ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-04-20 06:49:15 +00:00
Code Issues Packages Projects Releases Wiki Activity

Add element op (#1259)

Browse Source
This commit is contained in:
Rostyslav Geyyer
2024-04-26 12:55:45 -05:00
committed by GitHub
parent 5ae893c0d3
commit f044ff71fb
1 changed files with 20 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

20
include/ck/tensor_operation/gpu/element/element_wise_operation.hpp
View File

@@ -528,6 +528,26 @@ struct UnaryTypeConvert<ck::bhalf_t, float>
}
};
struct ConvInvscale
{
/// @brief Op to multiply convolution results by inverted scale factors
/// @param e Output after scaling
/// @param c Convolution result
/// @param d0 Input scale factor
/// @param d1 Weights scale factor
/// @param d2 Output scale factor
template <typename E, typename C, typename D0, typename D1, typename D2>
__host__ __device__ void
operator()(E& e, const C& c, const D0& d0, const D1& d1, const D2& d2) const;
template <>
__host__ __device__ void operator()<f8_t, float, float, float, float>(
f8_t& e, const float& c, const float& d0, const float& d1, const float& d2) const
{
e = type_convert<f8_t>(c / d0 / d1 / d2);
};
};
} // namespace element_wise
} // namespace tensor_operation
} // namespace ck