/home/docs/checkouts/readthedocs.org/user_builds/advanced-micro-devices-composable-kernel/checkouts/develop/include/ck_tile/core/tensor/tile_window.hpp Source File

/home/docs/checkouts/readthedocs.org/user_builds/advanced-micro-devices-composable-kernel/checkouts/develop/include/ck_tile/core/tensor/tile_window.hpp Source File#

Composable Kernel: /home/docs/checkouts/readthedocs.org/user_builds/advanced-micro-devices-composable-kernel/checkouts/develop/include/ck_tile/core/tensor/tile_window.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/arch/utility.hpp"
 #include "ck_tile/core/algorithm/space_filling_curve.hpp"
 #include "ck_tile/core/config.hpp"
 #include "ck_tile/core/container/array.hpp"
 #include "ck_tile/core/container/sequence.hpp"
 #include "ck_tile/core/container/tuple.hpp"
 #include "ck_tile/core/container/container_helper.hpp"
 #include "ck_tile/core/tensor/static_distributed_tensor.hpp"
 #include "ck_tile/core/tensor/tensor_adaptor.hpp"
 #include "ck_tile/core/tensor/tile_distribution.hpp"
 #include "ck_tile/core/tensor/tile_window_base.hpp"
 #include "ck_tile/core/utility/functional.hpp"
 #include "ck_tile/core/utility/type_traits.hpp"
  
 namespace ck_tile {
  
 template <typename BottomTensorView_,
           typename WindowLengths_,
           typename StaticTileDistribution_,
           index_t NumCoord>
 struct tile_window_with_static_distribution
     : public tile_window_with_tile_dstr_base<
           tile_window_with_static_distribution<BottomTensorView_,
                                                WindowLengths_,
                                                StaticTileDistribution_,
                                                NumCoord>,
           BottomTensorView_,
           WindowLengths_,
           StaticTileDistribution_>
 {
     using Base = tile_window_with_tile_dstr_base<
         tile_window_with_static_distribution<BottomTensorView_,
                                              WindowLengths_,
                                              StaticTileDistribution_,
                                              NumCoord>,
         BottomTensorView_,
         WindowLengths_,
         StaticTileDistribution_>;
  
     static constexpr auto I0 = number<0>{};
     static constexpr auto I1 = number<1>{};
     static_assert(NumCoord == 1);
  
     static_assert(Base::Traits::NumAccess % NumCoord == 0,
                   "wrong! # of access is not divisible by NumCoord");
     static constexpr index_t NumAccessPerCoord = Base::Traits::NumAccess / NumCoord;
  
     CK_TILE_DEVICE constexpr tile_window_with_static_distribution() = default;
  
     CK_TILE_DEVICE constexpr tile_window_with_static_distribution(
         const typename Base::BottomTensorView& bottom_tensor_view,
         const typename Base::WindowLengths& window_lengths,
         const typename Base::BottomTensorIndex& window_origin,
         const typename Base::TileDstr& tile_distribution)
         : pre_computed_coords_{}
     {
  
         this->window_origin_                       = window_origin;
         this->window_lengths_                      = window_lengths;
         this->bottom_tensor_view_                  = bottom_tensor_view;
         this->tile_dstr_                           = tile_distribution;
         const auto window_adaptor_thread_coord_tmp = make_tensor_adaptor_coordinate(
             tile_distribution.get_ps_ys_to_xs_adaptor(),
             container_concat(detail::get_partition_index(tile_distribution),
                              array<index_t, Base::NDimY>{0}));
  
         typename Base::BottomTensorIndex bottom_tensor_thread_origin_idx_tmp =
             window_origin + window_adaptor_thread_coord_tmp.get_bottom_index();
  
         const auto bottom_tensor_thread_coord_tmp = make_tensor_coordinate(
             bottom_tensor_view.get_tensor_descriptor(), bottom_tensor_thread_origin_idx_tmp);
  
         // pre-compute NumCoord (WindowAdaptorCoord, BottomTensorCoord) bundles to speed up
         // future load/store() calls (might allocate more registers)
         using Traits = typename Base::Traits;
         using SFC_Ys = typename Traits::SFC_Ys;
  
         static_for<0, NumCoord, 1>{}([&](auto iCoord) {
             auto window_adaptor_thread_coord = window_adaptor_thread_coord_tmp;
             auto bottom_tensor_thread_coord  = bottom_tensor_thread_coord_tmp;
  
             constexpr auto idx_diff_ys =
                 SFC_Ys::get_step_between(number<0>{}, number<iCoord * NumAccessPerCoord>{});
  
             constexpr auto idx_diff_ps_ys = container_concat(
                 generate_tuple([&](auto) { return number<0>{}; }, number<Base::NDimP>{}),
                 idx_diff_ys);
  
             Base::move_window_adaptor_and_bottom_tensor_thread_coordinate(
                 window_adaptor_thread_coord, bottom_tensor_thread_coord, idx_diff_ps_ys);
  
             pre_computed_coords_(iCoord) =
                 make_tuple(window_adaptor_thread_coord, bottom_tensor_thread_coord);
         });
     }
  
     template <index_t i_access_unsupport_ = -1, bool oob_conditional_check = true>
     CK_TILE_DEVICE auto load(number<i_access_unsupport_>          = {},
                              bool_constant<oob_conditional_check> = {}) const
     {
         constexpr auto tile_dstr = typename Base::TileDstr{};
         auto dst_tensor = make_static_distributed_tensor<typename Base::DataType>(tile_dstr);
         load(dst_tensor, number<i_access_unsupport_>{}, bool_constant<oob_conditional_check>{});
         return dst_tensor;
     }
  
     template <typename DistributedTensor,
               index_t i_access_unsupport_ = -1,
               bool oob_conditional_check  = true>
     CK_TILE_DEVICE auto load(DistributedTensor& dst_tensor,
                              number<i_access_unsupport_>          = {},
                              bool_constant<oob_conditional_check> = {}) const
     {
         using Traits   = typename Base::Traits;
         using vector_t = typename Traits::vector_t;
         using SFC_Ys   = typename Traits::SFC_Ys;
  
         constexpr auto tile_dstr = typename Base::TileDstr{};
  
