
|
||||
//######################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | NBlock_NPerBlock| |
|
||||
/* Prefetch 2, consume enormous vgpr resource*/
|
||||
// 8 Waves
|
||||
DeviceGemmWmma_CShuffle< Col, Col, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 2, 256, 128, 128, 32, 8, 16, 16, 4, 2, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, 1, 1, S<1, 32, 1, 8>, 8>,
|
||||
// 4 Waves
|
||||
DeviceGemmWmma_CShuffle< Col, Col, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 2, 128, 128, 64, 64, 8, 16, 16, 4, 2, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, 1, 1, S<1, 32, 1, 4>, 8>,
|
||||
// 2 Waves
|
||||
DeviceGemmWmma_CShuffle< Col, Col, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 2, 64, 64, 32, 32, 8, 16, 16, 4, 1, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, 1, 1, S<1, 16, 1, 4>, 8>,
|
||||
// 1 Wave
|
||||
DeviceGemmWmma_CShuffle< Col, Col, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 2, 32, 16, 16, 32, 8, 16, 16, 1, 1, S<2, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
/* Prefetch 1, prefer larger KPerBlock value for better latency hiding*/
|
||||
// 8 Waves
|
||||
DeviceGemmWmma_CShuffle< Col, Col, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 256, 128, 256, 64, 8, 16, 16, 4, 4, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, 1, 1, S<1, 32, 1, 8>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Col, Col, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 256, 128, 128, 64, 8, 16, 16, 4, 2, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, 1, 1, S<1, 32, 1, 8>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Col, Col, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 256, 128, 160, 64, 8, 16, 16, 2, 5, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, S<8, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, 1, 1, S<1, 64, 1, 4>, 8>,
|
||||
// 4 Waves
|
||||
DeviceGemmWmma_CShuffle< Col, Col, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 128, 128, 128, 32, 8, 16, 16, 4, 4, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, 1, 1, S<1, 32, 1, 4>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Col, Col, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 128, 256, 64, 64, 8, 16, 16, 8, 2, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, 1, 1, S<1, 32, 1, 4>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Col, Col, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 128, 64, 256, 64, 8, 16, 16, 2, 8, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, 1, 1, S<1, 32, 1, 4>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Col, Col, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 128, 64, 80, 64, 8, 16, 16, 1, 5, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, S<8, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, 1, 1, S<1, 64, 1, 2>, 8>,
|
||||
// 2 Waves
|
||||
DeviceGemmWmma_CShuffle< Col, Col, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 64, 16, 64, 64, 8, 16, 16, 1, 2, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, 1, 1, S<1, 16, 1, 4>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Col, Col, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 64, 64, 32, 64, 8, 16, 16, 4, 1, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, 1, 1, S<1, 16, 1, 4>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Col, Col, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 64, 32, 64, 64, 8, 16, 16, 2, 2, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, 1, 1, S<1, 16, 1, 4>, 8>,
|
||||
// 1 Wave
|
||||
DeviceGemmWmma_CShuffle< Col, Col, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 32, 16, 32, 64, 8, 16, 16, 1, 2, S<2, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Col, Col, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 32, 16, 16, 64, 8, 16, 16, 1, 1, S<2, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, 1, 1, S<1, 16, 1, 2>, 8>
|
||||
// clang-format on
|
||||
>;
|
||||
|
||||
void add_device_gemm_wmma_f16_f16_f16_km_nk_mn_instances(
|
||||
std::vector<std::unique_ptr<
|
||||
DeviceGemm<Col, Col, Row, F16, F16, F16, PassThrough, PassThrough, PassThrough>>>&
|
||||
instances)
|
||||
{
|
||||
add_device_operation_instances(instances, device_gemm_wmma_f16_f16_f16_km_nk_mn_instances{});
|
||||
}
|
||||
|
||||
} // namespace instance
|
||||
} // namespace device
|
||||
} // namespace tensor_operation
|
||||
} // namespace ck
|
||||
@@ -0,0 +1,158 @@
|
||||
// SPDX-License-Identifier: MIT
|
||||
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
|
||||
|
||||
#include <cstdlib>
|
||||
|
||||
#include "ck/ck.hpp"
|
||||
#include "ck/tensor_operation/gpu/device/tensor_layout.hpp"
|
||||
#include "ck/tensor_operation/gpu/device/gemm_specialization.hpp"
|
||||
#include "ck/tensor_operation/gpu/device/impl/device_gemm_wmma.hpp"
|
||||
#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
|
||||
|
||||
namespace ck {
|
||||
namespace tensor_operation {
|
||||
namespace device {
|
||||
namespace instance {
|
||||
|
||||
using F16 = ck::half_t;
|
||||
using F32 = float;
|
||||
|
||||
using Row = ck::tensor_layout::gemm::RowMajor;
|
||||
using Col = ck::tensor_layout::gemm::ColumnMajor;
|
||||
|
||||
template <ck::index_t... Is>
|
||||
using S = ck::Sequence<Is...>;
|
||||
|
||||
using PassThrough = ck::tensor_operation::element_wise::PassThrough;
|
||||
|
||||
// static constexpr auto GemmDefault = ck::tensor_operation::device::GemmSpecialization::Default;
|
||||
static constexpr auto GemmMNKPadding = ck::tensor_operation::device::GemmSpecialization::MNKPadding;
|
||||
|
||||
// Compilation parameters for a[m, k] * b[k, n] = c[m, n]
|
||||
using device_gemm_wmma_f16_f16_f16_mk_kn_mn_instances = std::tuple<
|
||||
// clang-format off
|
||||
//######################| ALayout| BLayout| CLayout| AData| BData| CData| AccData| CShuffle| A| B| C| GEMM| NumPrefetch| Block| MPer| NPer| KPer| K1| MPer| NPer| M| N| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds| BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CShuffle| CShuffle| CShuffleBlockTransfer| CShuffleBlockTransfer|
|
||||
//######################| | | | Type| Type| Type| Type| DataType| Elementwise| Elementwise| Elementwise|Specialization| | Size| Block| Block| Block| | WMMA| WMMA| Repeat| Repeat| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraM| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraN| MRepeat| MRepeat| ClusterLengths| ScalarPerVector|
|
||||
//######################| | | | | | | | | Operation| Operation| Operation| | | | | | | | | | | | Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| | Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerStore| PerStore| MBlock_MPerBlock| |
|
||||
//######################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | NBlock_NPerBlock| |
|
||||
/* Prefetch 2, consume enormous vgpr resource*/
|
||||
// 8 Waves
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 2, 256, 128, 128, 32, 8, 16, 16, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, 1, 1, S<1, 32, 1, 8>, 8>,
|
||||
// 4 Waves
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 2, 128, 128, 64, 64, 8, 16, 16, 4, 2, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, 1, 1, S<1, 32, 1, 4>, 8>,
|
||||
// 2 Waves
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 2, 64, 64, 32, 32, 8, 16, 16, 4, 1, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, 1, 1, S<1, 16, 1, 4>, 8>,
|
||||
// 1 Wave
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 2, 32, 16, 16, 32, 8, 16, 16, 1, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<2, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
/* Prefetch 1, prefer larger KPerBlock value for better latency hiding*/
|
||||
// 8 Waves
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 256, 128, 256, 64, 8, 16, 16, 4, 4, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, 1, 1, S<1, 32, 1, 8>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 256, 128, 128, 64, 8, 16, 16, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, 1, 1, S<1, 32, 1, 8>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 256, 128, 160, 64, 8, 16, 16, 2, 5, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<8, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, 1, 1, S<1, 64, 1, 4>, 8>,
|
||||
// 4 Waves
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 128, 128, 128, 32, 8, 16, 16, 4, 4, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, 1, 1, S<1, 32, 1, 4>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 128, 256, 64, 64, 8, 16, 16, 8, 2, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, 1, 1, S<1, 32, 1, 4>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 128, 64, 256, 64, 8, 16, 16, 2, 8, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, 1, 1, S<1, 32, 1, 4>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 128, 64, 80, 64, 8, 16, 16, 1, 5, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<8, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, 1, 1, S<1, 64, 1, 2>, 8>,
|
||||
// 2 Waves
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 64, 16, 64, 64, 8, 16, 16, 1, 2, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, 1, 1, S<1, 16, 1, 4>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 64, 64, 32, 64, 8, 16, 16, 4, 1, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, 1, 1, S<1, 16, 1, 4>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 64, 32, 64, 64, 8, 16, 16, 2, 2, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, 1, 1, S<1, 16, 1, 4>, 8>,
|
||||
// 1 Wave
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 32, 16, 32, 64, 8, 16, 16, 1, 2, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<2, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 32, 16, 16, 64, 8, 16, 16, 1, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<2, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 1, 8, true, 1, 1, S<1, 16, 1, 2>, 8>
|
||||
#if 0
|
||||
/* Prefetch 2, consume enormous vgpr resource*/
|
||||
// 8 Waves
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 2, 256, 128, 128, 32, 8, 16, 16, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, true, 1, 1, S<1, 32, 1, 8>, 8>,
|
||||
// 4 Waves
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 2, 128, 128, 64, 64, 8, 16, 16, 4, 2, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, true, 1, 1, S<1, 32, 1, 4>, 8>,
|
||||
// 2 Waves
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 2, 64, 64, 32, 32, 8, 16, 16, 4, 1, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, true, 1, 1, S<1, 16, 1, 4>, 8>,
|
||||
// 1 Wave
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 2, 32, 16, 16, 32, 8, 16, 16, 1, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<2, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, true, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
/* Prefetch 1, prefer larger KPerBlock value for better latency hiding*/
