/home/docs/checkouts/readthedocs.org/user_builds/advanced-micro-devices-composable-kernel/checkouts/develop/include/ck_tile/ops/layernorm2d/kernel/layernorm2d_fwd_kernel.hpp Source File

/home/docs/checkouts/readthedocs.org/user_builds/advanced-micro-devices-composable-kernel/checkouts/develop/include/ck_tile/ops/layernorm2d/kernel/layernorm2d_fwd_kernel.hpp Source File#

Composable Kernel: /home/docs/checkouts/readthedocs.org/user_builds/advanced-micro-devices-composable-kernel/checkouts/develop/include/ck_tile/ops/layernorm2d/kernel/layernorm2d_fwd_kernel.hpp Source File
Go to the documentation of this file.
 // SPDX-License-Identifier: MIT
 // Copyright (c) 2018-2025, Advanced Micro Devices, Inc. All rights reserved.
  
 #pragma once
  
 #include "ck_tile/core.hpp"
 #include "ck_tile/ops/common.hpp"
 #include "ck_tile/ops/layernorm2d/pipeline/layernorm2d_fwd_traits.hpp"
  
 namespace ck_tile {
  
 // host side args
 struct Layernorm2dFwdHostArgs
 {
     const void* p_x;          // [m ,n], input, fp16/bf16
     const void* p_x_residual; // [m ,n], shortcut input, prec same as input, nullptr if not used
     const void* p_sm_scale;   // [1 ,n], smooth scale input, fp32, nullptr if not used
     const void* p_x_bias;     // [1, n], bias, prec same as input
     const void* p_gamma;      // [1, n], gamma, prec same as input
     const void* p_beta;       // [1, n], beta, prec same as input
  
     void* p_y;          // [m, n], output, fp16/bf16
     void* p_y_residual; // [m, n], shortcut output, prec same as input, nullptr if not used
     void* p_y_scale;    // [m, 1], output a dynamic quant per row, nullptr if not used
     void* p_mean;       // [m, 1], output mean, prec same as input, nullptr if not used
     void* p_invStd;     // [m, 1], output inv-stdvariance, prec same as input, nullptr if not used
  
     float epsilon;
  
     index_t m;
     index_t n;
     index_t x_stride;  // x row_stride
     index_t xr_stride; // x residule row stride
     index_t y_stride;  // y row stride
     index_t yr_stride; // y residule row stride
 };
  
 // TODO: Extract some type to wrapper class
 template <typename Pipeline_, typename Epilogue_>
 struct Layernorm2dFwd
 {
     using Pipeline = remove_cvref_t<Pipeline_>;
     using Epilogue = remove_cvref_t<Epilogue_>;
     using Problem  = typename Pipeline::Problem;
  
     using XDataType           = remove_cvref_t<typename Problem::XDataType>;
     using XBiasDataType       = remove_cvref_t<typename Problem::XBiasDataType>;
     using GammaDataType       = remove_cvref_t<typename Problem::GammaDataType>;
     using BetaDataType        = remove_cvref_t<typename Problem::BetaDataType>;
     using ComputeDataType     = remove_cvref_t<typename Problem::ComputeDataType>;
     using YDataType           = remove_cvref_t<typename Problem::YDataType>;
     using MeanDataType        = remove_cvref_t<typename Problem::MeanDataType>;
     using InvStdDataType      = remove_cvref_t<typename Problem::InvStdDataType>;
     using SmoothScaleDataType = remove_cvref_t<typename Problem::SmoothScaleDataType>;
     using YScaleDataType      = remove_cvref_t<typename Problem::YScaleDataType>;
  
