ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-20 12:59:49 +00:00
Code Issues Packages Projects Releases Wiki Activity

[CK_TILE] Row/Col quant gemm (#2729)

Browse Source 
* Add cshuffle epilogue test

* add the poc implementation to the epilogue and tests

* refactor cshuffle epilogue

* WIP: adding tensor/tile usage to scale_tile

* fix usage of tile_elementwise_inout

* add gemm_quant_kernel for generalizing gemm quant kernel

* Add problem specific to different quants, add QuantType to Traits

* Add quant_type to quant_kernel template parameters

* Create aq/bq_block_windows and views depending on QuantType

* Use tile windows as inputs in cshuffle epilogue

* Fix some issues in epilogue

* initial new example code for new general gemm quant kernel test

* Fix issues in kernel

* Add verification check for rowcol Quantmode

* use AccDataType instead of AQ in pipeline

* fix aquant preshuffle

* fix formatting

* some cleanup

* remove gemm_aquant_basic.cpp

* remove gemm_aquant_kernel.hpp

* fix tests for the renamed quant kernel

* fix formatting

* clean example files

* fix some merge conflicts

* fix preshufflequant rename issue

* fix some templates after merging with develop

* fix test preshuffle parameter

* fix formatting

* Unify bquant kernel to the common quant kernel

* remove bquant kernel also from common header

* fix formatting

* clean up commented code

* fix formatting config hpp

* fix merge mistake

* Non-const for movable windows

* fix formatting

* Fix grammar in README

Co-authored-by: spolifroni-amd <Sandra.Polifroni@amd.com>

* Remove #include<bit> and clean up example

* fix strides

* Add some descriptions for move_windows

---------

Co-authored-by: Mohsen Saffari <mohsen.saffari@amd.com>
Co-authored-by: spolifroni-amd <Sandra.Polifroni@amd.com>

[ROCm/composable_kernel commit: c6010f2953]
This commit is contained in:
Sami Remes
2025-09-05 02:17:12 +03:00
committed by GitHub
parent ecc4a470ec
commit abf4f7a7b2
23 changed files with 1837 additions and 1331 deletions
Download Patch File Download Diff File Expand all files Collapse all files

86
include/ck_tile/host/reference/reference_gemm.hpp
View File

@@ -115,6 +115,92 @@ CK_TILE_HOST void reference_gemm_quant(const HostTensor<ADataType>& a_m_k,
std::cout << std::endl;
}
template <typename ADataType,
typename AQDataType,
typename BDataType,
typename BQDataType,
typename AccDataType,
typename CDataType,
typename AElementOp = ck_tile::identity,
typename BElementOp = ck_tile::identity,
typename ACCElementOp = ck_tile::identity>
CK_TILE_HOST void reference_gemm_rowcol_quant(const HostTensor<ADataType>& a_m_k,
const HostTensor<AQDataType>& aq_m_1,
const HostTensor<BDataType>& b_k_n,
const HostTensor<BQDataType>& bq_1_n,
HostTensor<CDataType>& c_m_n,
const AElementOp& a_element_op = {},
const BElementOp& b_element_op = {},
const ACCElementOp& acc_element_op = {})
{
static_assert(std::is_same_v<ADataType, fp8_t> || std::is_same_v<ADataType, bf8_t>);
static_assert(std::is_same_v<BDataType, fp8_t> || std::is_same_v<BDataType, bf8_t>);
static_assert(std::is_same_v<AccDataType, float>);
static_assert(std::is_same_v<CDataType, float> || std::is_same_v<CDataType, ck_tile::half_t>);
static_assert(std::is_same_v<AQDataType, float> && std::is_same_v<BQDataType, float>);
const std::size_t M = a_m_k.get_length(0);
const std::size_t N = b_k_n.get_length(1);
const std::size_t K = a_m_k.get_length(1);
auto f_mn = [&](auto m, auto n) {
// Init accumulator
AccDataType v_acc = 0;
// Get row scale for A and column scale for B
float a_scale = aq_m_1(m, 0);
float b_scale = bq_1_n(0, n);
// Compute the dot product
for(std::size_t k = 0; k < K; ++k)
{
AccDataType v_a;
AccDataType v_b;
// Process A data
if constexpr(std::is_same_v<ADataType, pk_int4_t>)
{
const pk_int4_t pk_val = a_element_op(a_m_k(m, k));
const fp32x2_t fp32_val = pk_int4_t_to_fp32x2_t_signed_conversion(pk_val);
if(k % 2 == 1)
v_a = fp32_val.hi;
else
v_a = fp32_val.lo;
}
else
{
v_a = ck_tile::type_convert<AccDataType>(a_element_op(a_m_k(m, k)));
}
// Process B data
if constexpr(std::is_same_v<BDataType, pk_int4_t>)
{
const pk_int4_t pk_val = b_element_op(b_k_n(k, n));
const fp32x2_t fp32_val = pk_int4_t_to_fp32x2_t_signed_conversion(pk_val);
if(k % 2 == 1)
v_b = fp32_val.hi;
else
v_b = fp32_val.lo;
}
else if constexpr(std::is_same_v<BDataType, fp8_t>)
{
v_b = fp8_to_float_raw(b_element_op(b_k_n(k, n)));
}
else
{
v_b = ck_tile::type_convert<AccDataType>(b_element_op(b_k_n(k, n)));
}
v_acc += v_a * v_b;
}
v_acc = v_acc * a_scale * b_scale;
c_m_n(m, n) = ck_tile::type_convert<CDataType>(acc_element_op(v_acc));
};
make_ParallelTensorFunctor(f_mn, M, N)(std::thread::hardware_concurrency());
std::cout << std::endl;
}
template <typename ADataType,
typename BDataType,
typename AccDataType,