|
||||
// 8 Waves
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 256, 128, 256, 64, 8, 16, 16, 4, 4, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, true, 1, 1, S<1, 32, 1, 8>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 256, 128, 128, 64, 8, 16, 16, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, true, 1, 1, S<1, 32, 1, 8>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 256, 128, 160, 64, 8, 16, 16, 2, 5, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<8, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, true, 1, 1, S<1, 64, 1, 4>, 8>,
|
||||
// 4 Waves
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 128, 128, 128, 32, 8, 16, 16, 4, 4, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, true, 1, 1, S<1, 32, 1, 4>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 128, 256, 64, 64, 8, 16, 16, 8, 2, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, true, 1, 1, S<1, 32, 1, 4>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 128, 64, 256, 64, 8, 16, 16, 2, 8, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, true, 1, 1, S<1, 32, 1, 4>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 128, 64, 80, 64, 8, 16, 16, 1, 5, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<8, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, true, 1, 1, S<1, 64, 1, 2>, 8>,
|
||||
// 2 Waves
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 64, 16, 64, 64, 8, 16, 16, 1, 2, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, true, 1, 1, S<1, 16, 1, 4>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 64, 64, 32, 64, 8, 16, 16, 4, 1, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, true, 1, 1, S<1, 16, 1, 4>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 64, 32, 64, 64, 8, 16, 16, 2, 2, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, true, 1, 1, S<1, 16, 1, 4>, 8>,
|
||||
// 1 Wave
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 32, 16, 32, 64, 8, 16, 16, 1, 2, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<2, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, true, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 32, 16, 16, 64, 8, 16, 16, 1, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<2, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, true, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
/* Prefetch 2, consume enormous vgpr resource*/
|
||||
// 8 Waves
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 2, 256, 128, 128, 32, 8, 16, 16, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 8, true, 1, 1, S<1, 32, 1, 8>, 8>,
|
||||
// 4 Waves
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 2, 128, 128, 64, 64, 8, 16, 16, 4, 2, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 8, true, 1, 1, S<1, 32, 1, 4>, 8>,
|
||||
// 2 Waves
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 2, 64, 64, 32, 32, 8, 16, 16, 4, 1, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 8, true, 1, 1, S<1, 16, 1, 4>, 8>,
|
||||
// 1 Wave
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 2, 32, 16, 16, 32, 8, 16, 16, 1, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<2, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 8, true, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
/* Prefetch 1, prefer larger KPerBlock value for better latency hiding*/
|
||||
// 8 Waves
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 256, 128, 256, 64, 8, 16, 16, 4, 4, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 8, true, 1, 1, S<1, 32, 1, 8>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 256, 128, 128, 64, 8, 16, 16, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 8, true, 1, 1, S<1, 32, 1, 8>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 256, 128, 160, 64, 8, 16, 16, 2, 5, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<8, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 8, true, 1, 1, S<1, 64, 1, 4>, 8>,
|
||||
// 4 Waves
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 128, 128, 128, 32, 8, 16, 16, 4, 4, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 8, true, 1, 1, S<1, 32, 1, 4>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 128, 256, 64, 64, 8, 16, 16, 8, 2, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 8, true, 1, 1, S<1, 32, 1, 4>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 128, 64, 256, 64, 8, 16, 16, 2, 8, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 8, true, 1, 1, S<1, 32, 1, 4>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 128, 64, 80, 64, 8, 16, 16, 1, 5, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<8, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 8, true, 1, 1, S<1, 64, 1, 2>, 8>,
|
||||
// 2 Waves
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 64, 16, 64, 64, 8, 16, 16, 1, 2, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 8, true, 1, 1, S<1, 16, 1, 4>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 64, 64, 32, 64, 8, 16, 16, 4, 1, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 8, true, 1, 1, S<1, 16, 1, 4>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 64, 32, 64, 64, 8, 16, 16, 2, 2, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 8, true, 1, 1, S<1, 16, 1, 4>, 8>,
|
||||
// 1 Wave
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 32, 16, 32, 64, 8, 16, 16, 1, 2, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<2, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 8, true, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 32, 16, 16, 64, 8, 16, 16, 1, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<2, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 4, 8, true, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
/* Prefetch 2, consume enormous vgpr resource*/
|
||||
// 8 Waves
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 2, 256, 128, 128, 32, 8, 16, 16, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 8, 8, true, 1, 1, S<1, 32, 1, 8>, 8>,
|
||||
// 4 Waves
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 2, 128, 128, 64, 64, 8, 16, 16, 4, 2, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 8, 8, true, 1, 1, S<1, 32, 1, 4>, 8>,
|
||||
// 2 Waves
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 2, 64, 64, 32, 32, 8, 16, 16, 4, 1, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 8, 8, true, 1, 1, S<1, 16, 1, 4>, 8>,
|
||||
// 1 Wave
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 2, 32, 16, 16, 32, 8, 16, 16, 1, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<2, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 8, 8, true, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
/* Prefetch 1, prefer larger KPerBlock value for better latency hiding*/
|
||||
// 8 Waves
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 256, 128, 256, 64, 8, 16, 16, 4, 4, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 8, 8, true, 1, 1, S<1, 32, 1, 8>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 256, 128, 128, 64, 8, 16, 16, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 8, 8, true, 1, 1, S<1, 32, 1, 8>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 256, 128, 160, 64, 8, 16, 16, 2, 5, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<8, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 8, 8, true, 1, 1, S<1, 64, 1, 4>, 8>,
|
||||
// 4 Waves
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 128, 128, 128, 32, 8, 16, 16, 4, 4, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 8, 8, true, 1, 1, S<1, 32, 1, 4>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 128, 256, 64, 64, 8, 16, 16, 8, 2, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 8, 8, true, 1, 1, S<1, 32, 1, 4>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 128, 64, 256, 64, 8, 16, 16, 2, 8, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 8, 8, true, 1, 1, S<1, 32, 1, 4>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 128, 64, 80, 64, 8, 16, 16, 1, 5, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<8, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 8, 8, true, 1, 1, S<1, 64, 1, 2>, 8>,
|
||||
// 2 Waves
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 64, 16, 64, 64, 8, 16, 16, 1, 2, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 8, 8, true, 1, 1, S<1, 16, 1, 4>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 64, 64, 32, 64, 8, 16, 16, 4, 1, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 8, 8, true, 1, 1, S<1, 16, 1, 4>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 64, 32, 64, 64, 8, 16, 16, 2, 2, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 8, 8, true, 1, 1, S<1, 16, 1, 4>, 8>,
|
||||
// 1 Wave
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 32, 16, 32, 64, 8, 16, 16, 1, 2, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<2, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 8, 8, true, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Row, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 32, 16, 16, 64, 8, 16, 16, 1, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<2, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 8, 8, true, 1, 1, S<1, 16, 1, 2>, 8>
|
||||
#endif
|
||||
// clang-format on
|
||||
>;
|
||||
|
||||
void add_device_gemm_wmma_f16_f16_f16_mk_kn_mn_instances(
|
||||
std::vector<std::unique_ptr<
|
||||
DeviceGemm<Row, Row, Row, F16, F16, F16, PassThrough, PassThrough, PassThrough>>>&
|
||||
instances)
|
||||
{
|
||||
add_device_operation_instances(instances, device_gemm_wmma_f16_f16_f16_mk_kn_mn_instances{});
|
||||
}
|
||||
|
||||
} // namespace instance
|
||||
} // namespace device
|
||||
} // namespace tensor_operation
|
||||
} // namespace ck
|
||||
@@ -0,0 +1,78 @@
|
||||
// SPDX-License-Identifier: MIT
|
||||
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
|
||||
|
||||
#include <cstdlib>
|
||||
|
||||
#include "ck/ck.hpp"
|
||||
#include "ck/tensor_operation/gpu/device/tensor_layout.hpp"
|
||||
#include "ck/tensor_operation/gpu/device/gemm_specialization.hpp"
|
||||
#include "ck/tensor_operation/gpu/device/impl/device_gemm_wmma.hpp"
|
||||
#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
|
||||
|
||||
namespace ck {
|
||||
namespace tensor_operation {
|
||||
namespace device {
|
||||
namespace instance {
|
||||
|
||||
using F16 = ck::half_t;
|
||||
using F32 = float;
|
||||
|
||||
using Row = ck::tensor_layout::gemm::RowMajor;
|
||||
using Col = ck::tensor_layout::gemm::ColumnMajor;
|
||||
|
||||
template <ck::index_t... Is>
|
||||
using S = ck::Sequence<Is...>;
|
||||
|
||||
using PassThrough = ck::tensor_operation::element_wise::PassThrough;
|
||||
|
||||
// static constexpr auto GemmDefault = ck::tensor_operation::device::GemmSpecialization::Default;
|
||||
static constexpr auto GemmMNKPadding = ck::tensor_operation::device::GemmSpecialization::MNKPadding;
|
||||
|
||||
// Compilation parameters for a[m, k] * b[n, k] = c[m, n]
|
||||