     // for simplicity, shortcut input/output type is same as X
     using XResidualDataType = XDataType;
     using YResidualDataType = XDataType;
  
     static constexpr bool kHasGamma       = !std::is_same_v<GammaDataType, null_type>;
     static constexpr bool kHasBeta        = !std::is_same_v<BetaDataType, null_type>;
     static constexpr bool kSaveMeanInvStd = Problem::Traits::kSaveMeanInvStd;
     static constexpr bool kSaveMean       = Problem::Traits::kSaveMeanInvStd;
     static constexpr bool kSaveInvStd     = Problem::Traits::kSaveMeanInvStd;
  
     static constexpr index_t Block_M  = Problem::BlockShape::Block_M;
     static constexpr index_t Block_N  = Problem::BlockShape::Block_N;
     static constexpr bool kPadM       = false; // always no need to pad along M
     static constexpr bool kPadN       = Problem::Traits::kPadN;
     static constexpr bool kTwoPass    = Problem::Traits::kTwoPass;
     static constexpr auto kXbias      = Problem::Traits::kXbias;
     static constexpr auto kFusedAdd   = Problem::Traits::kFusedAdd;
     static constexpr auto kFusedQuant = Problem::Traits::kFusedQuant;
  
     static constexpr index_t ThreadPerWarp_N = Problem::BlockShape::ThreadPerWarp_N;
     static constexpr index_t Vector_N        = Problem::BlockShape::Vector_N;
     static constexpr index_t Repeat_N        = Problem::BlockShape::Repeat_N;
     static constexpr index_t kBlockSize      = Problem::BlockShape::BlockSize;
     static constexpr auto I0                 = number<0>{};
     static constexpr auto I1                 = number<1>{};
  
     struct Kargs
     {
         const void* p_x;          // [m ,n], input, fp16/bf16
         const void* p_x_residual; // [m ,n], shortcut input, prec same as input, nullptr if not used
         const void* p_sm_scale;   // [1 ,n], smooth scale input, fp32, nullptr if not used
         const void* p_x_bias;     // [1, n], bias, prec same as input
         const void* p_gamma;      // [1, n], gamma, prec same as input
         const void* p_beta;       // [1, n], beta, prec same as input
  
         void* p_y;          // [m, n], output, fp16/bf16
         void* p_y_residual; // [m, n], shortcut output, prec same as input, nullptr if not used
         void* p_y_scale;    // [m, 1], output a dynamic quant per row, nullptr if not used
  
         void* p_mean;   // [m, 1], output mean, prec same as input, nullptr if not used
         void* p_invStd; // [m, 1], output inv-stdvariance, prec same as input, nullptr if not used
  
         float epsilon;
  
         index_t m;
         index_t n;
         index_t x_stride;  // x row_stride
         index_t xr_stride; // x residule row stride
         index_t y_stride;  // y row stride
         index_t yr_stride; // y residule row stride
     };
     using Hargs = Layernorm2dFwdHostArgs;
  
     CK_TILE_HOST static constexpr Kargs MakeKargs(const Hargs& hargs)
     {
         return Kargs{hargs.p_x,
                      hargs.p_x_residual,
                      hargs.p_sm_scale,
                      hargs.p_x_bias,
                      hargs.p_gamma,
                      hargs.p_beta,
                      hargs.p_y,
                      hargs.p_y_residual,
                      hargs.p_y_scale,
                      hargs.p_mean,
                      hargs.p_invStd,
                      hargs.epsilon,
                      hargs.m,
                      hargs.n,
                      hargs.x_stride,
                      hargs.xr_stride,
                      hargs.y_stride,
                      hargs.yr_stride};
     }
  
     CK_TILE_HOST static constexpr auto GridSize(const Hargs& hargs)
     {
         return dim3(integer_divide_ceil(hargs.m, Block_M));
     }
  
     CK_TILE_HOST static constexpr auto BlockSize() { return Problem::BlockShape::BlockSize; }
  
     // clang-format off
     template <typename T> struct t2s;
     template <> struct t2s<float> { static constexpr const char * name = "fp32"; };
     template <> struct t2s<ck_tile::fp16_t> { static constexpr const char * name = "fp16"; };
     template <> struct t2s<ck_tile::bf16_t> { static constexpr const char * name = "bf16"; };
     template <> struct t2s<ck_tile::fp8_t> { static constexpr const char * name = "fp8"; };
     template <> struct t2s<ck_tile::bf8_t> { static constexpr const char * name = "bf8"; };
     template <> struct t2s<ck_tile::int8_t> { static constexpr const char * name = "int8"; };
     // clang-format on
  