203
include/ck_tile/ops/epilogue/cshuffle_epilogue.hpp
View File

@@ -6,6 +6,9 @@
#include "ck_tile/core.hpp"
#include "ck_tile/ops/gemm/warp/warp_gemm_dispatcher.hpp"
#include "ck_tile/ops/common/tensor_layout.hpp"
#include "ck_tile/ops/elementwise/unary_element_wise_operation.hpp"
#include <optional>
namespace ck_tile {
@@ -210,8 +213,12 @@ struct CShuffleEpilogue
KPerXdl,
isCTransposed>;
using CWarpDstr = typename WG::CWarpDstr;
using CWarpTensor = typename WG::CWarpTensor;
using CWarpDstr = typename WG::CWarpDstr;
using CWarpTensor = typename WG::CWarpTensor;
using CWarpDstrEncoding = typename WG::CWarpDstrEncoding;
using SFC = space_filling_curve<sequence<kMPerBlock, kNPerBlock>,
sequence<0, 1>,
sequence<MPerIterationShuffle, NPerIterationShuffle>>;
template <typename Problem>
CK_TILE_HOST_DEVICE static constexpr auto MakeLdsBlockDescriptor()
@@ -257,11 +264,120 @@ struct CShuffleEpilogue
return MPerIterationShuffle * NPerIterationShuffle * sizeof(ODataType);
}
template <typename ODramWindow, typename OAccTile, typename DsDramWindows>
template <auto iAccess, typename LdsTile, typename ScaleM, typename ScaleN>
CK_TILE_DEVICE void
scale_tile(LdsTile& lds_tile, ScaleM& scale_m_window, ScaleN& scale_n_window)
{
// Load tiles
const auto scale_m_tile = load_tile(scale_m_window);
const auto scale_n_tile = load_tile(scale_n_window);
// Compute element-wise product in-place i.e. lds_tile = lds_tile * scale_m * scale_n
tile_elementwise_inout(
element_wise::MultiDMultiply{}, lds_tile, lds_tile, scale_m_tile, scale_n_tile);
// Move scale windows
constexpr index_t num_access = SFC::get_num_of_access();
if constexpr(iAccess != num_access - 1)
{
constexpr auto step = SFC::get_forward_step(iAccess);
move_tile_window(scale_m_window, {step.at(number<0>{}), step.at(number<1>{})});
move_tile_window(scale_n_window, {step.at(number<0>{}), step.at(number<1>{})});
}
}
template <auto iAccess, typename OAccTile, typename LdsTile>
CK_TILE_DEVICE void slice_acc_tile(const OAccTile& o_acc_tile, LdsTile& lds_tile)
{
constexpr auto idx_y_start = SFC::get_index(iAccess);
constexpr auto mIter = number<idx_y_start.at(number<0>{}) / (MPerIterationShuffle)>{};
constexpr auto nIter = number<idx_y_start.at(number<1>{}) / (NPerIterationShuffle)>{};
constexpr auto c_warp_y_lengths =
to_sequence(CWarpDstr{}.get_ys_to_d_descriptor().get_lengths());
constexpr auto c_warp_y_index_zeros = uniform_sequence_gen_t<CWarpDstr::NDimY, 0>{};
lds_tile.get_thread_buffer() = o_acc_tile.get_y_sliced_thread_data(
merge_sequences(
sequence<mIter * NumMXdlPerWavePerShuffle, nIter * NumNXdlPerWavePerShuffle>{},
c_warp_y_index_zeros),
merge_sequences(sequence<NumMXdlPerWavePerShuffle, NumNXdlPerWavePerShuffle>{},
c_warp_y_lengths));
}
template <typename LdsTile, typename InLdsWindow>
CK_TILE_DEVICE void cast_lds_tile(LdsTile& lds_tile, InLdsWindow& in_lds_window)
{
const auto c_warptile_in_tensor_casted = cast_tile<ODataType>(lds_tile);
store_tile(in_lds_window, c_warptile_in_tensor_casted);
}
template <typename DramWindows, typename COutTensor>
CK_TILE_DEVICE void apply_d_tensors(DramWindows& d_dram_windows, COutTensor& c_out_tensor)
{
const auto ds_tensor = generate_tuple(
[&](auto idx) { return load_tile(d_dram_windows[idx]); }, number<NumDTensor>{});
const auto c_ds_tiles = concat_tuple_of_reference(
tie(c_out_tensor, c_out_tensor),
generate_tie([&](auto idx) -> const auto& { return ds_tensor[idx]; },
number<NumDTensor>{}));
tile_elementwise_inout_unpack(typename Problem::CDElementwise{}, c_ds_tiles);
}
template <typename OutDramWindow, typename COutTensor>
CK_TILE_DEVICE void store_to_dram(OutDramWindow& out_dram_window,
const COutTensor& c_out_tensor)
{
if constexpr(MemoryOperation == memory_operation_enum::set)
{
store_tile(out_dram_window, c_out_tensor);
}
else
{
update_tile(out_dram_window, c_out_tensor);
}
}
/**
* @brief Move both the output and D tensors windows for the next access.
*/
template <auto iAccess, typename OutDramWindow, typename DDramWindows>
CK_TILE_DEVICE void move_windows(OutDramWindow& out_dram_window, DDramWindows& d_dram_windows)
{
constexpr index_t num_access = SFC::get_num_of_access();
if constexpr(iAccess != num_access - 1)
{
constexpr auto step = SFC::get_forward_step(iAccess);
// move the output dram window
move_tile_window(out_dram_window, {step.at(number<0>{}), step.at(number<1>{})});
// move windows for each of the D matrices (inputs for element-wise)
static_for<0, NumDTensor, 1>{}([&](auto idx) {
move_tile_window(d_dram_windows[idx], {step.at(number<0>{}), step.at(number<1>{})});
});
}
}
// TODO: Check if there would be nicer ways to overload rather than with EmptyScale or nullptr_t
struct EmptyScale
{
};
template <typename ODramWindow,
typename OAccTile,
typename DsDramWindows,
typename ScaleM = EmptyScale,
typename ScaleN = EmptyScale>
CK_TILE_DEVICE auto operator()(ODramWindow& out_dram_window,
const OAccTile& o_acc_tile,
const DsDramWindows& ds_dram_windows,
void* p_smem)
void* p_smem,
const ScaleM& scale_m = {},
const ScaleN& scale_n = {})
{
constexpr auto LdsTileDistr = make_static_tile_distribution(MakeLdsDistributionEncode());
@@ -282,9 +398,6 @@ struct CShuffleEpilogue
make_tuple(number<MPerIterationShuffle>{}, number<NPerIterationShuffle>{}),
{0, 0});
using SFC = space_filling_curve<sequence<kMPerBlock, kNPerBlock>,
sequence<0, 1>,
sequence<MPerIterationShuffle, NPerIterationShuffle>>;
constexpr index_t num_access = SFC::get_num_of_access();
static_assert(std::is_same_v<ELayout, tensor_layout::gemm::RowMajor>,
@@ -306,60 +419,46 @@ struct CShuffleEpilogue
},
number<NumDTensor>{});
constexpr auto c_warp_y_lengths =
to_sequence(CWarpDstr{}.get_ys_to_d_descriptor().get_lengths());
constexpr auto c_warp_y_index_zeros = uniform_sequence_gen_t<CWarpDstr::NDimY, 0>{};
constexpr bool has_scales =
!std::is_same<ScaleM, EmptyScale>::value && !std::is_same<ScaleN, EmptyScale>::value;
auto scale_m_window = [&]() {
if constexpr(has_scales)
{
return make_tile_window(scale_m, lds_tile.get_tile_distribution());
}
else
{
return EmptyScale{};
}
}();
auto scale_n_window = [&]() {
if constexpr(has_scales)
{
return make_tile_window(scale_n, lds_tile.get_tile_distribution());
}
else
{
return EmptyScale{};
}
}();
static_for<0, num_access, 1>{}([&](auto iAccess) {
block_sync_lds();
constexpr auto idx_y_start = SFC::get_index(iAccess);
slice_acc_tile<iAccess>(o_acc_tile, lds_tile);
constexpr auto mIter = number<idx_y_start.at(number<0>{}) / (MPerIterationShuffle)>{};
constexpr auto nIter = number<idx_y_start.at(number<1>{}) / (NPerIterationShuffle)>{};
if constexpr(has_scales)
{
scale_tile<iAccess>(lds_tile, scale_m_window, scale_n_window);
}
lds_tile.get_thread_buffer() = o_acc_tile.get_y_sliced_thread_data(
merge_sequences(
sequence<mIter * NumMXdlPerWavePerShuffle, nIter * NumNXdlPerWavePerShuffle>{},
c_warp_y_index_zeros),
merge_sequences(sequence<NumMXdlPerWavePerShuffle, NumNXdlPerWavePerShuffle>{},
c_warp_y_lengths));
const auto c_warptile_in_tensor_casted = cast_tile<ODataType>(lds_tile);
store_tile(in_lds_window, c_warptile_in_tensor_casted);
cast_lds_tile(lds_tile, in_lds_window);
block_sync_lds();
auto c_out_tensor = load_tile(make_tile_window(out_lds_window, dram_tile_distribution));
const auto ds_tensor = generate_tuple(
[&](auto idx) { return load_tile(d_dram_windows[idx]); }, number<NumDTensor>{});
const auto c_ds_tiles = concat_tuple_of_reference(
tie(c_out_tensor, c_out_tensor),
generate_tie([&](auto idx) -> const auto& { return ds_tensor[idx]; },
number<NumDTensor>{}));
tile_elementwise_inout_unpack(typename Problem::CDElementwise{}, c_ds_tiles);
if constexpr(MemoryOperation == memory_operation_enum::set)
{
store_tile(out_dram_window, c_out_tensor);
}
else
{
update_tile(out_dram_window, c_out_tensor);
}
if constexpr(iAccess != num_access - 1)
{
constexpr auto step = SFC::get_forward_step(iAccess);
move_tile_window(out_dram_window, {step.at(number<0>{}), step.at(number<1>{})});
static_for<0, NumDTensor, 1>{}([&](auto idx) {
move_tile_window(d_dram_windows[idx],
{step.at(number<0>{}), step.at(number<1>{})});
});
}
apply_d_tensors(d_dram_windows, c_out_tensor);
store_to_dram(out_dram_window, c_out_tensor);
move_windows<iAccess>(out_dram_window, d_dram_windows);
});
}
};

3
include/ck_tile/ops/gemm_group_quant.hpp
View File

@@ -5,8 +5,7 @@
#include "ck_tile/ops/gemm_group_quant/block/block_universal_gemm_as_aquant_bs_cr.hpp"
#include "ck_tile/ops/gemm_group_quant/block/block_universal_gemm_as_bs_bquant_cr.hpp"
#include "ck_tile/ops/gemm_group_quant/kernel/gemm_aquant_kernel.hpp"
#include "ck_tile/ops/gemm_group_quant/kernel/gemm_bquant_kernel.hpp"
#include "ck_tile/ops/gemm_group_quant/kernel/gemm_quant_kernel.hpp"
#include "ck_tile/ops/gemm_group_quant/pipeline/gemm_aquant_pipeline_ag_bg_cr_base.hpp"
#include "ck_tile/ops/gemm_group_quant/pipeline/gemm_aquant_pipeline_ag_bg_cr_policy.hpp"
#include "ck_tile/ops/gemm_group_quant/pipeline/gemm_aquant_pipeline_ag_bg_cr_v3.hpp"

