ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-05 14:11:29 +00:00
Code Issues Packages Projects Releases Wiki Activity

add vectorloads on non-k dim for memory pipelines (#1856)

Browse Source
This commit is contained in:
jakpiase
2025-02-06 21:12:13 +01:00
committed by Sam Wu
parent 82cda34dfe
commit 9b5dfba242
3 changed files with 251 additions and 72 deletions
Download Patch File Download Diff File Expand all files Collapse all files

30
test/ck_tile/gemm/test_gemm_pipeline.cpp
View File

@@ -1,5 +1,5 @@
// SPDX-License-Identifier: MIT
// Copyright (c) 2024, Advanced Micro Devices, Inc. All rights reserved.
// Copyright (c) 2024-2025, Advanced Micro Devices, Inc. All rights reserved.
#include <tuple>
@@ -14,28 +14,26 @@ using Row = ck_tile::tensor_layout::gemm::RowMajor;
using Col = ck_tile::tensor_layout::gemm::ColumnMajor;
using Intrawave = ck_tile::integral_constant<ck_tile::GemmPipelineScheduler,
ck_tile::GemmPipelineScheduler::Intrawave>;
// using Interwave = ck_tile::integral_constant<ck_tile::GemmPipelineScheduler,
// ck_tile::GemmPipelineScheduler::Interwave>;
// using Mem = ck_tile::integral_constant<GemmPipelineType, GemmPipelineType::Mem>;
using Comp = ck_tile::integral_constant<GemmPipelineType, GemmPipelineType::Comp>;
// TODO: Enable Memory pipeline, when it would be updated for vector loads on non-K major tensors.
using Interwave = ck_tile::integral_constant<ck_tile::GemmPipelineScheduler,
ck_tile::GemmPipelineScheduler::Interwave>;
using Mem = ck_tile::integral_constant<GemmPipelineType, GemmPipelineType::Mem>;
using Comp = ck_tile::integral_constant<GemmPipelineType, GemmPipelineType::Comp>;
// clang-format off
using KernelTypes = ::testing::Types<
// ALayout, BLayout, CLayout, ADataType, BDataType, AccDataType, CDataType, GemmPipelineScheduler, PipelineType
// std::tuple< Row, Row, Row, F16, F16, F32, F16, Intrawave, Mem>,
std::tuple< Row, Row, Row, F16, F16, F32, F16, Intrawave, Mem>,
std::tuple< Row, Row, Row, F16, F16, F32, F16, Intrawave, Comp>,
// std::tuple< Row, Row, Row, F16, F16, F32, F16, Interwave, Mem>,
// std::tuple< Row, Col, Row, F16, F16, F32, F16, Intrawave, Mem>,
std::tuple< Row, Row, Row, F16, F16, F32, F16, Interwave, Mem>,
std::tuple< Row, Col, Row, F16, F16, F32, F16, Intrawave, Mem>,
std::tuple< Row, Col, Row, F16, F16, F32, F16, Intrawave, Comp>,
// std::tuple< Row, Col, Row, F16, F16, F32, F16, Interwave, Mem>,
// std::tuple< Col, Row, Row, F16, F16, F32, F16, Intrawave, Mem>,
std::tuple< Row, Col, Row, F16, F16, F32, F16, Interwave, Mem>,
std::tuple< Col, Row, Row, F16, F16, F32, F16, Intrawave, Mem>,
std::tuple< Col, Row, Row, F16, F16, F32, F16, Intrawave, Comp>,
// std::tuple< Col, Row, Row, F16, F16, F32, F16, Interwave, Mem>,
// std::tuple< Col, Col, Row, F16, F16, F32, F16, Intrawave, Mem>,
std::tuple< Col, Col, Row, F16, F16, F32, F16, Intrawave, Comp>
// std::tuple< Col, Col, Row, F16, F16, F32, F16, Interwave, Mem>
std::tuple< Col, Row, Row, F16, F16, F32, F16, Interwave, Mem>,
std::tuple< Col, Col, Row, F16, F16, F32, F16, Intrawave, Mem>,
std::tuple< Col, Col, Row, F16, F16, F32, F16, Intrawave, Comp>,
std::tuple< Col, Col, Row, F16, F16, F32, F16, Interwave, Mem>
>;
// clang-format on