         // loop over thread tensor space [y0, y1, ...]
         static_for<0, NumCoord, 1>{}([&](auto iCoord) {
             auto window_adaptor_thread_coord = pre_computed_coords_[iCoord][I0];
             auto bottom_tensor_thread_coord  = pre_computed_coords_[iCoord][I1];
  
             static_for<0, NumAccessPerCoord, 1>{}([&](auto iCoordAccess) {
                 constexpr auto iAccess = number<iCoord * NumAccessPerCoord + iCoordAccess>{};
  
                 // data index [y0, y1, ...]
                 constexpr auto idx_ys_start = SFC_Ys::get_index(iAccess);
  
                 // read from bottom tensor
                 const vector_t vec_value =
                     this->get_bottom_tensor_view().template get_vectorized_elements<vector_t>(
                         bottom_tensor_thread_coord, 0, bool_constant<oob_conditional_check>{});
                 // write into distributed tensor
                 static_for<0, Traits::ScalarPerVector, Traits::PackedSize>{}([&](auto j) {
                     constexpr auto idx_ys = generate_tuple(
                         [&](auto jj) {
                             return jj == Traits::VectorDimY ? (idx_ys_start[jj] + j)
                                                             : idx_ys_start[jj];
                         },
                         number<Base::NDimY>{});
  
                     constexpr index_t d =
                         tile_dstr.get_ys_to_d_descriptor().calculate_offset(idx_ys) /
                         Traits::PackedSize;
  
                     dst_tensor.get_thread_buffer().template at<d>() =
                         vec_value
                             .template get_as<typename Base::DataType>()[j / Traits::PackedSize];
                 });
                 // move thread coordinate
                 if constexpr(iCoordAccess != (NumAccessPerCoord - 1))
                 {
                     constexpr auto idx_diff_ys = SFC_Ys::get_forward_step(iAccess);
  
                     constexpr auto idx_diff_ps_ys = container_concat(
                         generate_tuple([&](auto) { return number<0>{}; }, number<Base::NDimP>{}),
                         idx_diff_ys);
  
                     Base::move_window_adaptor_and_bottom_tensor_thread_coordinate(
                         window_adaptor_thread_coord, bottom_tensor_thread_coord, idx_diff_ps_ys);
                 }
             });
         });
     }
  
     template <typename DstTile,
               index_t i_access_unsupport_ = -1,
               bool oob_conditional_check  = true,
               bool pre_nop                = false>
     CK_TILE_DEVICE void load_raw(DstTile& dst_tensor,
                                  number<i_access_unsupport_>          = {},
                                  bool_constant<oob_conditional_check> = {},
                                  bool_constant<pre_nop>               = {}) const
     {
         using Traits   = typename Base::Traits;
         using vector_t = typename Traits::vector_t;
         using SFC_Ys   = typename Traits::SFC_Ys;
         static constexpr index_t YElementSize =
             typename Base::TileDstr{}.get_ys_to_d_descriptor().get_element_space_size();
         static_assert(YElementSize % (Traits::PackedSize * Traits::ScalarPerVector) == 0);
         using vectorized_tbuf =
             array<vector_t, YElementSize / (Traits::PackedSize * Traits::ScalarPerVector)>;
  
         constexpr auto tile_dstr = typename Base::TileDstr{};
  
         auto& dst_vec_tbuf = reinterpret_cast<vectorized_tbuf&>(dst_tensor.get_thread_buffer());
  
         // loop over thread tensor space [y0, y1, ...]
         static_for<0, NumCoord, 1>{}([&](auto iCoord) {
             auto window_adaptor_thread_coord = pre_computed_coords_[iCoord][I0];
             auto bottom_tensor_thread_coord  = pre_computed_coords_[iCoord][I1];
  
             static_for<0, NumAccessPerCoord, 1>{}([&](auto iCoordAccess) {
                 constexpr auto iAccess  = number<iCoord * NumAccessPerCoord + iCoordAccess>{};
                 constexpr auto pre_nop_ = [&]() {
                     if constexpr(pre_nop && iCoord == 0 && iCoordAccess == 0)
                         return bool_constant<true>{};
                     else
                         return bool_constant<false>{};
                 }();
  