     // in byte
     CK_TILE_HOST_DEVICE static constexpr index_t GetSmemSize() { return Pipeline::GetSmemSize(); }
  
     CK_TILE_HOST static std::string GetName()
     {
 #define _SS_ std::string
 #define _TS_ std::to_string
         // clang-format off
         using S_ = typename Problem::BlockShape;
         auto surfix = [&] () {
             std::string n;
             if (kXbias != Layernorm2dXBiasEnum::NO_BIAS) n += _SS_("_") + Layernorm2dXBiasEnumName<kXbias>::name;
             if (kFusedAdd != Layernorm2dFusedAddEnum::NO_ADD) n += _SS_("_") + Layernorm2dFusedAddEnumName<kFusedAdd>::name;
             if (kFusedQuant != Layernorm2dFusedQuantEnum::NO_SWEEP) n += _SS_("_") + Layernorm2dFusedQuantEnumName<kFusedQuant>::name;
             if (kPadN) n += "_pn";
             if (kSaveMeanInvStd) n += "_mv";
             // if (kTwoPass) n += "_2p";
             return n; }();
  
         auto prec_str = [&] () {
             std::string base_str = _SS_(t2s<XDataType>::name);
             if (!std::is_same_v<XDataType, YDataType>) {
                 base_str += _SS_("_") + _SS_(t2s<YDataType>::name);
             }
             if (kFusedQuant == Layernorm2dFusedQuantEnum::SMOOTH_DYNAMIC_QUANT) {
                 base_str += _SS_("_sx") + _SS_(t2s<SmoothScaleDataType>::name);
                 base_str += _SS_("_sy") + _SS_(t2s<YScaleDataType>::name);
             }
             if (kFusedQuant == Layernorm2dFusedQuantEnum::DYNAMIC_QUANT) {
                 base_str += _SS_("_sy") + _SS_(t2s<YScaleDataType>::name);
             }
             return base_str;
         }();
  
         return _SS_("layernorm2d_fwd_") + _SS_(prec_str) + "_" +
              _TS_(S_::Block_M) + "x" + _TS_(S_::Block_N) + "_" + _TS_(S_::WarpPerBlock_M) + "x" + _TS_(S_::WarpPerBlock_N) + "_" +
              _TS_(S_::Warp_M) + "x" + _TS_(S_::Warp_N) + "_" + _TS_(S_::Vector_M) + "x" + _TS_(S_::Vector_N) + "_" +
              _SS_(Pipeline::name) + surfix;
         // clang-format on
 #undef _SS_
 #undef _TS_
     }
  
     CK_TILE_DEVICE void operator()(Kargs kargs) const
     {
         const auto iM = get_block_id() * Block_M;
  
         const auto x_window = [&]() {
             const auto tmp_ = make_naive_tensor_view<address_space_enum::global>(
                 static_cast<const XDataType*>(kargs.p_x),
                 make_tuple(kargs.m, kargs.n),
                 make_tuple(kargs.x_stride, 1),
                 number<Vector_N>{},
                 number<1>{});
  
             // NOTE: we don't do any pad in this kernel for loading, assume that inside kernel will
             // check the max count dynamically
             const auto tmp2_ = pad_tensor_view(
                 tmp_, make_tuple(number<Block_M>{}, number<Block_N>{}), sequence<false, false>{});
             return make_tile_window(
                 tmp2_, make_tuple(number<Block_M>{}, number<Block_N>{}), {iM, 0});
         }();
  
         const auto x_residual_window = [&]() {
             if constexpr(kFusedAdd == Layernorm2dFusedAddEnum::PRE_ADD_STORE ||
                          kFusedAdd == Layernorm2dFusedAddEnum::PRE_ADD)
             {
                 const auto tmp_ = make_naive_tensor_view<address_space_enum::global>(
                     static_cast<const XResidualDataType*>(kargs.p_x_residual),
                     make_tuple(kargs.m, kargs.n),
                     make_tuple(kargs.xr_stride, 1),
                     number<Vector_N>{},
                     number<1>{});
  