679
include/ck_tile/ops/gemm_group_quant/kernel/gemm_bquant_kernel.hpp
View File

@@ -1,679 +0,0 @@
// SPDX-License-Identifier: MIT
// Copyright (c) 2025, Advanced Micro Devices, Inc. All rights reserved.
#pragma once
#include <iostream>
#include <string>
#include "ck_tile/core.hpp"
#include "ck_tile/ops/common.hpp"
#include "ck_tile/host/concat.hpp"
namespace ck_tile {
struct BQuantGemmProblem
{
CK_TILE_HOST BQuantGemmProblem() = default;
CK_TILE_HOST BQuantGemmProblem(index_t M_,
index_t N_,
index_t K_,
index_t QK_,
index_t stride_A_,
index_t stride_B_,
index_t stride_C_,
index_t stride_BQ_)
: M(M_),
N(N_),
K(K_),
QK(QK_),
stride_A(stride_A_),
stride_B(stride_B_),
stride_C(stride_C_),
stride_BQ(stride_BQ_)
{
}
index_t M;
index_t N;
index_t K;
index_t QK;
index_t stride_A;
index_t stride_B;
index_t stride_C;
index_t stride_BQ;
};
struct BQuantGemmHostArgs : public BQuantGemmProblem
{
CK_TILE_HOST BQuantGemmHostArgs() = default;
CK_TILE_HOST BQuantGemmHostArgs(const void* a_ptr_,
const void* b_ptr_,
void* c_ptr_,
const void* bq_ptr_,
index_t k_batch_,
index_t M_,
index_t N_,
index_t K_,
index_t QK_,
index_t stride_A_,
index_t stride_B_,
index_t stride_C_,
index_t stride_BQ_)
: BQuantGemmProblem(M_, N_, K_, QK_, stride_A_, stride_B_, stride_C_, stride_BQ_),
a_ptr(a_ptr_),
b_ptr(b_ptr_),
bq_ptr(bq_ptr_),
c_ptr(c_ptr_),
k_batch(k_batch_)
{
}
const void* a_ptr;
const void* b_ptr;
const void* bq_ptr;
void* c_ptr;
index_t k_batch;
};
struct BQuantGemmKernelArgs
{
const void* a_ptr;
const void* b_ptr;
const void* bq_ptr;
void* c_ptr;
index_t M;
index_t N;
index_t K;
index_t QK;
index_t stride_A;
index_t stride_B;
index_t stride_C;
index_t stride_BQ;
index_t k_batch;
};
template <typename TilePartitioner_, typename GemmPipeline_, typename EpiloguePipeline_>
struct BQuantGemmKernel
{
using TilePartitioner = remove_cvref_t<TilePartitioner_>;
using GemmPipeline = remove_cvref_t<GemmPipeline_>;
using EpiloguePipeline = remove_cvref_t<EpiloguePipeline_>;
using ALayout = remove_cvref_t<typename GemmPipeline::ALayout>;
using BLayout = remove_cvref_t<typename GemmPipeline::BLayout>;
using BQLayout = remove_cvref_t<typename GemmPipeline::BQLayout>;
using CLayout = remove_cvref_t<typename GemmPipeline::CLayout>;
static constexpr index_t kBlockSize = GemmPipeline::BlockSize;
using ADataType = remove_cvref_t<typename GemmPipeline::ADataType>;
using BDataType = remove_cvref_t<typename GemmPipeline::BDataType>;
using BQDataType = remove_cvref_t<typename GemmPipeline::BQDataType>;
using CDataType = remove_cvref_t<typename EpiloguePipeline::ODataType>;
static constexpr auto I0 = number<0>();
static constexpr auto I1 = number<1>();
static constexpr auto I2 = number<2>();
static constexpr auto I3 = number<3>();
[[nodiscard]] CK_TILE_HOST static const std::string GetName()
{
// clang-format off
return concat('_', "gemm", gemm_prec_str<ADataType, BDataType>, GemmPipeline::GetName());
// clang-format on
}
CK_TILE_HOST static constexpr auto GridSize(index_t M, index_t N, index_t KBatch)
{
return dim3(TilePartitioner::GridSize(M, N), 1, KBatch);
}
CK_TILE_HOST static constexpr auto BlockSize() { return dim3(kBlockSize); }
CK_TILE_HOST static constexpr BQuantGemmKernelArgs
MakeKernelArgs(const BQuantGemmHostArgs& hostArgs)
{
return BQuantGemmKernelArgs{hostArgs.a_ptr,
hostArgs.b_ptr,
hostArgs.bq_ptr,
hostArgs.c_ptr,
hostArgs.M,
hostArgs.N,
hostArgs.K,
hostArgs.QK,
hostArgs.stride_A,
hostArgs.stride_B,
hostArgs.stride_C,
hostArgs.stride_BQ,
hostArgs.k_batch};
}
CK_TILE_HOST_DEVICE static constexpr index_t GetSmemSize()
{
return max(GemmPipeline::GetSmemSize(), EpiloguePipeline::GetSmemSize());
}
struct SplitKBatchOffset
{
__device__ SplitKBatchOffset(const BQuantGemmKernelArgs& kargs,
const std::size_t k_id = blockIdx.z)
{
constexpr auto K1 = TilePartitioner::BlockGemmShape::WarpTile::at(number<2>{});
const index_t K_t = __builtin_amdgcn_readfirstlane(kargs.k_batch * K1);
const index_t KRead = __builtin_amdgcn_readfirstlane((kargs.K + K_t - 1) / K_t * K1);
if constexpr(std::is_same_v<tensor_layout::gemm::RowMajor, ALayout>)
{
a_k_split_offset = __builtin_amdgcn_readfirstlane(k_id * KRead);
}
else if constexpr(std::is_same_v<tensor_layout::gemm::ColumnMajor, ALayout>)
{
a_k_split_offset = __builtin_amdgcn_readfirstlane(k_id * KRead * kargs.stride_A);
}
if constexpr(std::is_same_v<tensor_layout::gemm::RowMajor, BLayout>)
{
b_k_split_offset = __builtin_amdgcn_readfirstlane(k_id * KRead * kargs.stride_B);
}
else if constexpr(std::is_same_v<tensor_layout::gemm::ColumnMajor, BLayout>)
{
b_k_split_offset = __builtin_amdgcn_readfirstlane(k_id * KRead);
}
if(k_id < static_cast<uint32_t>(kargs.k_batch - 1))
{
splitted_k = __builtin_amdgcn_readfirstlane(KRead);
}
else
{
splitted_k = __builtin_amdgcn_readfirstlane(kargs.K - KRead * (kargs.k_batch - 1));
}
}
index_t a_k_split_offset;
index_t b_k_split_offset;
index_t splitted_k;
};
CK_TILE_HOST static bool IsSupportedArgument(const BQuantGemmKernelArgs& kargs)
{
if(kargs.k_batch != 1)
{
if(ck_tile::EnvIsEnabled(CK_TILE_ENV(CK_TILE_LOGGING)))
{
CK_TILE_ERROR("Conditions not met for Kbatch >1 !");
}
return false;
}
static_assert(std::is_same_v<BQLayout, tensor_layout::gemm::ColumnMajor>);
if(kargs.QK % GemmPipeline::GetVectorSizeBQ() != 0)
{
if(ck_tile::EnvIsEnabled(CK_TILE_ENV(CK_TILE_LOGGING)))
{
CK_TILE_ERROR("K is not a multiple of vector load size for A tensor!");
}
return false;
}
if constexpr(std::is_same_v<ALayout, tensor_layout::gemm::RowMajor>)
{
if(kargs.K % (TilePartitioner::KPerBlock * kargs.k_batch) != 0 &&
GemmPipeline::kPadK == false)
{
if(ck_tile::EnvIsEnabled(CK_TILE_ENV(CK_TILE_LOGGING)))
{
CK_TILE_ERROR("Can't support K that is not a multiple of k_batch * KPerBlock "
"without padding!");
}
return false;
}
if(kargs.K % GemmPipeline::GetVectorSizeA() != 0)
{
if(ck_tile::EnvIsEnabled(CK_TILE_ENV(CK_TILE_LOGGING)))
{
CK_TILE_ERROR("K is not a multiple of vector load size for A tensor!");
}
return false;
}
}
else
{
if(kargs.M % TilePartitioner::MPerBlock != 0 && GemmPipeline::kPadM == false)
{
if(ck_tile::EnvIsEnabled(CK_TILE_ENV(CK_TILE_LOGGING)))
{
CK_TILE_ERROR(
"Can't support M that is not a multiple of MPerBlock without padding!");
}
return false;
}
if(kargs.M % GemmPipeline::GetVectorSizeA() != 0)
{
if(ck_tile::EnvIsEnabled(CK_TILE_ENV(CK_TILE_LOGGING)))
{
CK_TILE_ERROR("M is not a multiple of vector load size for A tensor!");
}
return false;
}
}
if constexpr(std::is_same_v<BLayout, tensor_layout::gemm::RowMajor>)
{
if(kargs.N % TilePartitioner::NPerBlock != 0 && GemmPipeline::kPadN == false)
{
if(ck_tile::EnvIsEnabled(CK_TILE_ENV(CK_TILE_LOGGING)))
{
CK_TILE_ERROR(
"Can't support N that is not a multiple of NPerBlock without padding!");
}
return false;
}
if(kargs.N % GemmPipeline::GetVectorSizeB() != 0)
{
if(ck_tile::EnvIsEnabled(CK_TILE_ENV(CK_TILE_LOGGING)))
{
CK_TILE_ERROR("N is not a multiple of vector load size for B tensor!");
}
return false;
}
}
else
{
if(kargs.K % (TilePartitioner::KPerBlock * kargs.k_batch) != 0 &&
GemmPipeline::kPadK == false)
{
if(ck_tile::EnvIsEnabled(CK_TILE_ENV(CK_TILE_LOGGING)))
{
CK_TILE_ERROR("Can't support K that is not a multiple of k_batch * KPerBlock "
"without padding!");
}
return false;
}
if(kargs.K % GemmPipeline::GetVectorSizeB() != 0)
{
if(ck_tile::EnvIsEnabled(CK_TILE_ENV(CK_TILE_LOGGING)))
{
CK_TILE_ERROR("K is not a multiple of vector load size for B tensor!");
}
return false;
}
}
if constexpr(std::is_same_v<CLayout, tensor_layout::gemm::RowMajor>)
{
if(kargs.N % TilePartitioner::NPerBlock != 0 && GemmPipeline::kPadN == false)
{
if(ck_tile::EnvIsEnabled(CK_TILE_ENV(CK_TILE_LOGGING)))
{
CK_TILE_ERROR(
"Can't support N that is not a multiple of NPerBlock without padding!");
}
return false;
}
if(kargs.N % EpiloguePipeline::GetVectorSizeC() != 0)
{
if(ck_tile::EnvIsEnabled(CK_TILE_ENV(CK_TILE_LOGGING)))
{
CK_TILE_ERROR("N is not a multiple of vector load size for C tensor!");
}
return false;
}
}
else
{
if(kargs.M % TilePartitioner::MPerBlock != 0 && GemmPipeline::kPadM == false)
{
if(ck_tile::EnvIsEnabled(CK_TILE_ENV(CK_TILE_LOGGING)))
{
CK_TILE_ERROR(
"Can't support M that is not a multiple of MPerBlock without padding!");
}
return false;
}
if(kargs.M % EpiloguePipeline::GetVectorSizeC() != 0)
{
if(ck_tile::EnvIsEnabled(CK_TILE_ENV(CK_TILE_LOGGING)))
{
CK_TILE_ERROR("M is not a multiple of vector load size for C tensor!");
}
return false;
}
}
return true;
}
template <memory_operation_enum DstInMemOp = memory_operation_enum::set>
CK_TILE_DEVICE static auto MakeGemmTensorViews(const ADataType* a_ptr,
const BDataType* b_ptr,
const BQDataType* bq_ptr,
CDataType* c_ptr,
const BQuantGemmKernelArgs& kargs,
const SplitKBatchOffset& splitk_batch_offset)
{
static_assert(!TilePartitioner::BlockGemmShape::PermuteA, "Not implemented!");
const auto& a_tensor_view = [&]() {
if constexpr(std::is_same_v<ALayout, tensor_layout::gemm::RowMajor>)
{
return make_naive_tensor_view<address_space_enum::global>(
a_ptr,
make_tuple(kargs.M, splitk_batch_offset.splitted_k),
make_tuple(kargs.stride_A, 1),
number<GemmPipeline::GetVectorSizeA()>{},
number<1>{});
}
else
{
return make_naive_tensor_view<address_space_enum::global>(
a_ptr,
make_tuple(splitk_batch_offset.splitted_k, kargs.M),
make_tuple(kargs.stride_A, 1),
number<GemmPipeline::GetVectorSizeA()>{},
number<1>{});
}
}();
const auto& bq_tensor_view = [&]() {
static_assert(std::is_same_v<BQLayout, tensor_layout::gemm::ColumnMajor>);
return make_naive_tensor_view<address_space_enum::global>(