                 // data index [y0, y1, ...]
                 constexpr auto idx_ys_start = SFC_Ys::get_index(iAccess);
                 constexpr index_t d =
                     tile_dstr.get_ys_to_d_descriptor().calculate_offset(idx_ys_start) /
                     Traits::PackedSize;
                 static_assert(d % Traits::ScalarPerVector == 0);
  
                 this->get_bottom_tensor_view().template get_vectorized_elements_raw<vector_t>(
                     dst_vec_tbuf.template at<d / Traits::ScalarPerVector>(),
                     bottom_tensor_thread_coord,
                     0 ,
                     bool_constant<oob_conditional_check>{},
                     pre_nop_);
 #if CK_TILE_WORKAROUND_ROCM_6_1_SCRATCH_MEMORY_ISSUE || \
     CK_TILE_WORKAROUND_ROCM_6_2_SCRATCH_MEMORY_ISSUE
                 asm volatile(
                     ""); // this is starting from rocm-6.2, but same sympton, reuse this flag
 #endif
                 // move thread coordinate
                 if constexpr(iCoordAccess != (NumAccessPerCoord - 1))
                 {
                     constexpr auto idx_diff_ys = SFC_Ys::get_forward_step(iAccess);
  
                     constexpr auto idx_diff_ps_ys = container_concat(
                         generate_tuple([&](auto) { return number<0>{}; }, number<Base::NDimP>{}),
                         idx_diff_ys);
  
                     Base::move_window_adaptor_and_bottom_tensor_thread_coordinate(
                         window_adaptor_thread_coord, bottom_tensor_thread_coord, idx_diff_ps_ys);
                 }
             });
         });
     }
  
     // TODO: currently async load only implemented in inline asm
     template <typename LdsTileWindow_,
               index_t i_access_unsupport_ = -1,
               bool oob_conditional_check  = true,
               bool pre_nop                = false>
     CK_TILE_DEVICE auto async_load_raw(LdsTileWindow_&& lds_tile,
                                        number<i_access_unsupport_>          = {},
                                        bool_constant<oob_conditional_check> = {},
                                        bool_constant<pre_nop>               = {}) const
     {
         using LdsTileWindow = remove_cvref_t<LdsTileWindow_>;
         // using LdsTensorView = typename LdsTileWindow::BottomTensorView;
         using LdsDataType = typename LdsTileWindow::DataType;
         // using LdsDescriptor = typename LdsTileWindow::BottomTensorDesc;
  
         // issues * warps * lanes
         static_assert(LdsTileWindow::get_num_of_dimension() == 3); // TODO: hard coded
  
         const index_t size_per_buf =
             lds_tile.get_bottom_tensor_view().get_tensor_descriptor().calculate_offset(
                 make_tuple(number<0>{}, number<0>{}, number<0>{})) *
             sizeof(LdsDataType);
  
         const index_t size_per_wave =
             lds_tile.get_bottom_tensor_view().get_tensor_descriptor().calculate_offset(
                 make_tuple(number<0>{}, number<1>{}, number<0>{})) *
                 sizeof(LdsDataType) -
             size_per_buf;
  
         const index_t size_per_issue =
             lds_tile.get_bottom_tensor_view().get_tensor_descriptor().calculate_offset(
                 make_tuple(number<1>{}, number<0>{}, number<0>{})) *
                 sizeof(LdsDataType) -
             size_per_buf;
  
         // Use VALU so the compiler can optimize redundant/repeated computations
         const index_t m0_init_value =
             size_per_buf + size_per_wave * get_warp_id(/*ReturnSgpr=*/bool_constant<false>{});
         m0_set_with_memory(
             __builtin_amdgcn_readfirstlane(m0_init_value)); // This should be wave independent
  
         using Traits = typename Base::Traits;
  
         using vector_t = typename Traits::vector_t;
         using SFC_Ys   = typename Traits::SFC_Ys;
  
         LdsDataType* smem = lds_tile.get_bottom_tensor_view().get_buffer_view().p_data_;
  
         // loop over thread tensor space [y0, y1, ...]
         static_for<0, NumCoord, 1>{}([&](auto iCoord) {
             auto window_adaptor_thread_coord = pre_computed_coords_[iCoord][I0];
             auto bottom_tensor_thread_coord  = pre_computed_coords_[iCoord][I1];
  
             static_for<0, NumAccessPerCoord, 1>{}([&](auto iCoordAccess) {
                 constexpr auto iAccess  = number<iCoord * NumAccessPerCoord + iCoordAccess>{};
                 constexpr auto pre_nop_ = [&]() {
                     if constexpr(pre_nop && iCoord == 0 && iCoordAccess == 0)
                         return bool_constant<true>{};
                     else
                         return bool_constant<false>{};
                 }();
  
                 // read from bottom tensor
                 this->get_bottom_tensor_view().template async_get_vectorized_elements_raw<vector_t>(
                     smem, bottom_tensor_thread_coord, 0, pre_nop_);
  
                 // move thread coordinate
                 if constexpr(iCoordAccess != (NumAccessPerCoord - 1))
                 {
                     constexpr auto idx_diff_ys = SFC_Ys::get_forward_step(iAccess);
  
                     constexpr auto idx_diff_ps_ys = container_concat(
                         generate_tuple([&](auto) { return number<0>{}; }, number<Base::NDimP>{}),
                         idx_diff_ys);
  
                     Base::move_window_adaptor_and_bottom_tensor_thread_coordinate(
                         window_adaptor_thread_coord, bottom_tensor_thread_coord, idx_diff_ps_ys);
  
                     m0_inc_with_memory(size_per_issue);
                 }
             });
         });
     }
  
     template <typename LdsTileWindow_,
               index_t i_access_unsupport_ = -1,
               bool oob_conditional_check  = true>
     CK_TILE_DEVICE auto async_load(LdsTileWindow_&& lds_tile,
                                    number<i_access_unsupport_>          = {},
                                    bool_constant<oob_conditional_check> = {}) const
     {
         using LdsTileWindow = remove_cvref_t<LdsTileWindow_>;
         using LdsDataType   = typename LdsTileWindow::DataType;
         using Traits        = typename Base::Traits;
  
         using vector_t = typename Traits::vector_t;
         using SFC_Ys   = typename Traits::SFC_Ys;
  