                 // NOTE: we don't do any pad in this kernel for loading, assume that inside kernel
                 // will check the max count dynamically
                 const auto tmp2_ = pad_tensor_view(tmp_,
                                                    make_tuple(number<Block_M>{}, number<Block_N>{}),
                                                    sequence<false, false>{});
                 return make_tile_window(
                     tmp2_, make_tuple(number<Block_M>{}, number<Block_N>{}), {iM, 0});
             }
             else
             {
                 return make_null_tile_window(make_tuple(number<Block_M>{}, number<Block_N>{}));
             }
         }();
  
         const auto x_bias_window = [&]() {
             if constexpr(kXbias == Layernorm2dXBiasEnum::ADD_BIAS)
             {
                 const auto tmp_ = make_naive_tensor_view<address_space_enum::global>(
                     static_cast<const XBiasDataType*>(kargs.p_x_bias),
                     make_tuple(kargs.n),
                     make_tuple(1),
                     number<Vector_N>{},
                     number<1>{});
  
                 const auto tmp2_ =
                     pad_tensor_view(tmp_, make_tuple(number<Block_N>{}), sequence<false>{});
  
                 return make_tile_window(tmp2_, make_tuple(number<Block_N>{}), {0});
             }
             else
             {
                 return make_null_tile_window(make_tuple(number<Block_N>{}));
             }
         }();
  
         const auto gamma_window = [&]() {
             const auto tmp_ = make_naive_tensor_view<address_space_enum::global>(
                 static_cast<const GammaDataType*>(kargs.p_gamma),
                 make_tuple(kargs.n),
                 make_tuple(1),
                 number<Vector_N>{},
                 number<1>{});
  
             const auto tmp2_ =
                 pad_tensor_view(tmp_, make_tuple(number<Block_N>{}), sequence<false>{});
  
             return make_tile_window(tmp2_, make_tuple(number<Block_N>{}), {0});
         }();
  
         const auto beta_window = [&]() {
             const auto tmp_ = make_naive_tensor_view<address_space_enum::global>(
                 static_cast<const BetaDataType*>(kargs.p_beta),
                 make_tuple(kargs.n),
                 make_tuple(1),
                 number<Vector_N>{},
                 number<1>{});
  
             const auto tmp2_ =
                 pad_tensor_view(tmp_, make_tuple(number<Block_N>{}), sequence<false>{});
             return make_tile_window(tmp2_, make_tuple(number<Block_M>{}, number<Block_N>{}), {0});
         }();
  
         auto y_window = [&]() {
             auto tmp_ = make_naive_tensor_view<address_space_enum::global>(
                 static_cast<YDataType*>(kargs.p_y),
                 make_tuple(kargs.m, kargs.n),
                 make_tuple(kargs.y_stride, 1),
                 number<Vector_N>{},
                 number<1>{});
  
             auto tmp2_ = pad_tensor_view(
                 tmp_, make_tuple(number<Block_M>{}, number<Block_N>{}), sequence<kPadM, kPadN>{});
             return make_tile_window(
                 tmp2_, make_tuple(number<Block_M>{}, number<Block_N>{}), {iM, 0});
         }();
  
         auto y_residual_window = [&]() {
             if constexpr(kFusedAdd == Layernorm2dFusedAddEnum::PRE_ADD_STORE)
             {
                 auto tmp_ = make_naive_tensor_view<address_space_enum::global>(
                     static_cast<YResidualDataType*>(kargs.p_y_residual),
                     make_tuple(kargs.m, kargs.n),
                     make_tuple(kargs.yr_stride, 1),
                     number<Vector_N>{},
                     number<1>{});
  
