max_propagate_nan Class — pytorch Architecture
Architecture documentation for the max_propagate_nan class in SharedReduceOps.h from the pytorch codebase.
Entity Profile
Source Code
aten/src/ATen/native/SharedReduceOps.h lines 25–35
template <typename scalar_t>
inline C10_DEVICE scalar_t max_propagate_nan(scalar_t a, scalar_t b) {
#if defined(__HIPCC__)
// TODO: remove this special case for HIP when issue is fixed:
// https://github.com/ROCm/hip/issues/2209
scalar_t max = at::_isnan(a) ? a : (at::_isnan(b) ? b : std::max(a, b));
#else
scalar_t max = at::_isnan(b) ? b : std::max(a, b);
#endif
return max;
}
Source
Analyze Your Own Codebase
Get architecture documentation, dependency graphs, and domain analysis for your codebase in minutes.
Try Supermodel Free