         // Precompute invariant values outside loops
         const auto window_origin       = lds_tile.get_window_origin();
         const auto& bottom_tensor_view = lds_tile.get_bottom_tensor_view();
         const auto& tensor_descriptor  = bottom_tensor_view.get_tensor_descriptor();
         auto smem_base_ptr             = bottom_tensor_view.get_buffer_view().p_data_;
  
         static_for<0, NumCoord, 1>{}([&](auto iCoord) {
             auto window_adaptor_thread_coord = pre_computed_coords_[iCoord][I0];
             auto bottom_tensor_thread_coord  = pre_computed_coords_[iCoord][I1];
  
             static_for<0, NumAccessPerCoord, 1>{}([&](auto iCoordAccess) {
                 constexpr auto iAccess = number<iCoord * NumAccessPerCoord + iCoordAccess>{};
  
                 // Use precomputed window origin
                 auto lds_bottom_tensor_thread_idx =
                     window_origin + window_adaptor_thread_coord.get_bottom_index();
  
                 // Use precomputed tensor descriptor
                 const auto lds_coord =
                     make_tensor_coordinate(tensor_descriptor, lds_bottom_tensor_thread_idx);
  
                 // Calculate SMEM address using base pointer
                 CK_TILE_LDS_ADDR LdsDataType* smem = smem_base_ptr + lds_coord.get_offset();
  
                 // Write into bottom tensor
                 this->get_bottom_tensor_view().template async_get_vectorized_elements<vector_t>(
                     smem,
                     bottom_tensor_thread_coord,
                     number<0>{},
                     bool_constant<oob_conditional_check>{});
  
                 // Move thread coordinate if not last access
                 if constexpr(iCoordAccess != (NumAccessPerCoord - 1))
                 {
                     constexpr auto idx_diff_ys    = SFC_Ys::get_forward_step(iAccess);
                     constexpr auto idx_diff_ps_ys = container_concat(
                         generate_tuple([&](auto) { return number<0>{}; }, number<Base::NDimP>{}),
                         idx_diff_ys);
  
                     Base::move_window_adaptor_and_bottom_tensor_thread_coordinate(
                         window_adaptor_thread_coord, bottom_tensor_thread_coord, idx_diff_ps_ys);
                 }
             });
         });
     }
  
     template <typename Policy, index_t i_access_unsupport_ = -1, bool oob_conditional_check = true>
     CK_TILE_DEVICE auto load_transpose() const
     {
         constexpr auto tile_dstr = typename Base::TileDstr{};
         auto dst_tensor = make_static_distributed_tensor<typename Base::DataType>(tile_dstr);
         this->template load_transpose<Policy>(
             dst_tensor, number<i_access_unsupport_>{}, bool_constant<oob_conditional_check>{});
         return dst_tensor;
     }
  
     template <typename Policy,
               typename DistributedTensor,
               index_t i_access_unsupport_ = -1,
               bool oob_conditional_check  = true>
     CK_TILE_DEVICE auto load_transpose(DistributedTensor& dst_tensor,
                                        number<i_access_unsupport_>          = {},
                                        bool_constant<oob_conditional_check> = {}) const
     {
         using Traits   = typename Base::Traits;
         using vector_t = typename Traits::vector_t;
         using SFC_Ys   = typename Traits::SFC_Ys;
  
         constexpr auto tile_dstr = typename Base::TileDstr{};
  
         constexpr auto group_func = Policy::group_func;
  
         // loop over thread tensor space [y0, y1, ...]
         static_for<0, NumCoord, 1>{}([&](auto iCoord) {
             auto window_adaptor_thread_coord = pre_computed_coords_[iCoord][I0];
             auto bottom_tensor_thread_coord  = pre_computed_coords_[iCoord][I1];
  
             static_for<0, NumAccessPerCoord, 1>{}([&](auto iCoordAccess) {
                 constexpr auto iAccess = number<iCoord * NumAccessPerCoord + iCoordAccess>{};
  
                 // data index [y0, y1, ...]
                 constexpr auto idx_ys_start = SFC_Ys::get_index(iAccess);
  
                 // read from bottom tensor
                 const vector_t vec_value =
                     this->get_bottom_tensor_view()
                         .template get_transpose_vectorized_elements<vector_t>(
                             bottom_tensor_thread_coord, 0);
                 // write into distributed tensor
                 static_for<0, Traits::ScalarPerVector, 1>{}([&](auto j) {
                     constexpr auto orig_idx_ys = generate_tuple(
                         [&](auto jj) {
                             return jj == Traits::VectorDimY ? (idx_ys_start[jj] + j)
                                                             : idx_ys_start[jj];
                         },
                         number<Base::NDimY>{});
  
                     constexpr auto grouped_idx_ys = group_func(orig_idx_ys);
  
                     constexpr index_t linear_distributed_index =
                         tile_dstr.get_ys_to_d_descriptor().calculate_offset(grouped_idx_ys);
  
                     dst_tensor.get_thread_buffer().template at<linear_distributed_index>() =
                         vec_value.template get_as<typename Base::DataType>()[j];
                 });
                 // move thread coordinate
                 if constexpr(iCoordAccess != (NumAccessPerCoord - 1))
                 {
                     constexpr auto idx_diff_ys = SFC_Ys::get_forward_step(iAccess);
  
                     constexpr auto idx_diff_ps_ys = container_concat(
                         generate_tuple([&](auto) { return number<0>{}; }, number<Base::NDimP>{}),
                         idx_diff_ys);
  