bq_ptr,
make_tuple(kargs.N, kargs.QK),
make_tuple(kargs.stride_BQ, 1),
number<GemmPipeline::GetVectorSizeBQ()>{},
number<1>{});
}();
const auto& b_tensor_view = [&]() {
if constexpr(std::is_same_v<BLayout, tensor_layout::gemm::RowMajor>)
{
if constexpr(TilePartitioner::BlockGemmShape::PermuteB)
{
constexpr index_t K1 = GemmPipeline::GetSmemPackB();
const index_t K0 = splitk_batch_offset.splitted_k / K1;
constexpr index_t VectorSizeB = std::min(K1, GemmPipeline::GetVectorSizeB());
const auto b_k0_n_k1_desc =
make_naive_tensor_descriptor(make_tuple(K0, kargs.N, K1),
make_tuple(kargs.N * K1, K1, I1),
number<VectorSizeB>{},
number<1>{});
const auto b_n_k_desc = transform_tensor_descriptor(
b_k0_n_k1_desc,
make_tuple(make_merge_transform(make_tuple(K0, K1)),
make_pass_through_transform(kargs.N)),
make_tuple(sequence<0, 2>{}, sequence<1>{}),
make_tuple(sequence<0>{}, sequence<1>{}));
return make_tensor_view<address_space_enum::global>(b_ptr, b_n_k_desc);
}
else
{
return make_naive_tensor_view<address_space_enum::global>(
b_ptr,
make_tuple(splitk_batch_offset.splitted_k, kargs.N),
make_tuple(kargs.stride_B, 1),
number<GemmPipeline::GetVectorSizeB()>{},
number<1>{});
}
}
else
{
if constexpr(TilePartitioner::BlockGemmShape::PermuteB)
{
constexpr index_t K1 = GemmPipeline::GetSmemPackB();
const index_t K0 = splitk_batch_offset.splitted_k / K1;
constexpr index_t VectorSizeB = std::min(K1, GemmPipeline::GetVectorSizeB());
const auto b_k0_n_k1_desc =
make_naive_tensor_descriptor(make_tuple(K0, kargs.N, K1),
make_tuple(kargs.N * K1, K1, I1),
number<VectorSizeB>{},
number<1>{});
const auto b_n_k_desc = transform_tensor_descriptor(
b_k0_n_k1_desc,
make_tuple(make_merge_transform(make_tuple(K0, K1)),
make_pass_through_transform(kargs.N)),
make_tuple(sequence<0, 2>{}, sequence<1>{}),
make_tuple(sequence<1>{}, sequence<0>{}));
return make_tensor_view<address_space_enum::global>(b_ptr, b_n_k_desc);
}
else
{
return make_naive_tensor_view<address_space_enum::global>(
b_ptr,
make_tuple(kargs.N, splitk_batch_offset.splitted_k),
make_tuple(kargs.stride_B, 1),
number<GemmPipeline::GetVectorSizeB()>{},
number<1>{});
}
}
}();
// TODO: enable vector write for C in ColMajor
const auto& c_tensor_view = [&]() {
if constexpr(std::is_same_v<CLayout, tensor_layout::gemm::RowMajor>)
{
return make_naive_tensor_view<address_space_enum::global, DstInMemOp>(
c_ptr,
make_tuple(kargs.M, kargs.N),
make_tuple(kargs.stride_C, 1),
number<EpiloguePipeline::GetVectorSizeC()>{},
number<1>{});
}
else
{
return make_naive_tensor_view<address_space_enum::global, DstInMemOp>(
c_ptr,
make_tuple(kargs.M, kargs.N),
make_tuple(1, kargs.stride_C),
number<1>{},
number<1>{});
}
}();
return make_tuple(a_tensor_view, bq_tensor_view, b_tensor_view, c_tensor_view);
}
template <typename TensorView>
CK_TILE_DEVICE static auto MakeGemmPadViews(const TensorView& views)
{
const auto& a_pad_view = [&]() {
const auto& a_tensor_view = views.at(I0);
if constexpr(std::is_same_v<ALayout, tensor_layout::gemm::RowMajor>)
{
return pad_tensor_view(a_tensor_view,
make_tuple(number<TilePartitioner::MPerBlock>{},
number<TilePartitioner::KPerBlock>{}),
sequence<false, GemmPipeline::kPadK>{});
}
else
{
return pad_tensor_view(a_tensor_view,
make_tuple(number<TilePartitioner::KPerBlock>{},
number<TilePartitioner::MPerBlock>{}),
sequence<false, GemmPipeline::kPadM>{});
}
}();
const auto& bq_pad_view = [&]() {
const auto& bq_tensor_view = views.at(I1);
static_assert(std::is_same_v<BQLayout, tensor_layout::gemm::ColumnMajor>);
return pad_tensor_view(
bq_tensor_view,
make_tuple(number<TilePartitioner::NPerBlock>{},
number<TilePartitioner::KPerBlock / GemmPipeline::QuantGroupSize>{}),
// TODO: Add support for padding.
sequence<false, false>{});
}();
const auto& b_pad_view = [&]() {
const auto& b_tensor_view = views.at(I2);
if constexpr(std::is_same_v<BLayout, tensor_layout::gemm::ColumnMajor>)
{
return pad_tensor_view(b_tensor_view,
make_tuple(number<TilePartitioner::NPerBlock>{},
number<TilePartitioner::KPerBlock>{}),
sequence<false, GemmPipeline::kPadK>{});
}
else
{
return pad_tensor_view(b_tensor_view,
make_tuple(number<TilePartitioner::KPerBlock>{},
number<TilePartitioner::NPerBlock>{}),
sequence<false, GemmPipeline::kPadN>{});
}
}();
// TODO vector write in for C in ColMajor
const auto& c_pad_view = [&]() {
const auto& c_tensor_view = views.at(I3);
if constexpr(std::is_same_v<CLayout, tensor_layout::gemm::RowMajor>)
{
return pad_tensor_view(c_tensor_view,
make_tuple(number<TilePartitioner::MPerBlock>{},
number<TilePartitioner::NPerBlock>{}),
sequence<false, GemmPipeline::kPadN>{});
}
else
{
return pad_tensor_view(c_tensor_view,
make_tuple(number<TilePartitioner::MPerBlock>{},
number<TilePartitioner::NPerBlock>{}),
sequence<GemmPipeline::kPadM, false>{});
}
}();
return make_tuple(a_pad_view, bq_pad_view, b_pad_view, c_pad_view);
}
template <typename PadView>
CK_TILE_DEVICE static auto
MakeGemmTileWindows(const PadView& views, const index_t i_m, const index_t i_n)
{
const auto& a_pad_view = views.at(I0);
const auto& bq_pad_view = views.at(I1);
const auto& b_pad_view = views.at(I2);
const auto& c_pad_view = views.at(I3);
const auto& a_block_window = [&]() {
if constexpr(std::is_same_v<ALayout, tensor_layout::gemm::RowMajor>)
{
return make_tile_window(a_pad_view,
make_tuple(number<TilePartitioner::MPerBlock>{},
number<TilePartitioner::KPerBlock>{}),
{i_m, 0});
}
else
{
return make_tile_window(a_pad_view,
make_tuple(number<TilePartitioner::KPerBlock>{},
number<TilePartitioner::MPerBlock>{}),
{0, i_m});
}
}();
const auto& bq_block_window = [&]() {
static_assert(std::is_same_v<BQLayout, tensor_layout::gemm::ColumnMajor>);
return make_tile_window(
bq_pad_view,
make_tuple(number<TilePartitioner::NPerBlock>{},
number<TilePartitioner::KPerBlock / GemmPipeline::QuantGroupSize>{}),
{i_n, 0});
}();
const auto& b_block_window = [&]() {
if constexpr(std::is_same_v<BLayout, tensor_layout::gemm::ColumnMajor>)
{
return make_tile_window(b_pad_view,
make_tuple(number<TilePartitioner::NPerBlock>{},
number<TilePartitioner::KPerBlock>{}),
{i_n, 0});
}
else
{
return make_tile_window(b_pad_view,
make_tuple(number<TilePartitioner::KPerBlock>{},
number<TilePartitioner::NPerBlock>{}),
{0, i_n});
}
}();
auto c_block_window = make_tile_window(
c_pad_view,
make_tuple(number<TilePartitioner::MPerBlock>{}, number<TilePartitioner::NPerBlock>{}),
{i_m, i_n});
return make_tuple(a_block_window, bq_block_window, b_block_window, c_block_window);
}
/**
* @brief Runs single GEMM problem cooperatively by whole workgroup.
*
* @param a_ptr input A pointer
* @param b_ptr input B pointer
* @param bq_ptr input BQ pointer
* @param c_ptr output C pointer
* @param smem_ptr_0 The start memory pointer of the shared memory block.
* @param kargs GEMM kernel arguments
* @param splitk_batch_offset splitk_batch_offset Utility structure used to calculate k batch.
* @param block_idx_m The GEMM's output M dimension tile index processed by this workgroup.
* @param block_idx_n The GEMM's output N dimension tile index processed by this workgroup.
*
* @tparam DstInMemOp Destination memory operation (default: set).
*/
template <memory_operation_enum DstInMemOp = memory_operation_enum::set>
CK_TILE_DEVICE static void RunGemm(const ADataType* a_ptr,
const BDataType* b_ptr,
const BQDataType* bq_ptr,
CDataType* c_ptr,
void* smem_ptr_0,
const BQuantGemmKernelArgs& kargs,
const SplitKBatchOffset& splitk_batch_offset,
const index_t block_idx_m,
const index_t block_idx_n)
{
// Create Gemm tensor views, pad views and tile windows
const auto& gemm_tensor_views_tuple = MakeGemmTensorViews<DstInMemOp>(
a_ptr, b_ptr, bq_ptr, c_ptr, kargs, splitk_batch_offset);
const auto& gemm_pad_views = MakeGemmPadViews(gemm_tensor_views_tuple);
auto gemm_tile_windows = MakeGemmTileWindows(gemm_pad_views, block_idx_m, block_idx_n);
const index_t num_loop = __builtin_amdgcn_readfirstlane(
TilePartitioner::GetLoopNum(splitk_batch_offset.splitted_k));
// Run GEMM cooperatively by whole workgroup.
const auto& a_block_window = gemm_tile_windows.at(I0);
const auto& bq_block_window = gemm_tile_windows.at(I1);
const auto& b_block_window = gemm_tile_windows.at(I2);
const auto& c_block_tile = GemmPipeline{}.template operator()(
a_block_window, b_block_window, bq_block_window, num_loop, smem_ptr_0);
// Run Epilogue Pipeline
auto& c_block_window = gemm_tile_windows.at(I3);
EpiloguePipeline{}.template
operator()<decltype(c_block_window), decltype(c_block_tile), decltype(c_block_window)>(
c_block_window, c_block_tile, c_block_window, smem_ptr_0);
}
CK_TILE_DEVICE void operator()(BQuantGemmKernelArgs kargs) const
{
const auto blockId = __builtin_amdgcn_readfirstlane(blockIdx.x);
const auto [iM, iN] = TilePartitioner{kargs.M, kargs.N}.GetOutputTileIndex(blockId);
const index_t i_m = __builtin_amdgcn_readfirstlane(iM * TilePartitioner::MPerBlock);
const index_t i_n = __builtin_amdgcn_readfirstlane(iN * TilePartitioner::NPerBlock);
const SplitKBatchOffset splitk_batch_offset(kargs);
// options
const ADataType* a_ptr = static_cast<const ADataType*>(kargs.a_ptr);
const BDataType* b_ptr = static_cast<const BDataType*>(kargs.b_ptr);
const BQDataType* bq_ptr = static_cast<const BQDataType*>(kargs.bq_ptr);
CDataType* c_ptr = static_cast<CDataType*>(kargs.c_ptr);
// allocate LDS
__shared__ char smem_ptr_0[GetSmemSize()];
assert(kargs.k_batch == 1);
RunGemm(a_ptr, b_ptr, bq_ptr, c_ptr, smem_ptr_0, kargs, splitk_batch_offset, i_m, i_n);
}
};
} // namespace ck_tile