using device_gemm_wmma_f16_f16_f16_mk_nk_mn_instances = std::tuple<
|
||||
// clang-format off
|
||||
//######################| ALayout| BLayout| CLayout| AData| BData| CData| AccData| CShuffle| A| B| C| GEMM| NumPrefetch| Block| MPer| NPer| KPer| K1| MPer| NPer| M| N| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds| BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CShuffle| CShuffle| CShuffleBlockTransfer| CShuffleBlockTransfer|
|
||||
//######################| | | | Type| Type| Type| Type| DataType| Elementwise| Elementwise| Elementwise| Specialization| | Size| Block| Block| Block| | WMMA| WMMA| Repeat| Repeat| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraM| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraN| MRepeat| MRepeat| ClusterLengths| ScalarPerVector|
|
||||
//######################| | | | | | | | | Operation| Operation| Operation| | | | | | | | | | | | Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| | Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerStore| PerStore| MBlock_MPerBlock| |
|
||||
//######################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | NBlock_NPerBlock| |
|
||||
/* Prefetch 2, consume enormous vgpr resource*/
|
||||
// 8 Waves
|
||||
DeviceGemmWmma_CShuffle< Row, Col, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 2, 256, 128, 128, 32, 8, 16, 16, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, 1, 1, S<1, 32, 1, 8>, 8>,
|
||||
// 4 Waves
|
||||
DeviceGemmWmma_CShuffle< Row, Col, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 2, 128, 128, 64, 64, 8, 16, 16, 4, 2, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, 1, 1, S<1, 32, 1, 4>, 8>,
|
||||
// 2 Waves
|
||||
DeviceGemmWmma_CShuffle< Row, Col, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 2, 64, 64, 32, 32, 8, 16, 16, 4, 1, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, 1, 1, S<1, 16, 1, 4>, 8>,
|
||||
// 1 Wave
|
||||
DeviceGemmWmma_CShuffle< Row, Col, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 2, 32, 16, 16, 32, 8, 16, 16, 1, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
/* Prefetch 1, prefer larger KPerBlock value for better latency hiding*/
|
||||
// 8 Waves
|
||||
DeviceGemmWmma_CShuffle< Row, Col, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 256, 128, 256, 64, 8, 16, 16, 4, 4, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, 1, 1, S<1, 32, 1, 8>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Col, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 256, 128, 128, 64, 8, 16, 16, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, 1, 1, S<1, 32, 1, 8>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Col, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 256, 128, 160, 64, 8, 16, 16, 2, 5, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<8, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, 1, 1, S<1, 64, 1, 4>, 8>,
|
||||
// 4 Waves
|
||||
DeviceGemmWmma_CShuffle< Row, Col, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 128, 128, 128, 32, 8, 16, 16, 4, 4, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, 1, 1, S<1, 32, 1, 4>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Col, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 128, 256, 64, 64, 8, 16, 16, 8, 2, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, 1, 1, S<1, 32, 1, 4>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Col, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 128, 64, 256, 64, 8, 16, 16, 2, 8, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, 1, 1, S<1, 32, 1, 4>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Col, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 128, 64, 80, 64, 8, 16, 16, 1, 5, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<8, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, 1, 1, S<1, 64, 1, 2>, 8>,
|
||||
// 2 Waves
|
||||
DeviceGemmWmma_CShuffle< Row, Col, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 64, 16, 64, 64, 8, 16, 16, 1, 2, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, 1, 1, S<1, 16, 1, 4>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Col, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 64, 64, 32, 64, 8, 16, 16, 4, 1, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, 1, 1, S<1, 16, 1, 4>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Col, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 64, 32, 64, 64, 8, 16, 16, 2, 2, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, 1, 1, S<1, 16, 1, 4>, 8>,
|
||||
// 1 Wave
|
||||
DeviceGemmWmma_CShuffle< Row, Col, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 32, 16, 32, 64, 8, 16, 16, 1, 2, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
DeviceGemmWmma_CShuffle< Row, Col, Row, F16, F16, F16, F32, F16, PassThrough, PassThrough, PassThrough, GemmMNKPadding, 1, 32, 16, 16, 64, 8, 16, 16, 1, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, true, 1, 1, S<1, 16, 1, 2>, 8>
|
||||
// clang-format on
|
||||
>;
|
||||
|
||||
void add_device_gemm_wmma_f16_f16_f16_mk_nk_mn_instances(
|
||||
std::vector<std::unique_ptr<
|
||||
DeviceGemm<Row, Col, Row, F16, F16, F16, PassThrough, PassThrough, PassThrough>>>&
|
||||
instances)
|
||||
{
|
||||
add_device_operation_instances(instances, device_gemm_wmma_f16_f16_f16_mk_nk_mn_instances{});
|
||||
}
|
||||
|
||||
} // namespace instance
|
||||
} // namespace device
|
||||
} // namespace tensor_operation
|
||||
} // namespace ck
|
||||
@@ -36,32 +36,32 @@ static constexpr auto GemmMNKPadding = ck::tensor_operation::device::GemmSpecial
|
||||
// e[m, n] = bilinear(a[m, k] * b[k, n], d[m, n])
|
||||
using device_gemm_bilinear_wmma_c_shuffle_i8_i8_i8_i8_km_kn_mn_mn_instances = std::tuple<
|
||||
// clang-format off
|
||||
//################################| A| B| Ds| E| AData| BData| DsData| EData| AccData| CShuffle| A| B| CDE| GEMM| Block| MPer| NPer| K0Per| K1| MPer| NPer| MRepeat| NRepeat| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds| BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CShuffle| CShuffle| CBlockTransferClusterLengths| CBlockTransfer|
|
||||
//################################| Layout| Layout| Layout| Layout| Type| Type| Type| Type| Type| DataType| Elementwise| Elementwise| Elementwise| Specialization| Size| Block| Block| Block| | WMMA| WMMA| | | ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraM| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraN| MXdlPerWave| NXdlPerWave| _MBlock_MWaveMPerXdl| ScalarPerVector|
|
||||
//################################| | | | | | | | | | | Operation| Operation| Operation| | | | | | | | | | | Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| | Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerShuffle| PerShuffle| _NBlock_NWaveNPerXdl| _NWaveNPerXdl|
|
||||
//################################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Row, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmDefault, 256, 128, 128, 4, 16, 16, 16, 4, 2, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, 1, 2, S<1, 32, 1, 8>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Row, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmDefault, 128, 64, 64, 4, 16, 16, 16, 2, 2, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, 1, 2, S<1, 32, 1, 4>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Row, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmDefault, 64, 32, 32, 4, 16, 16, 16, 1, 2, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, 1, 2, S<1, 32, 1, 2>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Row, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmDefault, 32, 16, 16, 4, 16, 16, 16, 1, 1, S<4, 8, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, S<4, 8, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
//################################| A| B| Ds| E| AData| BData| AccData| CShuffle| DsData| EData| A| B| CDE| GEMM| Prefetch| Block| MPer| NPer| K0Per| K1| MPer| NPer| MRepeat| NRepeat| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds| BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CShuffle| CShuffle| CBlockTransferClusterLengths| CBlockTransfer|
|
||||
//################################| Layout| Layout| Layout| Layout| Type| Type| Type| DataType| Type| Type| Elementwise| Elementwise| Elementwise| Specialization| Stage| Size| Block| Block| Block| | WMMA| WMMA| | | ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraM| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraN| MXdlPerWave| NXdlPerWave| _MBlock_MWaveMPerXdl| ScalarPerVector|
|
||||
//################################| | | | | | | | | | | Operation| Operation| Operation| | | | | | | | | | | | Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| | Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerShuffle| PerShuffle| _NBlock_NWaveNPerXdl| _NWaveNPerXdl|
|
||||
//################################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Row, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmDefault, 1, 256, 128, 128, 64, 16, 16, 16, 4, 2, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, 1, 2, S<1, 32, 1, 8>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Row, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmDefault, 1, 128, 64, 64, 64, 16, 16, 16, 2, 2, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, 1, 2, S<1, 32, 1, 4>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Row, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmDefault, 1, 64, 32, 32, 64, 16, 16, 16, 1, 2, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, 1, 2, S<1, 32, 1, 2>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Row, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmDefault, 1, 32, 16, 16, 64, 16, 16, 16, 1, 1, S<4, 8, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, S<4, 8, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
|
||||
// M/N/K padding
|
||||
//################################| A| B| Ds| E| AData| BData| DsData| EData| AccData| CShuffle| A| B| CDE| GEMM| Block| MPer| NPer| K0Per| K1| MPer| NPer| MRepeat| NRepeat| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds| BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CShuffle| CShuffle| CBlockTransferClusterLengths| CBlockTransfer|
|
||||
//################################| Layout| Layout| Layout| Layout| Type| Type| Type| Type| Type| DataType| Elementwise| Elementwise| Elementwise| Specialization| Size| Block| Block| Block| | WMMA| WMMA| | | ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraM| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraN| MXdlPerWave| NXdlPerWave| _MBlock_MWaveMPerXdl| ScalarPerVector|
|
||||
//################################| | | | | | | | | | | Operation| Operation| Operation| | | | | | | | | | | Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| | Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerShuffle| PerShuffle| _NBlock_NWaveNPerXdl| _NWaveNPerXdl|
