ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-02 12:41:26 +00:00
Code Issues Packages Projects Releases Wiki Activity

layernorm2d forward (#1339)

Browse Source 
* Add layernorm2d forward

* Refind file path

* clang format

* Exclude ck_tile op from all

* use add_executable instead

* refactor layernorm2d_fwd example

---------

Co-authored-by: carlushuang <carlus.huang@amd.com>
This commit is contained in:
rocking
2024-06-24 08:45:52 +08:00
committed by GitHub
parent 05b10e0e5a
commit cb13839425
17 changed files with 975 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files

30
include/ck_tile/ops/layernorm2d/pipeline/block_layernorm2d_fwd_problem.hpp Normal file
View File

@@ -0,0 +1,30 @@
// SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
#pragma once
#include "ck_tile/core/utility/type_traits.hpp"
namespace ck_tile {
template <typename XDataType_,
typename GammaDataType_,
typename BetaDataType_,
typename ComputeDataType_,
typename YDataType_,
typename MeanDataType_,
typename InvStdDataType_,
typename BlockShape_>
struct BlockLayernorm2dFwdProblem
{
using XDataType = remove_cvref_t<XDataType_>;
using GammaDataType = remove_cvref_t<GammaDataType_>;
using BetaDataType = remove_cvref_t<BetaDataType_>;
using ComputeDataType = remove_cvref_t<ComputeDataType_>;
using YDataType = remove_cvref_t<YDataType_>;
using MeanDataType = remove_cvref_t<MeanDataType_>;
using InvStdDataType = remove_cvref_t<InvStdDataType_>;
using BlockShape = remove_cvref_t<BlockShape_>;
};
} // namespace ck_tile

35
include/ck_tile/ops/layernorm2d/pipeline/tile_layernorm2d_fwd_shape.hpp Normal file
View File

@@ -0,0 +1,35 @@
// SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
#pragma once
#include "ck_tile/core.hpp"
namespace ck_tile {
template <typename ThreadTile, // Sequence<...
typename WarpTile, // Sequence<...
typename BlockTile> // Sequence<...
struct TileLayernorm2dShape
{
static constexpr index_t kMPerThread = ThreadTile::at(number<0>{});
static constexpr index_t kNPerThread = ThreadTile::at(number<1>{});
static constexpr index_t kMPerWarp = WarpTile::at(number<0>{});
static constexpr index_t kNPerWarp = WarpTile::at(number<1>{});
static constexpr index_t kMThreadPerWarp = kMPerWarp / kMPerThread;
static constexpr index_t kNThreadPerWarp = kNPerWarp / kNPerThread;
static constexpr index_t kMPerBlock = BlockTile::at(number<0>{});
static constexpr index_t kNPerBlock = BlockTile::at(number<1>{});
static constexpr index_t kMWarpPerBlock = kMPerBlock / kMPerWarp;
static constexpr index_t kNWarpPerBlock = kNPerBlock / kNPerWarp;
// TODO - kNNumWarps can only be 1 if we don't support cross warp welford
static_assert(kNWarpPerBlock == 1);
static constexpr index_t kBlockSize = warpSize * kMWarpPerBlock * kNWarpPerBlock;
};
} // namespace ck_tile