                 auto tmp2_ = pad_tensor_view(tmp_,
                                              make_tuple(number<Block_M>{}, number<Block_N>{}),
                                              sequence<kPadM, kPadN>{});
                 return make_tile_window(
                     tmp2_, make_tuple(number<Block_M>{}, number<Block_N>{}), {iM, 0});
             }
             else
             {
                 return make_null_tile_window(make_tuple(number<Block_M>{}, number<Block_N>{}));
             }
         }();
  
         auto mean_window = [&]() {
             if constexpr(kSaveMean)
             {
                 const auto mean_m = [&]() {
                     const auto mean_dram_naive =
                         make_naive_tensor_view_packed<address_space_enum::global>(
                             static_cast<MeanDataType*>(kargs.p_mean),
                             make_tuple(kargs.m),
                             number<1>{});
  
                     return pad_tensor_view(
                         mean_dram_naive, make_tuple(number<Block_M>{}), sequence<kPadM>{});
                 }();
                 return make_tile_window(mean_m, make_tuple(number<Block_M>{}), {iM});
             }
             else
                 return make_null_tile_window(make_tuple(number<Block_M>{}));
         }();
  
         auto inv_std_window = [&]() {
             if constexpr(kSaveInvStd)
             {
                 const auto inv_std_m = [&]() {
                     const auto inv_std_dram_naive =
                         make_naive_tensor_view_packed<address_space_enum::global>(
                             static_cast<InvStdDataType*>(kargs.p_invStd),
                             make_tuple(kargs.m),
                             number<1>{});
  
                     return pad_tensor_view(
                         inv_std_dram_naive, make_tuple(number<Block_M>{}), sequence<kPadM>{});
                 }();
                 return make_tile_window(inv_std_m, make_tuple(number<Block_M>{}), {iM});
             }
             else
                 return make_null_tile_window(make_tuple(number<Block_M>{}));
         }();
  
         auto sm_scale_window = [&]() {
             if constexpr(kFusedQuant == Layernorm2dFusedQuantEnum::SMOOTH_DYNAMIC_QUANT)
             {
                 const auto win_ = [&]() {
                     const auto tmp_0_ = make_naive_tensor_view_packed<address_space_enum::global>(
                         static_cast<const SmoothScaleDataType*>(kargs.p_sm_scale),
                         make_tuple(kargs.n),
                         number<Vector_N>{});
  
                     return pad_tensor_view(tmp_0_,
                                            make_tuple(number<Block_N>{}),
                                            sequence<false>{}); // sm_scale no need pad
                 }();
                 return make_tile_window(win_, make_tuple(number<Block_N>{}), {0});
             }
             else
                 return make_null_tile_window(make_tuple(number<Block_N>{}));
         }();
  
         auto y_scale_window = [&]() {
             if constexpr(kFusedQuant == Layernorm2dFusedQuantEnum::SMOOTH_DYNAMIC_QUANT ||
                          kFusedQuant == Layernorm2dFusedQuantEnum::DYNAMIC_QUANT)
             {
                 const auto win_ = [&]() {
                     const auto tmp_0_ = make_naive_tensor_view_packed<address_space_enum::global>(
                         static_cast<YScaleDataType*>(kargs.p_y_scale),
                         make_tuple(kargs.m),
                         number<1>{});
  
                     return pad_tensor_view(
                         tmp_0_, make_tuple(number<Block_M>{}), sequence<kPadM>{});
                 }();
                 return make_tile_window(win_, make_tuple(number<Block_M>{}), {iM});
             }
             else
                 return make_null_tile_window(make_tuple(number<Block_M>{}));
         }();
  
         __shared__ char smem[GetSmemSize()];
  
         Pipeline{}(x_window,
                    x_residual_window,
                    x_bias_window,
                    gamma_window,
                    beta_window,
                    y_window,
                    y_residual_window,
                    mean_window,
                    inv_std_window,
                    sm_scale_window,
                    y_scale_window,
                    static_cast<const ComputeDataType>(kargs.epsilon),
                    kargs.n,
                    smem,
                    Epilogue{});
     }
 };
  
 } // namespace ck_tile