|
||||
//################################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Row, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 256, 128, 128, 4, 16, 16, 16, 4, 2, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, 1, 2, S<1, 32, 1, 8>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Row, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 128, 64, 64, 4, 16, 16, 16, 2, 2, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, 1, 2, S<1, 32, 1, 4>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Row, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 64, 32, 32, 4, 16, 16, 16, 1, 2, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, 1, 2, S<1, 32, 1, 2>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Row, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 32, 16, 16, 4, 16, 16, 16, 1, 1, S<4, 8, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, S<4, 8, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Row, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 256, 128, 128, 8, 8, 16, 16, 4, 2, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, 1, 2, S<1, 32, 1, 8>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Row, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 128, 64, 64, 8, 8, 16, 16, 2, 2, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, 1, 2, S<1, 32, 1, 4>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Row, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 64, 32, 32, 8, 8, 16, 16, 1, 2, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, 1, 2, S<1, 32, 1, 2>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Row, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 32, 16, 16, 8, 8, 16, 16, 1, 1, S<4, 8, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, S<4, 8, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Row, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 256, 128, 128, 8, 4, 16, 16, 4, 2, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 4, 1, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 4, 1, 1, 2, S<1, 32, 1, 8>, 4>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Row, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 128, 64, 64, 8, 4, 16, 16, 2, 2, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 4, 1, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 4, 1, 1, 2, S<1, 32, 1, 4>, 4>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Row, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 64, 32, 32, 8, 4, 16, 16, 1, 2, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 4, 1, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 4, 1, 1, 2, S<1, 32, 1, 2>, 4>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Row, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 32, 16, 16, 8, 4, 16, 16, 1, 1, S<4, 8, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 4, 1, S<4, 8, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 4, 1, 1, 1, S<1, 16, 1, 2>, 4>
|
||||
//################################| A| B| Ds| E| AData| BData| AccData| CShuffle| DsData| EData| A| B| CDE| GEMM| Prefetch| Block| MPer| NPer| K0Per| K1| MPer| NPer| MRepeat| NRepeat| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds| BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CShuffle| CShuffle| CBlockTransferClusterLengths| CBlockTransfer|
|
||||
//################################| Layout| Layout| Layout| Layout| Type| Type| Type| DataType| Type| Type| Elementwise| Elementwise| Elementwise| Specialization| Stage| Size| Block| Block| Block| | WMMA| WMMA| | | ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraM| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraN| MXdlPerWave| NXdlPerWave| _MBlock_MWaveMPerXdl| ScalarPerVector|
|
||||
//################################| | | | | | | | | | | Operation| Operation| Operation| | | | | | | | | | | | Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| | Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerShuffle| PerShuffle| _NBlock_NWaveNPerXdl| _NWaveNPerXdl|
|
||||
//################################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Row, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 256, 128, 128, 64, 16, 16, 16, 4, 2, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, 1, 2, S<1, 32, 1, 8>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Row, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 128, 64, 64, 64, 16, 16, 16, 2, 2, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, 1, 2, S<1, 32, 1, 4>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Row, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 64, 32, 32, 64, 16, 16, 16, 1, 2, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, 1, 2, S<1, 32, 1, 2>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Row, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 32, 16, 16, 64, 16, 16, 16, 1, 1, S<4, 8, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, S<4, 8, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Row, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 256, 128, 128, 64, 8, 16, 16, 4, 2, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, 1, 2, S<1, 32, 1, 8>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Row, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 128, 64, 64, 64, 8, 16, 16, 2, 2, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, 1, 2, S<1, 32, 1, 4>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Row, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 64, 32, 32, 64, 8, 16, 16, 1, 2, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, 1, 2, S<1, 32, 1, 2>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Row, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 32, 16, 16, 64, 8, 16, 16, 1, 1, S<4, 8, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, S<4, 8, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Row, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 256, 128, 128, 32, 4, 16, 16, 4, 2, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 4, 1, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 4, 1, 1, 2, S<1, 32, 1, 8>, 4>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Row, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 128, 64, 64, 32, 4, 16, 16, 2, 2, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 4, 1, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 4, 1, 1, 2, S<1, 32, 1, 4>, 4>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Row, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 64, 32, 32, 32, 4, 16, 16, 1, 2, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 4, 1, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 4, 1, 1, 2, S<1, 32, 1, 2>, 4>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Row, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 32, 16, 16, 32, 4, 16, 16, 1, 1, S<4, 8, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 4, 1, S<4, 8, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 4, 1, 1, 1, S<1, 16, 1, 2>, 4>
|
||||
|
||||
// clang-format on
|
||||
>;
|
||||
|
||||
@@ -36,32 +36,32 @@ static constexpr auto GemmMNKPadding = ck::tensor_operation::device::GemmSpecial
|
||||
// e[m, n] = bilinear(a[m, k] * b[k, n], d[m, n])
|
||||
using device_gemm_bilinear_wmma_c_shuffle_i8_i8_i8_i8_km_nk_mn_mn_instances = std::tuple<
|
||||
// clang-format off
|
||||
//################################| A| B| Ds| E| AData| BData| DsData| EData| AccData| CShuffle| A| B| CDE| GEMM| Block| MPer| NPer| K0Per| K1| MPer| NPer| MRepeat| NRepeat| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds| BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CShuffle| CShuffle| CBlockTransferClusterLengths| CBlockTransfer|
|
||||
//################################| Layout| Layout| Layout| Layout| Type| Type| Type| Type| Type| DataType| Elementwise| Elementwise| Elementwise| Specialization| Size| Block| Block| Block| | WMMA| WMMA| | | ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraM| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraN| MXdlPerWave| NXdlPerWave| _MBlock_MWaveMPerXdl| ScalarPerVector|
|
||||
//################################| | | | | | | | | | | Operation| Operation| Operation| | | | | | | | | | | Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| | Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerShuffle| PerShuffle| _NBlock_NWaveNPerXdl| _NWaveNPerXdl|
|
||||
//################################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmDefault, 256, 128, 128, 4, 16, 16, 16, 4, 2, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, 1, 2, S<1, 32, 1, 8>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmDefault, 128, 64, 64, 4, 16, 16, 16, 2, 2, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, 1, 2, S<1, 32, 1, 4>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmDefault, 64, 32, 32, 4, 16, 16, 16, 1, 2, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, 1, 2, S<1, 32, 1, 2>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmDefault, 32, 16, 16, 4, 16, 16, 16, 1, 1, S<4, 8, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
//################################| A| B| Ds| E| AData| BData| AccData| CShuffle| DsData| EData| A| B| CDE| GEMM| Prefetch| Block| MPer| NPer| K0Per| K1| MPer| NPer| MRepeat| NRepeat| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds| BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CShuffle| CShuffle| CBlockTransferClusterLengths| CBlockTransfer|
|
||||
//################################| Layout| Layout| Layout| Layout| Type| Type| Type| DataType| Type| Type| Elementwise| Elementwise| Elementwise| Specialization| Stage| Size| Block| Block| Block| | WMMA| WMMA| | | ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraM| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraN| MXdlPerWave| NXdlPerWave| _MBlock_MWaveMPerXdl| ScalarPerVector|
|
||||
//################################| | | | | | | | | | | Operation| Operation| Operation| | | | | | | | | | | | Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| | Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerShuffle| PerShuffle| _NBlock_NWaveNPerXdl| _NWaveNPerXdl|
|
||||
//################################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmDefault, 1, 256, 128, 128, 64, 16, 16, 16, 4, 2, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, 1, 2, S<1, 32, 1, 8>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmDefault, 1, 128, 64, 64, 64, 16, 16, 16, 2, 2, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, 1, 2, S<1, 32, 1, 4>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmDefault, 1, 64, 32, 32, 64, 16, 16, 16, 1, 2, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, 1, 2, S<1, 32, 1, 2>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmDefault, 1, 32, 16, 16, 64, 16, 16, 16, 1, 1, S<4, 8, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
|
||||
// M/N/K padding
|
||||
