ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-06-29 03:07:02 +00:00
Code Issues Packages Projects Releases Wiki Activity

update

Browse Source
This commit is contained in:
Muhammed Emin Ozturk
2024-10-11 12:01:07 -05:00
committed by Muhammed Emin Ozturk
parent bd68dd12b1
commit 7839419475
4 changed files with 14 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
example/01_gemm/gemm_xdl_bf16.cpp
View File

@@ -13,6 +13,7 @@ using CShuffleDataType = ck::bhalf_t;
using ALayout = Row;
using BLayout = Col;
// using BLayout = Row;
using CLayout = Row;
using AElementOp = PassThrough;

4
example/01_gemm/gemm_xdl_streamk.cpp
View File

@@ -14,8 +14,8 @@ using CDataType = ck::half_t;
using F16 = ck::half_t;
using ALayout = Row;
using BLayout = Row;
// using BLayout = Col;
// using BLayout = Row;
using BLayout = Col;
using CLayout = Row;
using AElementOp = PassThrough;

5
example/01_gemm/run_gemm_example_streamk_v2.inc
View File

@@ -181,6 +181,11 @@ bool run_gemm(const ProblemType& problem_size, const ExecutionConfig& config)
std::cout << "b_k_n: " << b_k_n.mDesc << std::endl;
std::cout << "c_m_n: " << c_m_n_host_result.mDesc << std::endl;
//Added By Emin
// Added By Emin
#ifdef BUILD_INT4_EXAMPLE
DeviceMem a_m_k_device_buf(sizeof(KernelADataType) * a_m_k.mDesc.GetElementSpaceSize());
DeviceMem b_k_n_device_buf(sizeof(KernelBDataType) * b_k_n.mDesc.GetElementSpaceSize());

8
script/cmake-ck-dev.sh
View File

@@ -15,8 +15,8 @@ else
fi
cmake \
-D CMAKE_PREFIX_PATH=/opt/rocm \
-D CMAKE_CXX_COMPILER=/opt/rocm/bin/hipcc \
-D CMAKE_PREFIX_PATH=/opt/rocm-6.2.1/ \
-D CMAKE_CXX_COMPILER=/opt/rocm-6.2.1/bin/hipcc \
-D CMAKE_CXX_FLAGS="-Xclang -mllvm -Xclang -enable-post-misched=0 -std=c++17 -O3 -ftemplate-backtrace-limit=0 -fPIE -Wno-gnu-line-marker" \
-D CMAKE_BUILD_TYPE=Release \
-D BUILD_DEV=ON \
@@ -25,3 +25,7 @@ cmake
-D USE_BITINT_EXTENSION_INT4=OFF \
$REST_ARGS \
${MY_PROJECT_SOURCE}
# -D CMAKE_PREFIX_PATH=/opt/rocm \
# -D CMAKE_CXX_COMPILER=/opt/rocm/bin/hipcc