                     Base::move_window_adaptor_and_bottom_tensor_thread_coordinate(
                         window_adaptor_thread_coord, bottom_tensor_thread_coord, idx_diff_ps_ys);
                 }
             });
         });
     }
  
     template <index_t i_access_unsupport_ = -1, bool oob_conditional_check = true>
     CK_TILE_DEVICE void store(const static_distributed_tensor<typename Base::DataType,
                                                               typename Base::TileDstr>& dstr_tensor,
                               number<i_access_unsupport_>          = {},
                               bool_constant<oob_conditional_check> = {}) const
     {
         using Traits = typename Base::Traits;
  
         using vector_t = typename Traits::vector_t;
         using SFC_Ys   = typename Traits::SFC_Ys;
  
         constexpr auto tile_dstr = typename Base::TileDstr{};
  
         // loop over thread tensor space [y0, y1, ...]
         static_for<0, NumCoord, 1>{}([&](auto iCoord) {
             auto window_adaptor_thread_coord = pre_computed_coords_[iCoord][I0];
             auto bottom_tensor_thread_coord  = pre_computed_coords_[iCoord][I1];
  
             static_for<0, NumAccessPerCoord, 1>{}([&](auto iCoordAccess) {
                 constexpr auto iAccess = number<iCoord * NumAccessPerCoord + iCoordAccess>{};
  
                 // data index [y0, y1, ...]
                 constexpr auto idx_ys_start = SFC_Ys::get_index(iAccess);
  
                 // read from distributed tensor
                 // vector_type_t vec;
                 vector_t vec_value;
  
                 static_for<0, Traits::ScalarPerVector, Traits::PackedSize>{}([&](auto j) {
                     constexpr auto idx_ys = generate_tuple(
                         [&](auto jj) {
                             return jj == Traits::VectorDimY ? (idx_ys_start[jj] + j)
                                                             : idx_ys_start[jj];
                         },
                         number<Base::NDimY>{});
  
                     constexpr index_t d =
                         tile_dstr.get_ys_to_d_descriptor().calculate_offset(idx_ys) /
                         Traits::PackedSize;
  
                     vec_value.template get_as<typename Base::DataType>()(j / Traits::PackedSize) =
                         dstr_tensor.get_thread_buffer().template at<d>();
                 });
  
                 // const vector_t vec_value = vec.template get_as<vector_t>().template at<0>();
  
                 // write into bottom tensor
                 this->get_bottom_tensor_view().template set_vectorized_elements<vector_t>(
                     bottom_tensor_thread_coord,
                     0,
                     vec_value,
                     bool_constant<oob_conditional_check>{});
  
                 // move thread coordinate
                 if constexpr(iCoordAccess != (NumAccessPerCoord - 1))
                 {
                     constexpr auto idx_diff_ys = SFC_Ys::get_forward_step(iAccess);
  
                     constexpr auto idx_diff_ps_ys = container_concat(
                         generate_tuple([&](auto) { return number<0>{}; }, number<Base::NDimP>{}),
                         idx_diff_ys);
  
                     Base::move_window_adaptor_and_bottom_tensor_thread_coordinate(
                         window_adaptor_thread_coord, bottom_tensor_thread_coord, idx_diff_ps_ys);
                 }
             });
         });
     }
  
     template <index_t i_access_unsupport_ = -1>
     CK_TILE_DEVICE void
     store_raw(const static_distributed_tensor<typename Base::DataType, typename Base::TileDstr>&
                   dstr_tensor,
               number<i_access_unsupport_> = {}) const
     {
         using Traits = typename Base::Traits;
  
         using vector_t = typename Traits::vector_t;
         using SFC_Ys   = typename Traits::SFC_Ys;
  
         constexpr auto tile_dstr                    = typename Base::TileDstr{};
         static constexpr bool oob_conditional_check = true;
  
         // loop over thread tensor space [y0, y1, ...]
         static_for<0, NumCoord, 1>{}([&](auto iCoord) {
             auto window_adaptor_thread_coord = pre_computed_coords_[iCoord][I0];
             auto bottom_tensor_thread_coord  = pre_computed_coords_[iCoord][I1];
  
             static_for<0, NumAccessPerCoord, 1>{}([&](auto iCoordAccess) {
                 constexpr auto iAccess = number<iCoord * NumAccessPerCoord + iCoordAccess>{};
  
                 // data index [y0, y1, ...]
                 constexpr auto idx_ys_start = SFC_Ys::get_index(iAccess);
  
                 // read from distributed tensor
                 vector_t vec_value;
                 static_for<0, Traits::ScalarPerVector, Traits::PackedSize>{}([&](auto j) {
                     constexpr auto idx_ys = generate_tuple(
                         [&](auto jj) {
                             return jj == Traits::VectorDimY ? (idx_ys_start[jj] + j)
                                                             : idx_ys_start[jj];
                         },
                         number<Base::NDimY>{});
                     constexpr index_t d =
                         tile_dstr.get_ys_to_d_descriptor().calculate_offset(idx_ys) /
                         Traits::PackedSize;
                     vec_value.template get_as<typename Base::DataType>()(j / Traits::PackedSize) =
                         dstr_tensor.get_thread_buffer().template at<d>();
                 });
  
                 // write into bottom tensor
                 this->get_bottom_tensor_view()
                     .template set_vectorized_elements_raw<vector_t, oob_conditional_check>(
                         bottom_tensor_thread_coord, 0, vec_value);
  