555
include/ck_tile/ops/gemm_group_quant/kernel/gemm_aquant_kernel.hpp → include/ck_tile/ops/gemm_group_quant/kernel/gemm_quant_kernel.hpp
View File

@@ -12,117 +12,218 @@
#include "ck_tile/core/numeric/integer.hpp"
#include "ck_tile/core/numeric/math.hpp"
#include "ck_tile/host/concat.hpp"
#include "ck_tile/ops/gemm_group_quant/pipeline/tile_gemm_quant_traits.hpp"
namespace ck_tile {
struct AQuantGemmProblem
namespace detail {
// Helper templates for safe type extraction
template <typename T, typename Default>
struct get_aq_layout_or
{
CK_TILE_HOST AQuantGemmProblem() = default;
CK_TILE_HOST AQuantGemmProblem(index_t M_,
index_t N_,
index_t K_,
index_t QK_,
index_t stride_A_,
index_t stride_B_,
index_t stride_C_,
index_t stride_AQ_)
using type = Default;
};
template <typename T, typename Default>
requires requires { typename T::AQLayout; }
struct get_aq_layout_or<T, Default>
{
using type = typename T::AQLayout;
};
template <typename T, typename Default>
struct get_bq_layout_or
{
using type = Default;
};
template <typename T, typename Default>
requires requires { typename T::BQLayout; }
struct get_bq_layout_or<T, Default>
{
using type = typename T::BQLayout;
};
template <typename T, typename Default>
struct get_aq_data_type_or
{
using type = Default;
};
template <typename T, typename Default>
requires requires { typename T::AQDataType; }
struct get_aq_data_type_or<T, Default>
{
using type = typename T::AQDataType;
};
template <typename T, typename Default>
struct get_bq_data_type_or
{
using type = Default;
};
template <typename T, typename Default>
requires requires { typename T::BQDataType; }
struct get_bq_data_type_or<T, Default>
{
using type = typename T::BQDataType;
};
template <typename T, typename Default>
struct get_preshuffle_or
{
using type = Default;
};
template <typename T, typename Default>
requires requires { typename T::PreshuffleQuant; }
struct get_preshuffle_or<T, Default>
{
using type = typename T::PreshuffleQuant;
};
} // namespace detail
struct QuantGemmProblem
{
CK_TILE_HOST QuantGemmProblem() = default;
CK_TILE_HOST QuantGemmProblem(index_t M_,
index_t N_,
index_t K_,
index_t QK_A_,
index_t QK_B_,
index_t stride_A_,
index_t stride_B_,
index_t stride_C_,
index_t stride_AQ_,
index_t stride_BQ_)
: M(M_),
N(N_),
K(K_),
QK(QK_),
QK_A(QK_A_),
QK_B(QK_B_),
stride_A(stride_A_),
stride_B(stride_B_),
stride_C(stride_C_),
stride_AQ(stride_AQ_)
stride_AQ(stride_AQ_),
stride_BQ(stride_BQ_)
{
}
index_t M;
index_t N;
index_t K;
index_t QK;
index_t QK_A;
index_t QK_B;
index_t stride_A;
index_t stride_B;
index_t stride_C;
index_t stride_AQ;
index_t stride_BQ;
};
struct AQuantGemmHostArgs : public AQuantGemmProblem
struct QuantGemmHostArgs : public QuantGemmProblem
{
CK_TILE_HOST AQuantGemmHostArgs() = default;
CK_TILE_HOST AQuantGemmHostArgs(const void* a_ptr_,
const void* b_ptr_,
void* c_ptr_,
const void* aq_ptr_,
index_t k_batch_,
index_t M_,
index_t N_,
index_t K_,
index_t QK_,
index_t stride_A_,
index_t stride_B_,
index_t stride_C_,
index_t stride_AQ_)
: AQuantGemmProblem(M_, N_, K_, QK_, stride_A_, stride_B_, stride_C_, stride_AQ_),
CK_TILE_HOST QuantGemmHostArgs() = default;
CK_TILE_HOST QuantGemmHostArgs(const void* a_ptr_,
const void* b_ptr_,
void* c_ptr_,
const void* aq_ptr_,
const void* bq_ptr_,
index_t k_batch_,
index_t M_,
index_t N_,
index_t K_,
index_t QK_A_,
index_t QK_B_,
index_t stride_A_,
index_t stride_B_,
index_t stride_C_,
index_t stride_AQ_,
index_t stride_BQ_)
: QuantGemmProblem(
M_, N_, K_, QK_A_, QK_B_, stride_A_, stride_B_, stride_C_, stride_AQ_, stride_BQ_),
a_ptr(a_ptr_),
b_ptr(b_ptr_),
aq_ptr(aq_ptr_),
bq_ptr(bq_ptr_),
c_ptr(c_ptr_),
k_batch(k_batch_)
{
}
const void* a_ptr;
const void* b_ptr;
const void* aq_ptr;
void* c_ptr;
index_t k_batch;
const void* a_ptr = nullptr;
const void* b_ptr = nullptr;
const void* aq_ptr = nullptr;
const void* bq_ptr = nullptr;
void* c_ptr = nullptr;
index_t k_batch = 0;
};
struct AQuantGemmKernelArgs
struct QuantGemmKernelArgs
{
const void* a_ptr;
const void* b_ptr;
const void* aq_ptr;
const void* bq_ptr;
void* c_ptr;
index_t M;
index_t N;
index_t K;
index_t QK;
index_t QK_A;
index_t QK_B;
index_t stride_A;
index_t stride_B;
index_t stride_C;
index_t stride_AQ;
index_t stride_BQ;
index_t k_batch;
};
template <typename TilePartitioner_, typename GemmPipeline_, typename EpiloguePipeline_>
struct AQuantGemmKernel
template <typename TilePartitioner_,
typename GemmPipeline_,
typename EpiloguePipeline_,
QuantType QuantType_>
struct QuantGemmKernel
{
using TilePartitioner = remove_cvref_t<TilePartitioner_>;
using GemmPipeline = remove_cvref_t<GemmPipeline_>;
using EpiloguePipeline = remove_cvref_t<EpiloguePipeline_>;
using ALayout = remove_cvref_t<typename GemmPipeline::ALayout>;
using AQLayout = remove_cvref_t<typename GemmPipeline::AQLayout>;
using BLayout = remove_cvref_t<typename GemmPipeline::BLayout>;
using CLayout = remove_cvref_t<typename GemmPipeline::CLayout>;
using TilePartitioner = remove_cvref_t<TilePartitioner_>;
using GemmPipeline = remove_cvref_t<GemmPipeline_>;
using EpiloguePipeline = remove_cvref_t<EpiloguePipeline_>;
using ALayout = remove_cvref_t<typename GemmPipeline::ALayout>;
using BLayout = remove_cvref_t<typename GemmPipeline::BLayout>;
using CLayout = remove_cvref_t<typename GemmPipeline::CLayout>;
using AQLayout = remove_cvref_t<
typename detail::get_aq_layout_or<GemmPipeline, typename GemmPipeline::ALayout>::type>;
using BQLayout = remove_cvref_t<
typename detail::get_bq_layout_or<GemmPipeline, typename GemmPipeline::BLayout>::type>;
static constexpr index_t kBlockSize = GemmPipeline::BlockSize;
static constexpr bool PreshuffleQuant = GemmPipeline::PreshuffleQuant;
static constexpr bool PreshuffleQuant = remove_cvref_t<
typename detail::get_preshuffle_or<GemmPipeline, std::false_type>::type>::value;
using ADataType = remove_cvref_t<typename GemmPipeline::ADataType>;
using AQDataType = remove_cvref_t<typename GemmPipeline::AQDataType>;
using BDataType = remove_cvref_t<typename GemmPipeline::BDataType>;
using CDataType = remove_cvref_t<typename EpiloguePipeline::ODataType>;
using ADataType = remove_cvref_t<typename GemmPipeline::ADataType>;
using BDataType = remove_cvref_t<typename GemmPipeline::BDataType>;
using CDataType = remove_cvref_t<typename EpiloguePipeline::ODataType>;
using AccDataType = remove_cvref_t<typename EpiloguePipeline::AccDataType>;
static constexpr auto I0 = number<0>();
static constexpr auto I1 = number<1>();
static constexpr auto I2 = number<2>();
static constexpr auto I3 = number<3>();
using AQDataType =
remove_cvref_t<typename detail::get_aq_data_type_or<GemmPipeline, AccDataType>::type>;
using BQDataType =
remove_cvref_t<typename detail::get_bq_data_type_or<GemmPipeline, AccDataType>::type>;
static constexpr auto I0 = number<0>(); // A Tensor
static constexpr auto I1 = number<1>(); // AQ Tensor
static constexpr auto I2 = number<2>(); // B Tensor
static constexpr auto I3 = number<3>(); // BQ Tensor
static constexpr auto I4 = number<4>(); // C Tensor
static constexpr auto kQuantType = QuantType_;
[[nodiscard]] CK_TILE_HOST static const std::string GetName()
{
// clang-format off
return concat('_', "gemm", gemm_prec_str<ADataType, BDataType>, GemmPipeline::GetName());
return concat('_', "gemm_quant", gemm_prec_str<ADataType, BDataType>, GemmPipeline::GetName());
// clang-format on
}
@@ -133,22 +234,25 @@ struct AQuantGemmKernel
CK_TILE_HOST static constexpr auto BlockSize() { return dim3(kBlockSize); }
CK_TILE_HOST static constexpr AQuantGemmKernelArgs
MakeKernelArgs(const AQuantGemmHostArgs& hostArgs)
CK_TILE_HOST static constexpr QuantGemmKernelArgs
MakeKernelArgs(const QuantGemmHostArgs& hostArgs)
{
return AQuantGemmKernelArgs{hostArgs.a_ptr,
hostArgs.b_ptr,
hostArgs.aq_ptr,
hostArgs.c_ptr,
hostArgs.M,
hostArgs.N,
hostArgs.K,
hostArgs.QK,
hostArgs.stride_A,
hostArgs.stride_B,
hostArgs.stride_C,
hostArgs.stride_AQ,
hostArgs.k_batch};
return QuantGemmKernelArgs{hostArgs.a_ptr,
hostArgs.b_ptr,
hostArgs.aq_ptr,
hostArgs.bq_ptr,
hostArgs.c_ptr,
hostArgs.M,
hostArgs.N,
hostArgs.K,
hostArgs.QK_A,
hostArgs.QK_B,
hostArgs.stride_A,
hostArgs.stride_B,
hostArgs.stride_C,
hostArgs.stride_AQ,
hostArgs.stride_BQ,
hostArgs.k_batch};
}
CK_TILE_HOST_DEVICE static constexpr index_t GetSmemSize()
@@ -158,7 +262,7 @@ struct AQuantGemmKernel
struct SplitKBatchOffset
{
__device__ SplitKBatchOffset(const AQuantGemmKernelArgs& kargs,
__device__ SplitKBatchOffset(const QuantGemmKernelArgs& kargs,
const std::size_t k_id = blockIdx.z)
{
constexpr auto K1 = TilePartitioner::BlockGemmShape::WarpTile::at(I2);
@@ -198,7 +302,7 @@ struct AQuantGemmKernel
index_t splitted_k;
};
CK_TILE_HOST static bool IsSupportedArgument(const AQuantGemmKernelArgs& kargs)
CK_TILE_HOST static bool IsSupportedArgument(const QuantGemmKernelArgs& kargs)
{
if(kargs.k_batch != 1)
{
@@ -209,14 +313,31 @@ struct AQuantGemmKernel
return false;
}
static_assert(std::is_same_v<AQLayout, tensor_layout::gemm::RowMajor>);
if(kargs.QK % GemmPipeline::GetVectorSizeAQ() != 0)
if constexpr(kQuantType == QuantType::AQuantGrouped)
{
if(ck_tile::EnvIsEnabled(CK_TILE_ENV(CK_TILE_LOGGING)))
static_assert(std::is_same_v<AQLayout, tensor_layout::gemm::RowMajor>);
if(kargs.QK_A % GemmPipeline::GetVectorSizeAQ() != 0)
{
CK_TILE_ERROR("K is not a multiple of vector load size for A tensor!");
if(ck_tile::EnvIsEnabled(CK_TILE_ENV(CK_TILE_LOGGING)))
{
CK_TILE_ERROR("K_A is not a multiple of vector load size for A tensor!");
}
return false;
}
}
// NOTE: no kernel currently uses BQuant like this:
if constexpr(kQuantType == QuantType::BQuantGrouped)
{
static_assert(std::is_same_v<BQLayout, tensor_layout::gemm::ColumnMajor>);
if(kargs.QK_B % GemmPipeline::GetVectorSizeBQ() != 0)
{
if(ck_tile::EnvIsEnabled(CK_TILE_ENV(CK_TILE_LOGGING)))
{
CK_TILE_ERROR("K_B is not a multiple of vector load size for B tensor!");
}
return false;
}
return false;
}
if constexpr(std::is_same_v<ALayout, tensor_layout::gemm::RowMajor>)
@@ -350,8 +471,9 @@ struct AQuantGemmKernel
CK_TILE_DEVICE static auto MakeGemmTensorViews(const ADataType* a_ptr,
const BDataType* b_ptr,
const AQDataType* aq_ptr,
const BQDataType* bq_ptr,
CDataType* c_ptr,
const AQuantGemmKernelArgs& kargs,
const QuantGemmKernelArgs& kargs,
const SplitKBatchOffset& splitk_batch_offset)
{
static_assert(!TilePartitioner::BlockGemmShape::PermuteA, "Not implemented!");
@@ -380,71 +502,85 @@ struct AQuantGemmKernel
return ck_tile::integer_least_multiple(length, alignment) - length;
};
const auto& make_preshuffled_aq_tensor_view = [&]() {
const auto aq_x = kargs.M * GemmPipeline::KPerBlockAQ;
const auto aq_y = kargs.QK / GemmPipeline::KPerBlockAQ;
const auto aq_desc =
make_naive_tensor_descriptor(make_tuple(aq_y, aq_x),
make_tuple(aq_x, 1),
number<GemmPipeline::GetVectorSizeAQ()>{},
number<1>{});
const auto block_tile_size = GemmPipeline::MPerBlock * GemmPipeline::KPerBlockAQ;
const auto aq_pad0_desc = transform_tensor_descriptor(
aq_desc,
make_tuple(make_pass_through_transform(aq_y),
make_right_pad_transform(aq_x, get_padding_size(aq_x, block_tile_size))),
make_tuple(sequence<0>{}, sequence<1>{}),
make_tuple(sequence<0>{}, sequence<1>{}));
const auto pad_aq_x = aq_pad0_desc.get_lengths()[I1];
const auto wave_tile_size =
TilePartitioner::BlockGemmShape::WarpTile::at(I0) * GemmPipeline::KPerBlockAQ;
const auto wave_tile_count_x = ck_tile::integer_divide_ceil(pad_aq_x, wave_tile_size);
const auto aq_unmerge_pad0_desc = transform_tensor_descriptor(
aq_pad0_desc,
make_tuple(make_pass_through_transform(aq_y),
make_unmerge_transform(make_tuple(wave_tile_count_x, wave_tile_size))),
make_tuple(sequence<0>{}, sequence<1>{}),
make_tuple(sequence<0>{}, sequence<1, 2>{}));
const auto aq_pad1_desc = transform_tensor_descriptor(
aq_unmerge_pad0_desc,
make_tuple(make_pass_through_transform(aq_y),
make_pass_through_transform(wave_tile_count_x),
make_right_pad_transform(
wave_tile_size, get_padding_size(wave_tile_size, get_warp_size()))),
make_tuple(sequence<0>{}, sequence<1>{}, sequence<2>{}),
make_tuple(sequence<0>{}, sequence<1>{}, sequence<2>{}));
const auto pad_wave_size =
ck_tile::integer_least_multiple(wave_tile_size, get_warp_size());
const auto aq_merge_pad1_desc = transform_tensor_descriptor(
aq_pad1_desc,