//################################| A| B| Ds| E| AData| BData| DsData| EData| AccData| CShuffle| A| B| CDE| GEMM| Block| MPer| NPer| K0Per| K1| MPer| NPer| MRepeat| NRepeat| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds| BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CShuffle| CShuffle| CBlockTransferClusterLengths| CBlockTransfer|
|
||||
//################################| Layout| Layout| Layout| Layout| Type| Type| Type| Type| Type| DataType| Elementwise| Elementwise| Elementwise| Specialization| Size| Block| Block| Block| | WMMA| WMMA| | | ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraM| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraN| MXdlPerWave| NXdlPerWave| _MBlock_MWaveMPerXdl| ScalarPerVector|
|
||||
//################################| | | | | | | | | | | Operation| Operation| Operation| | | | | | | | | | | Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| | Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerShuffle| PerShuffle| _NBlock_NWaveNPerXdl| _NWaveNPerXdl|
|
||||
//################################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 256, 128, 128, 4, 16, 16, 16, 4, 2, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, 1, 2, S<1, 32, 1, 8>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 128, 64, 64, 4, 16, 16, 16, 2, 2, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, 1, 2, S<1, 32, 1, 4>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 64, 32, 32, 4, 16, 16, 16, 1, 2, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, 1, 2, S<1, 32, 1, 2>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 32, 16, 16, 4, 16, 16, 16, 1, 1, S<4, 8, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 256, 128, 128, 8, 8, 16, 16, 4, 2, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, 1, 2, S<1, 32, 1, 8>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 128, 64, 64, 8, 8, 16, 16, 2, 2, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, 1, 2, S<1, 32, 1, 4>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 64, 32, 32, 8, 8, 16, 16, 1, 2, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, 1, 2, S<1, 32, 1, 2>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 32, 16, 16, 8, 8, 16, 16, 1, 1, S<4, 8, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 256, 128, 128, 8, 4, 16, 16, 4, 2, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 4, 1, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, 1, 1, 2, S<1, 32, 1, 8>, 4>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 128, 64, 64, 8, 4, 16, 16, 2, 2, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 4, 1, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, 1, 1, 2, S<1, 32, 1, 4>, 4>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 64, 32, 32, 8, 4, 16, 16, 1, 2, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 4, 1, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, 1, 1, 2, S<1, 32, 1, 2>, 4>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 32, 16, 16, 8, 4, 16, 16, 1, 1, S<4, 8, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 4, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, 1, 1, 1, S<1, 16, 1, 2>, 4>
|
||||
//################################| A| B| Ds| E| AData| BData| AccData| CShuffle| DsData| EData| A| B| CDE| GEMM| Prefetch| Block| MPer| NPer| K0Per| K1| MPer| NPer| MRepeat| NRepeat| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds| BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CShuffle| CShuffle| CBlockTransferClusterLengths| CBlockTransfer|
|
||||
//################################| Layout| Layout| Layout| Layout| Type| Type| Type| DataType| Type| Type| Elementwise| Elementwise| Elementwise| Specialization| Stage| Size| Block| Block| Block| | WMMA| WMMA| | | ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraM| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraN| MXdlPerWave| NXdlPerWave| _MBlock_MWaveMPerXdl| ScalarPerVector|
|
||||
//################################| | | | | | | | | | | Operation| Operation| Operation| | | | | | | | | | | | Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| | Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerShuffle| PerShuffle| _NBlock_NWaveNPerXdl| _NWaveNPerXdl|
|
||||
//################################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 256, 128, 128, 64, 16, 16, 16, 4, 2, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, 1, 2, S<1, 32, 1, 8>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 128, 64, 64, 64, 16, 16, 16, 2, 2, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, 1, 2, S<1, 32, 1, 4>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 64, 32, 32, 64, 16, 16, 16, 1, 2, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, 1, 2, S<1, 32, 1, 2>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 32, 16, 16, 64, 16, 16, 16, 1, 1, S<4, 8, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 256, 128, 128, 64, 8, 16, 16, 4, 2, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, 1, 2, S<1, 32, 1, 8>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 128, 64, 64, 64, 8, 16, 16, 2, 2, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, 1, 2, S<1, 32, 1, 4>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 64, 32, 32, 64, 8, 16, 16, 1, 2, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, 1, 2, S<1, 32, 1, 2>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 32, 16, 16, 64, 8, 16, 16, 1, 1, S<4, 8, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 256, 128, 128, 32, 4, 16, 16, 4, 2, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 4, 1, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, 1, 1, 2, S<1, 32, 1, 8>, 4>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 128, 64, 64, 32, 4, 16, 16, 2, 2, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 4, 1, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, 1, 1, 2, S<1, 32, 1, 4>, 4>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 64, 32, 32, 32, 4, 16, 16, 1, 2, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 4, 1, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, 1, 1, 2, S<1, 32, 1, 2>, 4>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Col, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 32, 16, 16, 32, 4, 16, 16, 1, 1, S<4, 8, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 4, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, 1, 1, 1, S<1, 16, 1, 2>, 4>
|
||||
|
||||
// clang-format on
|
||||
>;
|
||||
|
||||
@@ -36,32 +36,32 @@ static constexpr auto GemmMNKPadding = ck::tensor_operation::device::GemmSpecial
|
||||
// e[m, n] = bilinear(a[m, k] * b[k, n], d[m, n])
|
||||
using device_gemm_bilinear_wmma_c_shuffle_i8_i8_i8_i8_mk_kn_mn_mn_instances = std::tuple<
|
||||
// clang-format off
|
||||
//################################| A| B| Ds| E| AData| BData| DsData| EData| AccData| CShuffle| A| B| CDE| GEMM| Block| MPer| NPer| K0Per| K1| MPer| NPer| MRepeat| NRepeat| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds| BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CShuffle| CShuffle| CBlockTransferClusterLengths| CBlockTransfer|
|
||||
//################################| Layout| Layout| Layout| Layout| Type| Type| Type| Type| Type| DataType| Elementwise| Elementwise| Elementwise| Specialization| Size| Block| Block| Block| | WMMA| WMMA| | | ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraM| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraN| MXdlPerWave| NXdlPerWave| _MBlock_MWaveMPerXdl| ScalarPerVector|
|
||||
//################################| | | | | | | | | | | Operation| Operation| Operation| | | | | | | | | | | Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| | Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerShuffle| PerShuffle| _NBlock_NWaveNPerXdl| _NWaveNPerXdl|
|
||||
//################################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Row, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmDefault, 256, 128, 128, 4, 16, 16, 16, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, 1, 2, S<1, 32, 1, 8>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Row, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmDefault, 128, 64, 64, 4, 16, 16, 16, 2, 2, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, 1, 2, S<1, 32, 1, 4>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Row, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmDefault, 64, 32, 32, 4, 16, 16, 16, 1, 2, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, 1, 2, S<1, 32, 1, 2>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Row, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmDefault, 32, 16, 16, 4, 16, 16, 16, 1, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, S<4, 8, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
//################################| A| B| Ds| E| AData| BData| AccData| CShuffle| DsData| EData| A| B| CDE| GEMM| Prefetch| Block| MPer| NPer| K0Per| K1| MPer| NPer| MRepeat| NRepeat| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds| BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CShuffle| CShuffle| CBlockTransferClusterLengths| CBlockTransfer|
|
||||
//################################| Layout| Layout| Layout| Layout| Type| Type| Type| DataType| Type| Type| Elementwise| Elementwise| Elementwise| Specialization| Stage| Size| Block| Block| Block| | WMMA| WMMA| | | ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraM| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraN| MXdlPerWave| NXdlPerWave| _MBlock_MWaveMPerXdl| ScalarPerVector|
|
||||
//################################| | | | | | | | | | | Operation| Operation| Operation| | | | | | | | | | | | Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| | Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerShuffle| PerShuffle| _NBlock_NWaveNPerXdl| _NWaveNPerXdl|
|
||||
//################################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Row, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmDefault, 1, 256, 128, 128, 64, 16, 16, 16, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, 1, 2, S<1, 32, 1, 8>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Row, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmDefault, 1, 128, 64, 64, 64, 16, 16, 16, 2, 2, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, 1, 2, S<1, 32, 1, 4>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Row, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmDefault, 1, 64, 32, 32, 64, 16, 16, 16, 1, 2, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, 1, 2, S<1, 32, 1, 2>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Row, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmDefault, 1, 32, 16, 16, 64, 16, 16, 16, 1, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, S<4, 8, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
|
||||
// M/N/K padding
|
||||