                 // move thread coordinate
                 if constexpr(iCoordAccess != (NumAccessPerCoord - 1))
                 {
                     constexpr auto idx_diff_ys = SFC_Ys::get_forward_step(iAccess);
  
                     constexpr auto idx_diff_ps_ys = container_concat(
                         generate_tuple([&](auto) { return number<0>{}; }, number<Base::NDimP>{}),
                         idx_diff_ys);
  
                     Base::move_window_adaptor_and_bottom_tensor_thread_coordinate(
                         window_adaptor_thread_coord, bottom_tensor_thread_coord, idx_diff_ps_ys);
                 }
             });
         });
     }
  
     template <index_t i_access_unsupport_ = -1, bool oob_conditional_check = true>
     CK_TILE_DEVICE void
     update(const static_distributed_tensor<typename Base::DataType, typename Base::TileDstr>&
                dstr_tensor,
            number<i_access_unsupport_>          = {},
            bool_constant<oob_conditional_check> = {}) const
     {
         using Traits = typename Base::Traits;
  
         using vector_t = typename Traits::vector_t;
         using SFC_Ys   = typename Traits::SFC_Ys;
  
         constexpr auto tile_dstr = typename Base::TileDstr{};
  
         // loop over thread tensor space [y0, y1, ...]
         static_for<0, NumCoord, 1>{}([&](auto iCoord) {
             auto window_adaptor_thread_coord = pre_computed_coords_[iCoord][I0];
             auto bottom_tensor_thread_coord  = pre_computed_coords_[iCoord][I1];
  
             static_for<0, NumAccessPerCoord, 1>{}([&](auto iCoordAccess) {
                 constexpr auto iAccess = number<iCoord * NumAccessPerCoord + iCoordAccess>{};
  
                 // data index [y0, y1, ...]
                 constexpr auto idx_ys_start = SFC_Ys::get_index(iAccess);
  
                 // read from distributed tensor
                 vector_t vec_value;
  
                 static_for<0, Traits::ScalarPerVector, Traits::PackedSize>{}([&](auto j) {
                     constexpr auto idx_ys = generate_tuple(
                         [&](auto jj) {
                             return jj == Traits::VectorDimY ? (idx_ys_start[jj] + j)
                                                             : idx_ys_start[jj];
                         },
                         number<Base::NDimY>{});
  
                     constexpr index_t d =
                         tile_dstr.get_ys_to_d_descriptor().calculate_offset(idx_ys) /
                         Traits::PackedSize;
  
                     vec_value.template get_as<typename Base::DataType>()(j / Traits::PackedSize) =
                         dstr_tensor.get_thread_buffer().template at<d>();
                 });
  
                 // write into bottom tensor
                 this->get_bottom_tensor_view().template update_vectorized_elements<vector_t>(
                     bottom_tensor_thread_coord,
                     0,
                     vec_value,
                     bool_constant<oob_conditional_check>{});
  
                 // move thread coordinate
                 if constexpr(iCoordAccess != (NumAccessPerCoord - 1))
                 {
                     constexpr auto idx_diff_ys = SFC_Ys::get_forward_step(iAccess);
  
                     constexpr auto idx_diff_ps_ys = container_concat(
                         generate_tuple([&](auto) { return number<0>{}; }, number<Base::NDimP>{}),
                         idx_diff_ys);
  
                     Base::move_window_adaptor_and_bottom_tensor_thread_coordinate(
                         window_adaptor_thread_coord, bottom_tensor_thread_coord, idx_diff_ps_ys);
                 }
             });
         });
     }
  
     template <index_t i_access_unsupport_ = -1, bool oob_conditional_check = true, bool pre_nop>
     CK_TILE_DEVICE void
     update_raw(const static_distributed_tensor<typename Base::DataType, typename Base::TileDstr>&
                    dstr_tensor,
                number<i_access_unsupport_>          = {},
                bool_constant<oob_conditional_check> = {},
                bool_constant<pre_nop>               = {}) const
     {
         using Traits = typename Base::Traits;
  
         using vector_t = typename Traits::vector_t;
         using SFC_Ys   = typename Traits::SFC_Ys;
  
         constexpr auto tile_dstr = typename Base::TileDstr{};
  
         // loop over thread tensor space [y0, y1, ...]
         static_for<0, NumCoord, 1>{}([&](auto iCoord) {
             auto window_adaptor_thread_coord = pre_computed_coords_[iCoord][I0];
             auto bottom_tensor_thread_coord  = pre_computed_coords_[iCoord][I1];
  
             static_for<0, NumAccessPerCoord, 1>{}([&](auto iCoordAccess) {
                 constexpr auto iAccess = number<iCoord * NumAccessPerCoord + iCoordAccess>{};
  
                 // data index [y0, y1, ...]
                 constexpr auto idx_ys_start = SFC_Ys::get_index(iAccess);
  
                 // read from distributed tensor
                 vector_t vec_value;
  
                 static_for<0, Traits::ScalarPerVector, Traits::PackedSize>{}([&](auto j) {
                     constexpr auto idx_ys = generate_tuple(
                         [&](auto jj) {
                             return jj == Traits::VectorDimY ? (idx_ys_start[jj] + j)
                                                             : idx_ys_start[jj];
                         },
                         number<Base::NDimY>{});
  
                     constexpr index_t d =
                         tile_dstr.get_ys_to_d_descriptor().calculate_offset(idx_ys) /
                         Traits::PackedSize;
  
                     vec_value.template get_as<typename Base::DataType>()(j / Traits::PackedSize) =
                         dstr_tensor.get_thread_buffer().template at<d>();
                 });
  
