mirror of
https://github.com/ROCm/composable_kernel.git
synced 2026-05-13 09:45:56 +00:00
cd51732690
* Fix for lwpck-425, update BlockTransferSrcVectorDim
* Revert "Fix for lwpck-425, update BlockTransferSrcVectorDim"
This reverts commit fd24e280e2.
* Add Batched Gemm int8 test, expect it to fail
* Format
* Re-add the fix
45 lines
1.4 KiB
C++
45 lines
1.4 KiB
C++
// SPDX-License-Identifier: MIT
|
|
// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
|
|
|
|
#include <iostream>
|
|
|
|
#include "profiler/include/profile_batched_gemm_impl.hpp"
|
|
|
|
namespace {
|
|
using ADataType = int8_t;
|
|
using BDataType = int8_t;
|
|
using CDataType = int8_t;
|
|
|
|
using Row = ck::tensor_layout::gemm::RowMajor;
|
|
using Col = ck::tensor_layout::gemm::ColumnMajor;
|
|
} // namespace
|
|
|
|
int main()
|
|
{
|
|
int M = 256;
|
|
int N = 256;
|
|
int K = 128;
|
|
int BatchCount = 3;
|
|
|
|
bool pass = true;
|
|
|
|
pass = pass &&
|
|
ck::profiler::profile_batched_gemm_impl<ADataType, BDataType, CDataType, Row, Row, Row>(
|
|
true, 1, false, 1, M, N, K, K, N, N, M * K, K * N, M * N, BatchCount);
|
|
|
|
pass = pass &&
|
|
ck::profiler::profile_batched_gemm_impl<ADataType, BDataType, CDataType, Row, Col, Row>(
|
|
true, 1, false, 1, M, N, K, K, K, N, M * K, K * N, M * N, BatchCount);
|
|
|
|
pass = pass &&
|
|
ck::profiler::profile_batched_gemm_impl<ADataType, BDataType, CDataType, Col, Row, Row>(
|
|
true, 1, false, 1, M, N, K, M, N, N, M * K, K * N, M * N, BatchCount);
|
|
|
|
pass = pass &&
|
|
ck::profiler::profile_batched_gemm_impl<ADataType, BDataType, CDataType, Col, Col, Row>(
|
|
true, 1, false, 1, M, N, K, M, K, N, M * K, K * N, M * N, BatchCount);
|
|
|
|
std::cout << "test BatchedGEMM int8: " << (pass ? "Pass" : "Fail") << std::endl;
|
|
return pass ? 0 : 1;
|
|
}
|