make_tuple(make_merge_transform(make_tuple(aq_y, wave_tile_count_x)),
make_pass_through_transform(pad_wave_size)),
make_tuple(sequence<0, 1>{}, sequence<2>{}),
make_tuple(sequence<0>{}, sequence<1>{}));
return make_tensor_view<address_space_enum::global>(aq_ptr, aq_merge_pad1_desc);
};
const auto& aq_tensor_view = [&]() {
static_assert(std::is_same_v<AQLayout, tensor_layout::gemm::RowMajor>);
if constexpr(PreshuffleQuant)
if constexpr(kQuantType == QuantType::AQuantGrouped && PreshuffleQuant)
{
return make_preshuffled_aq_tensor_view();
static_assert(std::is_same_v<AQLayout, tensor_layout::gemm::RowMajor>);
const auto aq_x = kargs.M * GemmPipeline::KPerBlockAQ;
const auto aq_y = kargs.QK_A / GemmPipeline::KPerBlockAQ;
const auto aq_desc =
make_naive_tensor_descriptor(make_tuple(aq_y, aq_x),
make_tuple(aq_x, 1),
number<GemmPipeline::GetVectorSizeAQ()>{},
number<1>{});
const auto block_tile_size = GemmPipeline::MPerBlock * GemmPipeline::KPerBlockAQ;
const auto aq_pad0_desc = transform_tensor_descriptor(
aq_desc,
make_tuple(
make_pass_through_transform(aq_y),
make_right_pad_transform(aq_x, get_padding_size(aq_x, block_tile_size))),
make_tuple(sequence<0>{}, sequence<1>{}),
make_tuple(sequence<0>{}, sequence<1>{}));
const auto pad_aq_x = aq_pad0_desc.get_lengths()[I1];
const auto wave_tile_size =
TilePartitioner::BlockGemmShape::WarpTile::at(I0) * GemmPipeline::KPerBlockAQ;
const auto wave_tile_count_x =
ck_tile::integer_divide_ceil(pad_aq_x, wave_tile_size);
const auto aq_unmerge_pad0_desc = transform_tensor_descriptor(
aq_pad0_desc,
make_tuple(
make_pass_through_transform(aq_y),
make_unmerge_transform(make_tuple(wave_tile_count_x, wave_tile_size))),
make_tuple(sequence<0>{}, sequence<1>{}),
make_tuple(sequence<0>{}, sequence<1, 2>{}));
const auto aq_pad1_desc = transform_tensor_descriptor(
aq_unmerge_pad0_desc,
make_tuple(
make_pass_through_transform(aq_y),
make_pass_through_transform(wave_tile_count_x),
make_right_pad_transform(
wave_tile_size, get_padding_size(wave_tile_size, get_warp_size()))),
make_tuple(sequence<0>{}, sequence<1>{}, sequence<2>{}),
make_tuple(sequence<0>{}, sequence<1>{}, sequence<2>{}));
const auto pad_wave_size =
ck_tile::integer_least_multiple(wave_tile_size, get_warp_size());
const auto aq_merge_pad1_desc = transform_tensor_descriptor(
aq_pad1_desc,
make_tuple(make_merge_transform(make_tuple(aq_y, wave_tile_count_x)),
make_pass_through_transform(pad_wave_size)),
make_tuple(sequence<0, 1>{}, sequence<2>{}),
make_tuple(sequence<0>{}, sequence<1>{}));
return make_tensor_view<address_space_enum::global>(aq_ptr, aq_merge_pad1_desc);
}
else
else if constexpr(kQuantType == QuantType::AQuantGrouped && !PreshuffleQuant)
{
static_assert(std::is_same_v<AQLayout, tensor_layout::gemm::RowMajor>);
return make_naive_tensor_view<address_space_enum::global>(
aq_ptr,
make_tuple(kargs.M, kargs.QK),
make_tuple(kargs.M, kargs.QK_A),
make_tuple(kargs.stride_AQ, 1),
number<GemmPipeline::GetVectorSizeAQ()>{},
number<1>{});
}
else if constexpr(kQuantType == QuantType::RowColQuant)
{
return make_naive_tensor_view<address_space_enum::global>(
aq_ptr,
make_tuple(kargs.M, kargs.N),
make_tuple(1, 0), // broadcasting over n
number<1>{},
number<1>{});
}
else
{
return nullptr; // TODO: use some other "empty" type for this
}
}();
const auto& b_tensor_view = [&]() {
@@ -510,6 +646,32 @@ struct AQuantGemmKernel
}
}();
const auto& bq_tensor_view = [&]() {
if constexpr(kQuantType == QuantType::RowColQuant)
{
return make_naive_tensor_view<address_space_enum::global>(
bq_ptr,
make_tuple(kargs.M, kargs.N),
make_tuple(0, 1), // broadcasting over m
number<1>{},
number<1>{});
}
else if constexpr(kQuantType == QuantType::BQuantGrouped)
{
static_assert(std::is_same_v<BQLayout, tensor_layout::gemm::ColumnMajor>);
return make_naive_tensor_view<address_space_enum::global>(
bq_ptr,
make_tuple(kargs.N, kargs.QK_B),
make_tuple(kargs.stride_BQ, 1),
number<GemmPipeline::GetVectorSizeBQ()>{},
number<1>{});
}
else
{
return nullptr; // TODO: use some other "empty" type for this
}
}();
// TODO: enable vector write for C in ColMajor
const auto& c_tensor_view = [&]() {
if constexpr(std::is_same_v<CLayout, tensor_layout::gemm::RowMajor>)
@@ -532,7 +694,8 @@ struct AQuantGemmKernel
}
}();
return make_tuple(a_tensor_view, aq_tensor_view, b_tensor_view, c_tensor_view);
return make_tuple(
a_tensor_view, aq_tensor_view, b_tensor_view, bq_tensor_view, c_tensor_view);
}
template <typename TensorView>
@@ -556,6 +719,7 @@ struct AQuantGemmKernel
}
}();
// no padding
const auto& aq_pad_view = [&]() { return views.at(I1); }();
const auto& b_pad_view = [&]() {
@@ -576,9 +740,12 @@ struct AQuantGemmKernel
}
}();
// no padding
const auto& bq_pad_view = [&]() { return views.at(I3); }();
// TODO vector write in for C in ColMajor
const auto& c_pad_view = [&]() {
const auto& c_tensor_view = views.at(I3);
const auto& c_tensor_view = views.at(I4);
if constexpr(std::is_same_v<CLayout, tensor_layout::gemm::RowMajor>)
{
return pad_tensor_view(c_tensor_view,
@@ -595,7 +762,7 @@ struct AQuantGemmKernel
}
}();
return make_tuple(a_pad_view, aq_pad_view, b_pad_view, c_pad_view);
return make_tuple(a_pad_view, aq_pad_view, b_pad_view, bq_pad_view, c_pad_view);
}
template <typename PadView>
@@ -605,7 +772,8 @@ struct AQuantGemmKernel
const auto& a_pad_view = views.at(I0);
const auto& aq_pad_view = views.at(I1);
const auto& b_pad_view = views.at(I2);
const auto& c_pad_view = views.at(I3);
const auto& bq_pad_view = views.at(I3);
const auto& c_pad_view = views.at(I4);
const auto& a_block_window = [&]() {
if constexpr(std::is_same_v<ALayout, tensor_layout::gemm::RowMajor>)
@@ -625,14 +793,13 @@ struct AQuantGemmKernel
}();
const auto& aq_block_window = [&]() {
static_assert(std::is_same_v<AQLayout, tensor_layout::gemm::RowMajor>);
constexpr auto block_m = TilePartitioner::MPerBlock;
constexpr auto block_k = TilePartitioner::KPerBlock;
constexpr auto warp_m = TilePartitioner::BlockGemmShape::WarpTile::at(I0);
constexpr auto aqk_per_block =
TilePartitioner::KPerBlock / GemmPipeline::QuantGroupSize;
if constexpr(PreshuffleQuant)
if constexpr(kQuantType == QuantType::AQuantGrouped && PreshuffleQuant)
{
static_assert(std::is_same_v<AQLayout, tensor_layout::gemm::RowMajor>);
constexpr auto block_m = TilePartitioner::MPerBlock;
constexpr auto warp_m = TilePartitioner::BlockGemmShape::WarpTile::at(I0);
constexpr auto aqk_per_block =
TilePartitioner::KPerBlock / GemmPipeline::QuantGroupSize;
constexpr auto tile_window_width =
ck_tile::integer_least_multiple(warp_m * aqk_per_block, get_warp_size());
constexpr auto tile_window_height = block_m / warp_m;
@@ -642,13 +809,27 @@ struct AQuantGemmKernel
make_tuple(number<tile_window_height>{}, number<tile_window_width>{}),
{block_m_idx * tile_window_height, 0});
}
else
else if constexpr(kQuantType == QuantType::AQuantGrouped && !PreshuffleQuant)
{
static_assert(std::is_same_v<AQLayout, tensor_layout::gemm::RowMajor>);
constexpr auto block_m = TilePartitioner::MPerBlock;
constexpr auto block_k = TilePartitioner::KPerBlock;
return make_tile_window(
aq_pad_view,
make_tuple(number<block_m>{}, number<block_k / GemmPipeline::QuantGroupSize>{}),
{i_m, 0});
}
else if constexpr(kQuantType == QuantType::RowColQuant)
{
return make_tile_window(aq_pad_view,
make_tuple(number<TilePartitioner::MPerBlock>{},
number<TilePartitioner::NPerBlock>{}),
{i_m, i_n});
}
else
{
return nullptr; // TODO: use some other "empty" type?
}
}();
const auto& b_block_window = [&]() {
@@ -668,12 +849,36 @@ struct AQuantGemmKernel
}
}();
const auto& bq_block_window = [&]() {
if constexpr(kQuantType == QuantType::RowColQuant)
{
return make_tile_window(bq_pad_view,
make_tuple(number<TilePartitioner::MPerBlock>{},
number<TilePartitioner::NPerBlock>{}),
{i_m, i_n});
}
else if constexpr(kQuantType == QuantType::BQuantGrouped)
{
static_assert(std::is_same_v<BQLayout, tensor_layout::gemm::ColumnMajor>);
return make_tile_window(
bq_pad_view,
make_tuple(number<TilePartitioner::NPerBlock>{},
number<TilePartitioner::KPerBlock / GemmPipeline::QuantGroupSize>{}),
{i_n, 0});
}
else
{
return nullptr; // TODO: use some other "empty" type here
}
}();
auto c_block_window = make_tile_window(
c_pad_view,
make_tuple(number<TilePartitioner::MPerBlock>{}, number<TilePartitioner::NPerBlock>{}),
{i_m, i_n});
return make_tuple(a_block_window, aq_block_window, b_block_window, c_block_window);
return make_tuple(
a_block_window, aq_block_window, b_block_window, bq_block_window, c_block_window);
}
/**
@@ -695,16 +900,17 @@ struct AQuantGemmKernel
CK_TILE_DEVICE static void RunGemm(const ADataType* a_ptr,
const BDataType* b_ptr,
const AQDataType* aq_ptr,
const BQDataType* bq_ptr,
CDataType* c_ptr,
void* smem_ptr_0,
const AQuantGemmKernelArgs& kargs,
const QuantGemmKernelArgs& kargs,
const SplitKBatchOffset& splitk_batch_offset,
const index_t block_idx_m,
const index_t block_idx_n)
{
// Create Gemm tensor views, pad views and tile windows
const auto& gemm_tensor_views_tuple = MakeGemmTensorViews<DstInMemOp>(
a_ptr, b_ptr, aq_ptr, c_ptr, kargs, splitk_batch_offset);
a_ptr, b_ptr, aq_ptr, bq_ptr, c_ptr, kargs, splitk_batch_offset);
const auto& gemm_pad_views = MakeGemmPadViews(gemm_tensor_views_tuple);
auto gemm_tile_windows = MakeGemmTileWindows(gemm_pad_views, block_idx_m, block_idx_n);
@@ -713,22 +919,51 @@ struct AQuantGemmKernel
TilePartitioner::GetLoopNum(splitk_batch_offset.splitted_k));
// Run GEMM cooperatively by whole workgroup.
const auto& a_block_window = gemm_tile_windows.at(I0);
const auto& aq_block_window = gemm_tile_windows.at(I1);
const auto& b_block_window = gemm_tile_windows.at(I2);
const auto& a_block_window = gemm_tile_windows.at(I0);
const auto& b_block_window = gemm_tile_windows.at(I2);
const auto& c_block_tile = GemmPipeline{}.template operator()(
a_block_window, b_block_window, aq_block_window, kargs.M, num_loop, smem_ptr_0);
const auto& c_block_tile = [&]() {
if constexpr(kQuantType == QuantType::AQuantGrouped)
{
const auto& aq_block_window = gemm_tile_windows.at(I1);
return GemmPipeline{}.template operator()(
a_block_window, b_block_window, aq_block_window, kargs.M, num_loop, smem_ptr_0);
}
else if constexpr(kQuantType == QuantType::BQuantGrouped)
{
const auto& bq_block_window = gemm_tile_windows.at(I3);
return GemmPipeline{}.template operator()(
a_block_window, b_block_window, bq_block_window, num_loop, smem_ptr_0);
}
else if constexpr(kQuantType == QuantType::RowColQuant)
{
return GemmPipeline{}.template operator()(
a_block_window, b_block_window, num_loop, smem_ptr_0);
}
}();
// Run Epilogue Pipeline
auto& c_block_window = gemm_tile_windows.at(I3);
auto& c_block_window = gemm_tile_windows.at(I4);
EpiloguePipeline{}.template
operator()<decltype(c_block_window), decltype(c_block_tile), decltype(c_block_window)>(
c_block_window, c_block_tile, c_block_window, smem_ptr_0);
if constexpr(kQuantType == QuantType::AQuantGrouped ||
kQuantType == QuantType::BQuantGrouped)
{
EpiloguePipeline{}(c_block_window, c_block_tile, c_block_window, smem_ptr_0);
}
else if constexpr(kQuantType == QuantType::RowColQuant)
{
const auto& aq_block_window = gemm_tile_windows.at(I1);
const auto& bq_block_window = gemm_tile_windows.at(I3);
EpiloguePipeline{}(c_block_window,
c_block_tile,
c_block_window,
smem_ptr_0,
aq_block_window,
bq_block_window);
}
}
CK_TILE_DEVICE void operator()(AQuantGemmKernelArgs kargs) const
CK_TILE_DEVICE void operator()(QuantGemmKernelArgs kargs) const
{
const auto blockId = __builtin_amdgcn_readfirstlane(blockIdx.x);
const auto [iM, iN] = TilePartitioner{kargs.M, kargs.N}.GetOutputTileIndex(blockId);
@@ -740,13 +975,15 @@ struct AQuantGemmKernel
const ADataType* a_ptr = static_cast<const ADataType*>(kargs.a_ptr);
const BDataType* b_ptr = static_cast<const BDataType*>(kargs.b_ptr);
const AQDataType* aq_ptr = static_cast<const AQDataType*>(kargs.aq_ptr);
const BQDataType* bq_ptr = static_cast<const BQDataType*>(kargs.bq_ptr);
CDataType* c_ptr = static_cast<CDataType*>(kargs.c_ptr);
// allocate LDS
__shared__ char smem_ptr_0[GetSmemSize()];
assert(kargs.k_batch == 1);
RunGemm(a_ptr, b_ptr, aq_ptr, c_ptr, smem_ptr_0, kargs, splitk_batch_offset, i_m, i_n);
RunGemm(
a_ptr, b_ptr, aq_ptr, bq_ptr, c_ptr, smem_ptr_0, kargs, splitk_batch_offset, i_m, i_n);
}
};

