ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-04-20 14:59:17 +00:00
Code Issues Packages Projects Releases Wiki Activity

feature:tf32:add initial conv3d fwd kernel support (#2763)

Browse Source
This commit is contained in:
lym
2025-09-15 21:03:00 +08:00
committed by GitHub
parent e5d73da2da
commit c51102144f
44 changed files with 1085 additions and 175 deletions
Download Patch File Download Diff File Expand all files Collapse all files

16
example/01_gemm/common.hpp
View File

@@ -310,10 +310,14 @@ bool parse_cmd_args<ProblemSizeSplitK>(int argc,
return true;
}
template <typename DataType>
template <typename DataType, typename ComputeDataType = DataType>
inline __host__ __device__ constexpr double get_rtol()
{
if constexpr(std::is_same_v<DataType, float>)
if constexpr(std::is_same_v<DataType, float> && std::is_same_v<ComputeDataType, ck::tf32_t>)
{
return 1e-3;
}
else if constexpr(std::is_same_v<DataType, float>)
{
return 1e-3;
}
@@ -351,10 +355,14 @@ inline __host__ __device__ constexpr double get_rtol()
}
}
template <typename DataType>
template <typename DataType, typename ComputeDataType = DataType>
inline __host__ __device__ constexpr double get_atol()
{
if constexpr(std::is_same_v<DataType, float>)
if constexpr(std::is_same_v<DataType, float> && std::is_same_v<ComputeDataType, ck::tf32_t>)
{
return 1e-3;
}
else if constexpr(std::is_same_v<DataType, float>)
{
return 1e-3;
}