                 // write into bottom tensor
                 this->get_bottom_tensor_view().template update_vectorized_elements_raw<vector_t>(
                     bottom_tensor_thread_coord,
                     0,
                     vec_value,
                     bool_constant<oob_conditional_check>{},
                     bool_constant<pre_nop>{});
  
                 // move thread coordinate
                 if constexpr(iCoordAccess != (NumAccessPerCoord - 1))
                 {
                     constexpr auto idx_diff_ys = SFC_Ys::get_forward_step(iAccess);
  
                     constexpr auto idx_diff_ps_ys = container_concat(
                         generate_tuple([&](auto) { return number<0>{}; }, number<Base::NDimP>{}),
                         idx_diff_ys);
  
                     Base::move_window_adaptor_and_bottom_tensor_thread_coordinate(
                         window_adaptor_thread_coord, bottom_tensor_thread_coord, idx_diff_ps_ys);
                 }
             });
         });
     }
  
     // Custom move behavior
     CK_TILE_DEVICE void move_extended(const typename Base::BottomTensorIndex& step)
     {
         static_for<0, NumCoord, 1>{}([&](auto iCoord) {
             move_tensor_coordinate(this->bottom_tensor_view_.get_tensor_descriptor(),
                                    pre_computed_coords_(iCoord)(I1),
                                    step);
         });
     }
  
     CK_TILE_DEVICE void set_window_origin_extended(const typename Base::BottomTensorIndex&)
     {
         // TODO: this use less register for FA, but more register for GEMM
         // need investigation
         const auto window_adaptor_thread_coord_tmp = make_tensor_adaptor_coordinate(
             this->tile_dstr_.get_ps_ys_to_xs_adaptor(),
             container_concat(detail::get_partition_index(this->tile_dstr_),
                              array<index_t, Base::NDimY>{0}));
  
         typename Base::BottomTensorIndex bottom_tensor_thread_origin_idx_tmp =
             this->window_origin_ + window_adaptor_thread_coord_tmp.get_bottom_index();
  
         const auto bottom_tensor_thread_coord_tmp = make_tensor_coordinate(
             this->bottom_tensor_view_.get_tensor_descriptor(), bottom_tensor_thread_origin_idx_tmp);
  
         // pre-compute NumCoord (WindowAdaptorCoord, BottomTensorCoord) bundles to speed up
         // future load/store() calls (might allocate more registers)
         using Traits = typename Base::Traits;
         using SFC_Ys = typename Traits::SFC_Ys;
  
         static_for<0, NumCoord, 1>{}([&](auto iCoord) {
             auto window_adaptor_thread_coord = window_adaptor_thread_coord_tmp;
             auto bottom_tensor_thread_coord  = bottom_tensor_thread_coord_tmp;
  
             constexpr auto idx_diff_ys =
                 SFC_Ys::get_step_between(number<0>{}, number<iCoord * NumAccessPerCoord>{});
  
             constexpr auto idx_diff_ps_ys = container_concat(
                 generate_tuple([&](auto) { return number<0>{}; }, number<Base::NDimP>{}),
                 idx_diff_ys);
  
             Base::move_window_adaptor_and_bottom_tensor_thread_coordinate(
                 window_adaptor_thread_coord, bottom_tensor_thread_coord, idx_diff_ps_ys);
  
             pre_computed_coords_(iCoord) =
                 make_tuple(window_adaptor_thread_coord, bottom_tensor_thread_coord);
         });
     }
  
     // this contains:
     //   per-thread coordinate for window adaptor
     //   per-thread coordinate for bottom tensor
     array<tuple<typename Base::WindowAdaptorCoord, typename Base::BottomTensorCoord>, NumCoord>
         pre_computed_coords_;
 };
  
 // TODO: use strategy
 template <typename TensorView_,
           typename WindowLengths_,
           typename StaticTileDistribution_,
           index_t NumCoord = 1>
 CK_TILE_DEVICE constexpr auto
 make_tile_window(const TensorView_& tensor_view,
                  const WindowLengths_& window_lengths,
                  const multi_index<TensorView_::get_num_of_dimension()>& origin,
                  const StaticTileDistribution_& tile_distribution,
                  number<NumCoord> = {})
 {
     return tile_window_with_static_distribution<remove_cvref_t<TensorView_>,
                                                 remove_cvref_t<WindowLengths_>,
                                                 remove_cvref_t<StaticTileDistribution_>,
                                                 NumCoord>{
         tensor_view, window_lengths, origin, tile_distribution};
 }
  
 // this version can't be called in a constexpr context
 template <typename TensorView_,
           typename WindowLengths_,
           typename StaticTileDistribution_,
           index_t NumCoord = 1>
 CK_TILE_DEVICE auto
 make_tile_window_raw(const TensorView_& tensor_view,
                      const WindowLengths_& window_lengths,
                      const multi_index<TensorView_::get_num_of_dimension()>& origin,
                      const StaticTileDistribution_& tile_distribution,
                      number<NumCoord> = {})
 {
     auto w = tile_window_with_static_distribution<remove_cvref_t<TensorView_>,
                                                   remove_cvref_t<WindowLengths_>,
                                                   remove_cvref_t<StaticTileDistribution_>,
                                                   NumCoord>{
         tensor_view, window_lengths, origin, tile_distribution};
     w.init_raw();
     return w;
 }
  