//################################| A| B| Ds| E| AData| BData| DsData| EData| AccData| CShuffle| A| B| CDE| GEMM| Block| MPer| NPer| K0Per| K1| MPer| NPer| MRepeat| NRepeat| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds| BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CShuffle| CShuffle| CBlockTransferClusterLengths| CBlockTransfer|
|
||||
//################################| Layout| Layout| Layout| Layout| Type| Type| Type| Type| Type| DataType| Elementwise| Elementwise| Elementwise| Specialization| Size| Block| Block| Block| | WMMA| WMMA| | | ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraM| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraN| MXdlPerWave| NXdlPerWave| _MBlock_MWaveMPerXdl| ScalarPerVector|
|
||||
//################################| | | | | | | | | | | Operation| Operation| Operation| | | | | | | | | | | Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| | Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerShuffle| PerShuffle| _NBlock_NWaveNPerXdl| _NWaveNPerXdl|
|
||||
//################################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Row, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 256, 128, 128, 4, 16, 16, 16, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, 1, 2, S<1, 32, 1, 8>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Row, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 128, 64, 64, 4, 16, 16, 16, 2, 2, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, 1, 2, S<1, 32, 1, 4>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Row, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 64, 32, 32, 4, 16, 16, 16, 1, 2, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, 1, 2, S<1, 32, 1, 2>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Row, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 32, 16, 16, 4, 16, 16, 16, 1, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, S<4, 8, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Row, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 256, 128, 128, 8, 8, 16, 16, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, 1, 2, S<1, 32, 1, 8>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Row, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 128, 64, 64, 8, 8, 16, 16, 2, 2, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, 1, 2, S<1, 32, 1, 4>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Row, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 64, 32, 32, 8, 8, 16, 16, 1, 2, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, 1, 2, S<1, 32, 1, 2>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Row, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 32, 16, 16, 8, 8, 16, 16, 1, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 8, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Row, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 256, 128, 128, 8, 4, 16, 16, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, 1, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 4, 1, 1, 2, S<1, 32, 1, 8>, 4>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Row, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 128, 64, 64, 8, 4, 16, 16, 2, 2, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, 1, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 4, 1, 1, 2, S<1, 32, 1, 4>, 4>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Row, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 64, 32, 32, 8, 4, 16, 16, 1, 2, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, 1, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 4, 1, 1, 2, S<1, 32, 1, 2>, 4>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Row, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 32, 16, 16, 8, 4, 16, 16, 1, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, 1, S<4, 8, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 4, 1, 1, 1, S<1, 16, 1, 2>, 4>
|
||||
//################################| A| B| Ds| E| AData| BData| AccData| CShuffle| DsData| EData| A| B| CDE| GEMM| Prefetch| Block| MPer| NPer| K0Per| K1| MPer| NPer| MRepeat| NRepeat| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds| BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CShuffle| CShuffle| CBlockTransferClusterLengths| CBlockTransfer|
|
||||
//################################| Layout| Layout| Layout| Layout| Type| Type| Type| DataType| Type| Type| Elementwise| Elementwise| Elementwise| Specialization| Stage| Size| Block| Block| Block| | WMMA| WMMA| | | ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraM| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraN| MXdlPerWave| NXdlPerWave| _MBlock_MWaveMPerXdl| ScalarPerVector|
|
||||
//################################| | | | | | | | | | | Operation| Operation| Operation| | | | | | | | | | | | Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| | Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerShuffle| PerShuffle| _NBlock_NWaveNPerXdl| _NWaveNPerXdl|
|
||||
//################################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Row, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 256, 128, 128, 64, 16, 16, 16, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, 1, 2, S<1, 32, 1, 8>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Row, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 128, 64, 64, 64, 16, 16, 16, 2, 2, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, 1, 2, S<1, 32, 1, 4>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Row, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 64, 32, 32, 64, 16, 16, 16, 1, 2, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, 1, 2, S<1, 32, 1, 2>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Row, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 32, 16, 16, 64, 16, 16, 16, 1, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, S<4, 8, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 16, 1, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Row, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 256, 128, 128, 64, 8, 16, 16, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, 1, 2, S<1, 32, 1, 8>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Row, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 128, 64, 64, 64, 8, 16, 16, 2, 2, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, 1, 2, S<1, 32, 1, 4>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Row, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 64, 32, 32, 64, 8, 16, 16, 1, 2, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, 1, 2, S<1, 32, 1, 2>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Row, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 32, 16, 16, 64, 8, 16, 16, 1, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 8, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 8, 1, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Row, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 256, 128, 128, 32, 4, 16, 16, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, 1, S<4, 64, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 4, 1, 1, 2, S<1, 32, 1, 8>, 4>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Row, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 128, 64, 64, 32, 4, 16, 16, 2, 2, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, 1, S<4, 32, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 4, 1, 1, 2, S<1, 32, 1, 4>, 4>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Row, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 64, 32, 32, 32, 4, 16, 16, 1, 2, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, 1, S<4, 16, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 4, 1, 1, 2, S<1, 32, 1, 2>, 4>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Row, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 32, 16, 16, 32, 4, 16, 16, 1, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, 1, S<4, 8, 1>, S<0, 2, 1>, S<0, 2, 1>, 1, 2, 4, 1, 1, 1, S<1, 16, 1, 2>, 4>
|
||||
|
||||
// clang-format on
|
||||
>;
|
||||
|
||||
@@ -38,56 +38,56 @@ using device_gemm_bilinear_wmma_c_shuffle_i8_i8_i8_i8_mk_nk_mn_mn_instances = st
|
||||
// clang-format off
|
||||
// no padding
|
||||
// N % 16 == 0 && K % 16 == 0
|
||||
//################################| A| B| Ds| E| AData| BData| DsData| EData| AccData| CShuffle| A| B| CDE| GEMM| Block| MPer| NPer| K0Per| K1| MPer| NPer| MRepeat| NRepeat| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds| BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CShuffle| CShuffle| CBlockTransferClusterLengths| CBlockTransfer|
|
||||
//################################| Layout| Layout| Layout| Layout| Type| Type| Type| Type| Type| DataType| Elementwise| Elementwise| Elementwise| Specialization| Size| Block| Block| Block| | WMMA| WMMA| | | ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraM| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraN| MXdlPerWave| NXdlPerWave| _MBlock_MWaveMPerXdl| ScalarPerVector|
|
||||
//################################| | | | | | | | | | | Operation| Operation| Operation| | | | | | | | | | | Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| | Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerShuffle| PerShuffle| _NBlock_NWaveNPerXdl| _NWaveNPerXdl|
|
||||
//################################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmDefault, 256, 128, 128, 4, 16, 16, 16, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, 1, 2, S<1, 32, 1, 8>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmDefault, 128, 64, 64, 4, 16, 16, 16, 2, 2, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, 1, 2, S<1, 32, 1, 4>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmDefault, 64, 32, 32, 4, 16, 16, 16, 1, 2, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, 1, 2, S<1, 32, 1, 2>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmDefault, 32, 16, 16, 4, 16, 16, 16, 1, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
//################################| A| B| Ds| E| AData| BData| AccData| CShuffle| DsData| EData| A| B| CDE| GEMM| Prefetch| Block| MPer| NPer| K0Per| K1| MPer| NPer| MRepeat| NRepeat| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds| BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CShuffle| CShuffle| CBlockTransferClusterLengths| CBlockTransfer|
|
||||
//################################| Layout| Layout| Layout| Layout| Type| Type| Type| DataType| Type| Type| Elementwise| Elementwise| Elementwise| Specialization| Stage| Size| Block| Block| Block| | WMMA| WMMA| | | ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraM| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraN| MXdlPerWave| NXdlPerWave| _MBlock_MWaveMPerXdl| ScalarPerVector|
|
||||
//################################| | | | | | | | | | | Operation| Operation| Operation| | | | | | | | | | | | Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| | Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerShuffle| PerShuffle| _NBlock_NWaveNPerXdl| _NWaveNPerXdl|