161
include/ck_tile/ops/gemm_group_quant/pipeline/gemm_quant_pipeline_problem.hpp
View File

@@ -14,6 +14,7 @@ namespace ck_tile {
template <typename ADataType_,
typename AQDataType_,
typename BDataType_,
typename BQDataType_,
typename CDataType_,
typename BlockGemmShape_,
typename Traits_,
@@ -23,12 +24,12 @@ template <typename ADataType_,
GemmPipelineScheduler Scheduler_ = GemmPipelineScheduler::Intrawave,
bool HasHotLoop_ = true,
TailNumber TailNum_ = TailNumber::Full>
struct GemmAQuantPipelineProblemBase : public GemmPipelineProblemBase<ADataType_,
BDataType_,
CDataType_,
BlockGemmShape_,
Traits_,
ComputeDataType_>
struct GemmQuantPipelineProblemBase : public GemmPipelineProblemBase<ADataType_,
BDataType_,
CDataType_,
BlockGemmShape_,
Traits_,
ComputeDataType_>
{
using Base = GemmPipelineProblemBase<ADataType_,
BDataType_,
@@ -44,6 +45,7 @@ struct GemmAQuantPipelineProblemBase : public GemmPipelineProblemBase<ADataType_
using typename Base::CDataType;
using typename Base::ComputeDataType;
using AQDataType = remove_cvref_t<AQDataType_>;
using BQDataType = remove_cvref_t<BQDataType_>;
using BlockGemmShape = typename Base::BlockGemmShape;
@@ -63,6 +65,7 @@ struct GemmAQuantPipelineProblemBase : public GemmPipelineProblemBase<ADataType_
using Base::VectorLoadSize;
using AQLayout = remove_cvref_t<typename Traits::AQLayout>;
using BQLayout = remove_cvref_t<typename Traits::BQLayout>;
static constexpr uint32_t kQuantGroupSize = QuantGroupSize_;
static constexpr auto Scheduler = Scheduler_;
@@ -75,7 +78,7 @@ struct GemmAQuantPipelineProblemBase : public GemmPipelineProblemBase<ADataType_
[[nodiscard]] CK_TILE_HOST static const std::string GetName()
{
// clang-format off
return concat('_', "gemm_aquant_problem",
return concat('_', "gemm_quant_problem",
concat('x', VectorLoadSize, kBlockSize),
concat('x', kPadM, kPadN, kPadK),
Scheduler,
@@ -94,6 +97,13 @@ struct GemmAQuantPipelineProblemBase : public GemmPipelineProblemBase<ADataType_
static_assert(std::is_same_v<ALayout, tensor_layout::gemm::RowMajor>);
return kPadK ? 1 : GetAlignmentAQ();
}();
CK_TILE_HOST_DEVICE static constexpr auto GetAlignmentBQ()
{
return VectorLoadSize / sizeof(BQDataType);
}
static constexpr index_t VectorSizeBQ = []() { return kPadK ? 1 : GetAlignmentBQ(); }();
};
template <typename ADataType_,
@@ -108,18 +118,19 @@ template <typename ADataType_,
GemmPipelineScheduler Scheduler_ = GemmPipelineScheduler::Intrawave,
bool HasHotLoop_ = true,
TailNumber TailNum_ = TailNumber::Full>
using GemmAQuantPipelineProblem = GemmAQuantPipelineProblemBase<ADataType_,
AQDataType_,
BDataType_,
CDataType_,
BlockGemmShape_,
Traits_,
QuantGroupSize_,
TransposeC_,
ComputeDataType_,
Scheduler_,
HasHotLoop_,
TailNum_>;
using GemmAQuantPipelineProblem = GemmQuantPipelineProblemBase<ADataType_,
AQDataType_,
BDataType_,
void, // no BQDataType for AQuant
CDataType_,
BlockGemmShape_,
Traits_,
QuantGroupSize_,
TransposeC_,
ComputeDataType_,
Scheduler_,
HasHotLoop_,
TailNum_>;
template <typename ADataType_,
typename BDataType_,
@@ -132,96 +143,42 @@ template <typename ADataType_,
GemmPipelineScheduler Scheduler_ = GemmPipelineScheduler::Intrawave,
bool HasHotLoop_ = true,
TailNumber TailNum_ = TailNumber::Full>
struct GemmBQuantPipelineProblemBase : public GemmPipelineProblemBase<ADataType_,
BDataType_,
CDataType_,
BlockGemmShape_,
Traits_,
ComputeDataType_>
{
using Base = GemmPipelineProblemBase<ADataType_,
BDataType_,
CDataType_,
BlockGemmShape_,
Traits_,
ComputeDataType_>;
using Traits = typename Base::Traits;
using typename Base::ADataType;
using typename Base::BDataType;
using typename Base::CDataType;
using typename Base::ComputeDataType;
using BQDataType = remove_cvref_t<BQDataType_>;
using BlockGemmShape = typename Base::BlockGemmShape;
using typename Base::ALayout;
using typename Base::BLayout;
using typename Base::CLayout;
static constexpr bool TransposeC = Traits::TransposeC;
using Base::kBlockSize;
using Base::kPadK;
using Base::kPadM;
using Base::kPadN;
using Base::DoubleSmemBuffer;
using Base::VectorLoadSize;
using BQLayout = remove_cvref_t<typename Traits::BQLayout>;
static constexpr uint32_t kQuantGroupSize = QuantGroupSize_;
static constexpr auto Scheduler = Scheduler_;
static constexpr auto HasHotLoop = HasHotLoop_;
static constexpr auto TailNum = TailNum_;
static_assert(BlockGemmShape::kK % kQuantGroupSize == 0);
static_assert(Scheduler == GemmPipelineScheduler::Intrawave);
[[nodiscard]] CK_TILE_HOST static const std::string GetName()
{
// clang-format off
return concat('_', "gemm_bquant_problem",
concat('x', VectorLoadSize, kBlockSize),
concat('x', kPadM, kPadN, kPadK),
Scheduler,
"QuantGroupSize",
kQuantGroupSize);
// clang-format on
}
CK_TILE_HOST_DEVICE static constexpr auto GetAlignmentBQ()
{
return VectorLoadSize / sizeof(BQDataType);
}
static constexpr index_t VectorSizeBQ = []() { return kPadK ? 1 : GetAlignmentBQ(); }();
};
using GemmBQuantPipelineProblem = GemmQuantPipelineProblemBase<ADataType_,
void, // no AQDataType for BQuant
BDataType_,
BQDataType_,
CDataType_,
BlockGemmShape_,
Traits_,
QuantGroupSize_,
false, // no TransposeC
ComputeDataType_,
Scheduler_,
HasHotLoop_,
TailNum_>;
template <typename ADataType_,
typename BDataType_,
typename BQDataType_,
typename CDataType_,
typename AccDataType_,
typename BlockGemmShape_,
typename Traits_,
uint32_t QuantGroupSize_,
typename ComputeDataType_ = ADataType_,
bool TransposeC_ = false,
typename ComputeDataType_ = BDataType_,
GemmPipelineScheduler Scheduler_ = GemmPipelineScheduler::Intrawave,
bool HasHotLoop_ = true,
TailNumber TailNum_ = TailNumber::Full>
using GemmBQuantPipelineProblem = GemmBQuantPipelineProblemBase<ADataType_,
BDataType_,
BQDataType_,
CDataType_,
BlockGemmShape_,
Traits_,
QuantGroupSize_,
ComputeDataType_,
Scheduler_,
HasHotLoop_,
TailNum_>;
using GemmRowColQuantPipelineProblem = GemmQuantPipelineProblemBase<ADataType_,
AccDataType_,
BDataType_,
AccDataType_,
CDataType_,
BlockGemmShape_,
Traits_,
1, // no group size applicable
TransposeC_,
ComputeDataType_,
Scheduler_,
HasHotLoop_,
TailNum_>;
} // namespace ck_tile