 template <typename TensorView_,
           typename WindowLengths_,
           typename StaticTileDistribution_,
           index_t NumCoord>
 CK_TILE_DEVICE void move_tile_window(
     tile_window_with_static_distribution<TensorView_,
                                          WindowLengths_,
                                          StaticTileDistribution_,
                                          NumCoord>& window,
     const typename tile_window_with_static_distribution<TensorView_,
                                                         WindowLengths_,
                                                         StaticTileDistribution_,
                                                         NumCoord>::BottomTensorIndex& step)
 {
     window.move(step);
 }
  
 template <typename BottomTensorView_, typename WindowLengths_>
 struct tile_window_with_static_lengths
     : public tile_window_base<tile_window_with_static_lengths<BottomTensorView_, WindowLengths_>,
                               BottomTensorView_,
                               WindowLengths_>
 {
     using Base =
         tile_window_base<tile_window_with_static_lengths<BottomTensorView_, WindowLengths_>,
                          BottomTensorView_,
                          WindowLengths_>;
  
     CK_TILE_DEVICE constexpr tile_window_with_static_lengths() = default;
  
     CK_TILE_DEVICE constexpr tile_window_with_static_lengths(
         const typename Base::BottomTensorView& bottom_tensor_view,
         const typename Base::WindowLengths& window_lengths,
         const typename Base::BottomTensorIndex& window_origin)
     {
         this->window_origin_      = window_origin;
         this->window_lengths_     = window_lengths;
         this->bottom_tensor_view_ = bottom_tensor_view;
     }
 };
  
 template <typename TensorView_, typename WindowLengths_>
 CK_TILE_DEVICE constexpr auto
 make_tile_window(const TensorView_& tensor_view,
                  const WindowLengths_& window_lengths,
                  const multi_index<TensorView_::get_num_of_dimension()>& origin)
 {
     static_assert(ck_tile::is_known_at_compile_time<WindowLengths_>::value,
                   "wrong! lengths should be static");
  
     return tile_window_with_static_lengths<remove_cvref_t<TensorView_>,
                                            remove_cvref_t<WindowLengths_>>{
         tensor_view, window_lengths, origin};
 }
  
 // duplicate tile window and replace its origin
 template <typename TensorView, typename WindowLengths>
 CK_TILE_DEVICE constexpr auto
 make_tile_window(const tile_window_with_static_lengths<TensorView, WindowLengths>& tile_window,
                  const multi_index<TensorView::get_num_of_dimension()>& origin)
 {
     return tile_window_with_static_lengths<TensorView, WindowLengths>{
         tile_window.get_bottom_tensor_view(), tile_window.get_window_lengths(), origin};
 }
  
 template <typename TensorView, typename WindowLengths, typename StaticTileDistribution>
 CK_TILE_DEVICE constexpr auto
 make_tile_window(const tile_window_with_static_lengths<TensorView, WindowLengths>& tile_window,
                  const multi_index<TensorView::get_num_of_dimension()>& origin,
                  const StaticTileDistribution& tile_distribution)
 {
     return make_tile_window(tile_window.get_bottom_tensor_view(),
                             tile_window.get_window_lengths(),
                             origin,
                             tile_distribution);
 }
  
 template <typename TensorView, typename WindowLengths, typename StaticTileDistribution>
 CK_TILE_DEVICE constexpr auto
 make_tile_window(const tile_window_with_static_lengths<TensorView, WindowLengths>& tile_window,
                  const StaticTileDistribution& tile_distribution)
 {
     return make_tile_window(tile_window.get_bottom_tensor_view(),
                             tile_window.get_window_lengths(),
                             tile_window.get_window_origin(),
                             tile_distribution);
 }
  
 template <typename TensorView, typename WindowLengths, typename StaticTileDistribution>
 CK_TILE_DEVICE constexpr auto
 make_tile_window_raw(const tile_window_with_static_lengths<TensorView, WindowLengths>& tile_window,
                      const StaticTileDistribution& tile_distribution)
 {
     auto w = make_tile_window(tile_window.get_bottom_tensor_view(),
                               tile_window.get_window_lengths(),
                               tile_window.get_window_origin(),
                               tile_distribution);
     w.init_raw();
     return w;
 }
  
 template <typename TensorView_, typename WindowLengths_>
 CK_TILE_DEVICE void move_tile_window(
     tile_window_with_static_lengths<TensorView_, WindowLengths_>& window,
     const typename tile_window_with_static_lengths<TensorView_, WindowLengths_>::BottomTensorIndex&
         step)
 {
     window.move(step);
 }
  
 template <typename T>
 struct is_tile_window_with_static_distribution : std::false_type
 {
 };
  
 template <typename BottomTensorView_,
           typename WindowLengths_,
           typename StaticTileDistribution_,
           index_t NumCoord>
 struct is_tile_window_with_static_distribution<
     tile_window_with_static_distribution<BottomTensorView_,
                                          WindowLengths_,
                                          StaticTileDistribution_,
                                          NumCoord>> : std::true_type
 {
 };
  
 template <typename T>
 inline constexpr bool is_tile_window_with_static_distribution_v =
     is_tile_window_with_static_distribution<T>::value;
  
 template <typename T>
 struct is_tile_window_with_static_lengths : std::false_type
 {
 };
  
 template <typename BottomTensorView_, typename WindowLengths_>
 struct is_tile_window_with_static_lengths<
     tile_window_with_static_lengths<BottomTensorView_, WindowLengths_>> : std::true_type
 {
 };
  
 template <typename T>
 inline constexpr bool is_tile_window_with_static_lengths_v =
     is_tile_window_with_static_lengths<T>::value;
  
 } // namespace ck_tile