|
||||
//################################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmDefault, 1, 256, 128, 128, 64, 16, 16, 16, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, 1, 2, S<1, 32, 1, 8>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmDefault, 1, 128, 64, 64, 64, 16, 16, 16, 2, 2, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, 1, 2, S<1, 32, 1, 4>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmDefault, 1, 64, 32, 32, 64, 16, 16, 16, 1, 2, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, 1, 2, S<1, 32, 1, 2>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmDefault, 1, 32, 16, 16, 64, 16, 16, 16, 1, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
// M/N/K padding
|
||||
// N % 16 == 0 && K % 16 == 0
|
||||
//################################| A| B| Ds| E| AData| BData| DsData| EData| AccData| CShuffle| A| B| CDE| GEMM| Block| MPer| NPer| K0Per| K1| MPer| NPer| MRepeat| NRepeat| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds| BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CShuffle| CShuffle| CBlockTransferClusterLengths| CBlockTransfer|
|
||||
//################################| Layout| Layout| Layout| Layout| Type| Type| Type| Type| Type| DataType| Elementwise| Elementwise| Elementwise| Specialization| Size| Block| Block| Block| | WMMA| WMMA| | | ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraM| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraN| MXdlPerWave| NXdlPerWave| _MBlock_MWaveMPerXdl| ScalarPerVector|
|
||||
//################################| | | | | | | | | | | Operation| Operation| Operation| | | | | | | | | | | Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| | Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerShuffle| PerShuffle| _NBlock_NWaveNPerXdl| _NWaveNPerXdl|
|
||||
//################################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 256, 128, 128, 4, 16, 16, 16, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, 1, 2, S<1, 32, 1, 8>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 128, 64, 64, 4, 16, 16, 16, 2, 2, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, 1, 2, S<1, 32, 1, 4>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 64, 32, 32, 4, 16, 16, 16, 1, 2, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, 1, 2, S<1, 32, 1, 2>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 32, 16, 16, 4, 16, 16, 16, 1, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
//################################| A| B| Ds| E| AData| BData| AccData| CShuffle| DsData| EData| A| B| CDE| GEMM| Prefetch| Block| MPer| NPer| K0Per| K1| MPer| NPer| MRepeat| NRepeat| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds| BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CShuffle| CShuffle| CBlockTransferClusterLengths| CBlockTransfer|
|
||||
//################################| Layout| Layout| Layout| Layout| Type| Type| Type| DataType| Type| Type| Elementwise| Elementwise| Elementwise| Specialization| Stage| Size| Block| Block| Block| | WMMA| WMMA| | | ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraM| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraN| MXdlPerWave| NXdlPerWave| _MBlock_MWaveMPerXdl| ScalarPerVector|
|
||||
//################################| | | | | | | | | | | Operation| Operation| Operation| | | | | | | | | | | | Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| | Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerShuffle| PerShuffle| _NBlock_NWaveNPerXdl| _NWaveNPerXdl|
|
||||
//################################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 256, 128, 128, 64, 16, 16, 16, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, 1, 2, S<1, 32, 1, 8>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 128, 64, 64, 64, 16, 16, 16, 2, 2, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, 1, 2, S<1, 32, 1, 4>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 64, 32, 32, 64, 16, 16, 16, 1, 2, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, 1, 2, S<1, 32, 1, 2>, 16>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 32, 16, 16, 64, 16, 16, 16, 1, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 16, 16, 1, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
// M/N/K padding
|
||||
// N % 8 == 0 && K % 8 == 0
|
||||
//################################| A| B| Ds| E| AData| BData| DsData| EData| AccData| CShuffle| A| B| CDE| GEMM| Block| MPer| NPer| K0Per| K1| MPer| NPer| MRepeat| NRepeat| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds| BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CShuffle| CShuffle| CBlockTransferClusterLengths| CBlockTransfer|
|
||||
//################################| Layout| Layout| Layout| Layout| Type| Type| Type| Type| Type| DataType| Elementwise| Elementwise| Elementwise| Specialization| Size| Block| Block| Block| | WMMA| WMMA| | | ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraM| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraN| MXdlPerWave| NXdlPerWave| _MBlock_MWaveMPerXdl| ScalarPerVector|
|
||||
//################################| | | | | | | | | | | Operation| Operation| Operation| | | | | | | | | | | Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| | Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerShuffle| PerShuffle| _NBlock_NWaveNPerXdl| _NWaveNPerXdl|
|
||||
//################################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 256, 128, 128, 8, 8, 16, 16, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, 1, 2, S<1, 32, 1, 8>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 128, 64, 64, 8, 8, 16, 16, 2, 2, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, 1, 2, S<1, 32, 1, 4>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 64, 32, 32, 8, 8, 16, 16, 1, 2, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, 1, 2, S<1, 32, 1, 2>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 32, 16, 16, 8, 8, 16, 16, 1, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
//################################| A| B| Ds| E| AData| BData| AccData| CShuffle| DsData| EData| A| B| CDE| GEMM| Prefetch| Block| MPer| NPer| K0Per| K1| MPer| NPer| MRepeat| NRepeat| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds| BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CShuffle| CShuffle| CBlockTransferClusterLengths| CBlockTransfer|
|
||||
//################################| Layout| Layout| Layout| Layout| Type| Type| Type| DataType| Type| Type| Elementwise| Elementwise| Elementwise| Specialization| Stage| Size| Block| Block| Block| | WMMA| WMMA| | | ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraM| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraN| MXdlPerWave| NXdlPerWave| _MBlock_MWaveMPerXdl| ScalarPerVector|
|
||||
//################################| | | | | | | | | | | Operation| Operation| Operation| | | | | | | | | | | | Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| | Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerShuffle| PerShuffle| _NBlock_NWaveNPerXdl| _NWaveNPerXdl|
|
||||
//################################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 256, 128, 128, 64, 8, 16, 16, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, 1, 2, S<1, 32, 1, 8>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 128, 64, 64, 64, 8, 16, 16, 2, 2, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, 1, 2, S<1, 32, 1, 4>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 64, 32, 32, 64, 8, 16, 16, 1, 2, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, 1, 2, S<1, 32, 1, 2>, 8>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 32, 16, 16, 64, 8, 16, 16, 1, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, 1, 1, S<1, 16, 1, 2>, 8>,
|
||||
|
||||
// M/N/K padding
|
||||
// N % 8 == 0 && K % 8 == 0
|
||||
//################################| A| B| Ds| E| AData| BData| DsData| EData| AccData| CShuffle| A| B| CDE| GEMM| Block| MPer| NPer| K0Per| K1| MPer| NPer| MRepeat| NRepeat| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds| BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CShuffle| CShuffle| CBlockTransferClusterLengths| CBlockTransfer|
|
||||
//################################| Layout| Layout| Layout| Layout| Type| Type| Type| Type| Type| DataType| Elementwise| Elementwise| Elementwise| Specialization| Size| Block| Block| Block| | WMMA| WMMA| | | ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraM| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraN| MXdlPerWave| NXdlPerWave| _MBlock_MWaveMPerXdl| ScalarPerVector|
|
||||
//################################| | | | | | | | | | | Operation| Operation| Operation| | | | | | | | | | | Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| | Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerShuffle| PerShuffle| _NBlock_NWaveNPerXdl| _NWaveNPerXdl|
|
||||
//################################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 256, 128, 128, 8, 4, 16, 16, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, 1, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, 1, 1, 2, S<1, 32, 1, 8>, 4>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 128, 64, 64, 8, 4, 16, 16, 2, 2, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, 1, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, 1, 1, 2, S<1, 32, 1, 4>, 4>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 64, 32, 32, 8, 4, 16, 16, 1, 2, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, 1, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, 1, 1, 2, S<1, 32, 1, 2>, 4>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 32, 16, 16, 8, 4, 16, 16, 1, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, 1, 1, 1, S<1, 16, 1, 2>, 4>,
|
||||
//################################| A| B| Ds| E| AData| BData| AccData| CShuffle| DsData| EData| A| B| CDE| GEMM| Prefetch| Block| MPer| NPer| K0Per| K1| MPer| NPer| MRepeat| NRepeat| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds| BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CShuffle| CShuffle| CBlockTransferClusterLengths| CBlockTransfer|
|
||||
//################################| Layout| Layout| Layout| Layout| Type| Type| Type| DataType| Type| Type| Elementwise| Elementwise| Elementwise| Specialization| Stage| Size| Block| Block| Block| | WMMA| WMMA| | | ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraM| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraN| MXdlPerWave| NXdlPerWave| _MBlock_MWaveMPerXdl| ScalarPerVector|
|
||||
//################################| | | | | | | | | | | Operation| Operation| Operation| | | | | | | | | | | | Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| | Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerShuffle| PerShuffle| _NBlock_NWaveNPerXdl| _NWaveNPerXdl|
|
||||