44
include/ck_tile/ops/gemm_group_quant/pipeline/tile_gemm_quant_traits.hpp
View File

@@ -4,9 +4,17 @@
#pragma once
#include "ck_tile/core.hpp"
#include <cstdint>
namespace ck_tile {
enum struct QuantType : std::uint16_t
{
AQuantGrouped = 0,
BQuantGrouped = 1,
RowColQuant = 2
};
template <bool kPadM_,
bool kPadN_,
bool kPadK_,
@@ -14,19 +22,24 @@ template <bool kPadM_,
typename ALayout_,
typename BLayout_,
typename CLayout_,
typename AQLayout_ = ALayout_>
struct TileGemmAQuantTraits
QuantType QuantType_,
typename AQLayout_ = ALayout_,
typename BQLayout_ = BLayout_>
struct TileGemmQuantTraits
{
static constexpr bool kPadM = kPadM_;
static constexpr bool kPadN = kPadN_;
static constexpr bool kPadK = kPadK_;
static constexpr QuantType kQuantType = QuantType_;
static constexpr int _VectorSize = 16;
using ALayout = ALayout_;
using BLayout = BLayout_;
using CLayout = CLayout_;
using AQLayout = AQLayout_;
using BQLayout = BQLayout_;
static constexpr bool TransposeC = false;
static constexpr bool UseStructuredSparsity = false;
@@ -35,31 +48,4 @@ struct TileGemmAQuantTraits
static constexpr bool PreshuffleQuant = PreshuffleQuant_;
};
template <bool kPadM_,
bool kPadN_,
bool kPadK_,
bool PreshuffleQuant_,
typename ALayout_,
typename BLayout_,
typename CLayout_,
typename BQLayout_ = BLayout_>
struct TileGemmBQuantTraits
{
static constexpr bool kPadM = kPadM_;
static constexpr bool kPadN = kPadN_;
static constexpr bool kPadK = kPadK_;
static constexpr int _VectorSize = 16;
using ALayout = ALayout_;
using BLayout = BLayout_;
using CLayout = CLayout_;
using BQLayout = BQLayout_;
static constexpr bool TransposeC = false;
static constexpr bool UseStructuredSparsity = false;
static constexpr index_t NumWaveGroups = 1;
static constexpr bool PreshuffleQuant = PreshuffleQuant_;
};
} // namespace ck_tile