//################################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 256, 128, 128, 32, 4, 16, 16, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, 1, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, 1, 1, 2, S<1, 32, 1, 8>, 4>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 128, 64, 64, 32, 4, 16, 16, 2, 2, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, 1, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, 1, 1, 2, S<1, 32, 1, 4>, 4>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 64, 32, 32, 32, 4, 16, 16, 1, 2, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, 1, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, 1, 1, 2, S<1, 32, 1, 2>, 4>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 32, 16, 16, 32, 4, 16, 16, 1, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 4, 4, 1, 1, 1, S<1, 16, 1, 2>, 4>,
|
||||
|
||||
// M/N/K padding
|
||||
// N % 1 == 0 && K % 8 == 0
|
||||
//################################| A| B| Ds| E| AData| BData| DsData| EData| AccData| CShuffle| A| B| CDE| GEMM| Block| MPer| NPer| K0Per| K1| MPer| NPer| MRepeat| NRepeat| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds| BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CShuffle| CShuffle| CBlockTransferClusterLengths| CBlockTransfer|
|
||||
//################################| Layout| Layout| Layout| Layout| Type| Type| Type| Type| Type| DataType| Elementwise| Elementwise| Elementwise| Specialization| Size| Block| Block| Block| | WMMA| WMMA| | | ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraM| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraN| MXdlPerWave| NXdlPerWave| _MBlock_MWaveMPerXdl| ScalarPerVector|
|
||||
//################################| | | | | | | | | | | Operation| Operation| Operation| | | | | | | | | | | Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| | Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerShuffle| PerShuffle| _NBlock_NWaveNPerXdl| _NWaveNPerXdl|
|
||||
//################################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 256, 128, 128, 8, 8, 16, 16, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, 1, 2, S<1, 32, 1, 8>, 1>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 128, 64, 64, 8, 8, 16, 16, 2, 2, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, 1, 2, S<1, 32, 1, 4>, 1>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 64, 32, 32, 8, 8, 16, 16, 1, 2, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, 1, 2, S<1, 32, 1, 2>, 1>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I8_Tuple, I8, I32, I32, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 32, 16, 16, 8, 8, 16, 16, 1, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, 1, 1, S<1, 16, 1, 2>, 1>
|
||||
//################################| A| B| Ds| E| AData| BData| AccData| CShuffle| DsData| EData| A| B| CDE| GEMM| Prefetch| Block| MPer| NPer| K0Per| K1| MPer| NPer| MRepeat| NRepeat| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds| BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CShuffle| CShuffle| CBlockTransferClusterLengths| CBlockTransfer|
|
||||
//################################| Layout| Layout| Layout| Layout| Type| Type| Type| DataType| Type| Type| Elementwise| Elementwise| Elementwise| Specialization| Stage| Size| Block| Block| Block| | WMMA| WMMA| | | ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraM| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| AddExtraN| MXdlPerWave| NXdlPerWave| _MBlock_MWaveMPerXdl| ScalarPerVector|
|
||||
//################################| | | | | | | | | | | Operation| Operation| Operation| | | | | | | | | | | | Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| | Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerShuffle| PerShuffle| _NBlock_NWaveNPerXdl| _NWaveNPerXdl|
|
||||
//################################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 256, 128, 128, 64, 8, 16, 16, 4, 2, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, 1, 2, S<1, 32, 1, 8>, 1>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 128, 64, 64, 64, 8, 16, 16, 2, 2, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, 1, 2, S<1, 32, 1, 4>, 1>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 64, 32, 32, 64, 8, 16, 16, 1, 2, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<4, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, 1, 2, S<1, 32, 1, 2>, 1>,
|
||||
DeviceGemmMultipleD_Wmma_CShuffle< Row, Col, Row_Tuple, Row, I8, I8, I32, I32, I8_Tuple, I8, PassThrough, PassThrough, Bilinear, GemmMNKPadding, 1, 32, 16, 16, 64, 8, 16, 16, 1, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<2, 16, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, 1, 1, S<1, 16, 1, 2>, 1>
|
||||
|
||||
// clang-format on
|
||||
>;
|
||||
|
||||
@@ -1,16 +1,18 @@
|
||||
add_instance_library(device_grouped_conv2d_bwd_data_instance
|
||||
xdl/device_grouped_conv2d_bwd_data_xdl_gnhwc_gkyxc_gnhwk_f16_instance.cpp
|
||||
xdl/device_grouped_conv2d_bwd_data_xdl_gnhwc_gkyxc_gnhwk_bf16_instance.cpp
|
||||
xdl/device_grouped_conv2d_bwd_data_xdl_gnhwc_gkyxc_gnhwk_f32_instance.cpp
|
||||
xdl/device_grouped_conv2d_bwd_data_xdl_nhwgc_gkyxc_nhwgk_f16_instance.cpp
|
||||
xdl/device_grouped_conv2d_bwd_data_xdl_nhwgc_gkyxc_nhwgk_bf16_instance.cpp
|
||||
xdl/device_grouped_conv2d_bwd_data_xdl_nhwgc_gkyxc_nhwgk_f32_instance.cpp
|
||||
add_instance_library(
|
||||
device_grouped_conv2d_bwd_data_instance
|
||||
xdl/device_grouped_conv2d_bwd_data_xdl_gnhwc_gkyxc_gnhwk_f16_instance.cpp
|
||||
xdl/device_grouped_conv2d_bwd_data_xdl_gnhwc_gkyxc_gnhwk_bf16_instance.cpp
|
||||
xdl/device_grouped_conv2d_bwd_data_xdl_gnhwc_gkyxc_gnhwk_f32_instance.cpp
|
||||
xdl/device_grouped_conv2d_bwd_data_xdl_nhwgc_gkyxc_nhwgk_f16_instance.cpp
|
||||
xdl/device_grouped_conv2d_bwd_data_xdl_nhwgc_gkyxc_nhwgk_bf16_instance.cpp
|
||||
xdl/device_grouped_conv2d_bwd_data_xdl_nhwgc_gkyxc_nhwgk_f32_instance.cpp
|
||||
|
||||
wmma/device_grouped_conv2d_bwd_data_wmma_gnhwc_gkyxc_gnhwk_f16_1x1s1p0_instance.cpp
|
||||
wmma/device_grouped_conv2d_bwd_data_wmma_nhwgc_gkyxc_nhwgk_f16_1x1s1p0_instance.cpp
|
||||
wmma/device_grouped_conv2d_bwd_data_wmma_gnhwc_gkyxc_gnhwk_i8_1x1s1p0_instance.cpp
|
||||
wmma/device_grouped_conv2d_bwd_data_wmma_nhwgc_gkyxc_nhwgk_i8_1x1s1p0_instance.cpp
|
||||
wmma/device_grouped_conv2d_bwd_data_wmma_gnhwc_gkyxc_gnhwk_f16_instance.cpp
|
||||
wmma/device_grouped_conv2d_bwd_data_wmma_nhwgc_gkyxc_nhwgk_f16_instance.cpp
|
||||
wmma/device_grouped_conv2d_bwd_data_wmma_gnhwc_gkyxc_gnhwk_i8_instance.cpp
|
||||
wmma/device_grouped_conv2d_bwd_data_wmma_nhwgc_gkyxc_nhwgk_i8_instance.cpp)
|
||||
wmma/device_grouped_conv2d_bwd_data_wmma_gnhwc_gkyxc_gnhwk_f16_1x1s1p0_instance.cpp
|
||||
wmma/device_grouped_conv2d_bwd_data_wmma_nhwgc_gkyxc_nhwgk_f16_1x1s1p0_instance.cpp
|
||||
wmma/device_grouped_conv2d_bwd_data_wmma_gnhwc_gkyxc_gnhwk_i8_1x1s1p0_instance.cpp
|
||||
wmma/device_grouped_conv2d_bwd_data_wmma_nhwgc_gkyxc_nhwgk_i8_1x1s1p0_instance.cpp
|
||||
wmma/device_grouped_conv2d_bwd_data_wmma_gnhwc_gkyxc_gnhwk_f16_instance.cpp
|
||||
wmma/device_grouped_conv2d_bwd_data_wmma_nhwgc_gkyxc_nhwgk_f16_instance.cpp
|
||||
wmma/device_grouped_conv2d_bwd_data_wmma_gnhwc_gkyxc_gnhwk_i8_instance.cpp
|
||||
wmma/device_grouped_conv2d_bwd_data_wmma_nhwgc_gkyxc_nhwgk_i8_instance.cpp
|
||||
)
|
||||
|
||||
@@ -17,21 +17,21 @@ add_instance_library(device_grouped_conv2d_fwd_instance
|
||||
dl/device_grouped_conv2d_fwd_dl_nhwgc_gkyxc_nhwgk_f32_instance.cpp
|
||||
# WMMA
|
||||
# GNHWC, GKYXC, GNHWK
|
||||
wmma/device_grouped_conv2d_fwd_wmma_gnhwc_gkyxc_gnhwk_f16_instance.cpp
|
||||
wmma/device_grouped_conv2d_fwd_wmma_gnhwc_gkyxc_gnhwk_i8_instance.cpp
|
||||
wmma/device_grouped_conv2d_fwd_wmma_gnhwc_gkyxc_gnhwk_f16_1x1p0_instance.cpp
|
||||
wmma/device_grouped_conv2d_fwd_wmma_gnhwc_gkyxc_gnhwk_i8_1x1p0_instance.cpp
|
||||
wmma/device_grouped_conv2d_fwd_wmma_gnhwc_gkyxc_gnhwk_f16_1x1s1p0_instance.cpp
|
||||
wmma/device_grouped_conv2d_fwd_wmma_gnhwc_gkyxc_gnhwk_i8_1x1s1p0_instance.cpp
|
||||
wmma/device_grouped_conv2d_fwd_wmma_gnhwc_gkyxc_gnhwk_f16_oddc_instance.cpp
|
||||
wmma/device_grouped_conv2d_fwd_wmma_gnhwc_gkyxc_gnhwk_i8_oddc_instance.cpp
|
||||
# NHWGC, GKYXC, NHWGK
|
||||
wmma/device_grouped_conv2d_fwd_wmma_nhwgc_gkyxc_nhwgk_f16_instance.cpp
|
||||
wmma/device_grouped_conv2d_fwd_wmma_nhwgc_gkyxc_nhwgk_i8_instance.cpp
|
||||
wmma/device_grouped_conv2d_fwd_wmma_nhwgc_gkyxc_nhwgk_f16_1x1p0_instance.cpp
|
||||
wmma/device_grouped_conv2d_fwd_wmma_nhwgc_gkyxc_nhwgk_i8_1x1p0_instance.cpp
|
||||
wmma/device_grouped_conv2d_fwd_wmma_nhwgc_gkyxc_nhwgk_f16_1x1s1p0_instance.cpp
|
||||
wmma/device_grouped_conv2d_fwd_wmma_nhwgc_gkyxc_nhwgk_i8_1x1s1p0_instance.cpp
|
||||
wmma/device_grouped_conv2d_fwd_wmma_nhwgc_gkyxc_nhwgk_f16_oddc_instance.cpp
|
||||
wmma/device_grouped_conv2d_fwd_wmma_nhwgc_gkyxc_nhwgk_i8_oddc_instance.cpp
|
||||
wmma/device_grouped_conv2d_fwd_wmma_gnhwc_gkyxc_gnhwk_f16_instance.cpp
|
||||
wmma/device_grouped_conv2d_fwd_wmma_gnhwc_gkyxc_gnhwk_i8_instance.cpp
|
||||
wmma/device_grouped_conv2d_fwd_wmma_gnhwc_gkyxc_gnhwk_f16_1x1p0_instance.cpp
|
||||
wmma/device_grouped_conv2d_fwd_wmma_gnhwc_gkyxc_gnhwk_i8_1x1p0_instance.cpp
|
||||
wmma/device_grouped_conv2d_fwd_wmma_gnhwc_gkyxc_gnhwk_f16_1x1s1p0_instance.cpp
|
||||
wmma/device_grouped_conv2d_fwd_wmma_gnhwc_gkyxc_gnhwk_i8_1x1s1p0_instance.cpp
|
||||
wmma/device_grouped_conv2d_fwd_wmma_gnhwc_gkyxc_gnhwk_f16_oddc_instance.cpp
|
||||
wmma/device_grouped_conv2d_fwd_wmma_gnhwc_gkyxc_gnhwk_i8_oddc_instance.cpp
|
||||
## NHWGC, GKYXC, NHWGK
|
||||
wmma/device_grouped_conv2d_fwd_wmma_nhwgc_gkyxc_nhwgk_f16_instance.cpp
|
||||
wmma/device_grouped_conv2d_fwd_wmma_nhwgc_gkyxc_nhwgk_i8_instance.cpp
|
||||
wmma/device_grouped_conv2d_fwd_wmma_nhwgc_gkyxc_nhwgk_f16_1x1p0_instance.cpp
|
||||
wmma/device_grouped_conv2d_fwd_wmma_nhwgc_gkyxc_nhwgk_i8_1x1p0_instance.cpp
|
||||
wmma/device_grouped_conv2d_fwd_wmma_nhwgc_gkyxc_nhwgk_f16_1x1s1p0_instance.cpp
|
||||
wmma/device_grouped_conv2d_fwd_wmma_nhwgc_gkyxc_nhwgk_i8_1x1s1p0_instance.cpp
|
||||
wmma/device_grouped_conv2d_fwd_wmma_nhwgc_gkyxc_nhwgk_f16_oddc_instance.cpp
|
||||
wmma/device_grouped_conv2d_fwd_wmma_nhwgc_gkyxc_nhwgk_i8_oddc_instance.cpp
|
||||
)
|
||||
|
||||
@@ -22,7 +22,8 @@ set(GROUPED_CONV3D_FWD
|
||||
wmma/device_grouped_conv3d_fwd_wmma_gndhwc_gkzyxc_gndhwk_f16_oddc_instance.cpp
|
||||
wmma/device_grouped_conv3d_fwd_wmma_gndhwc_gkzyxc_gndhwk_i8_oddc_instance.cpp
|
||||
wmma/device_grouped_conv3d_fwd_wmma_ndhwgc_gkzyxc_ndhwgk_f16_oddc_instance.cpp
|
||||
wmma/device_grouped_conv3d_fwd_wmma_ndhwgc_gkzyxc_ndhwgk_i8_oddc_instance.cpp)
|
||||
wmma/device_grouped_conv3d_fwd_wmma_ndhwgc_gkzyxc_ndhwgk_i8_oddc_instance.cpp
|
||||
)
|
||||
|
||||
if((DTYPES MATCHES "fp8" AND DTYPES MATCHES "fp16") OR NOT DEFINED DTYPES)
|
||||
list(APPEND GROUPED_CONV3D_FWD
|
||||
|
||||
Reference in New Issue
Block a user