dst_n Class — pytorch Architecture
Architecture documentation for the dst_n class in vec_convert.h from the pytorch codebase.
Entity Profile
Source Code
aten/src/ATen/cpu/vec/vec_convert.h lines 9–28
template <
typename dst_t,
int dst_n,
typename src_t,
int src_n,
typename Enabled = void>
struct VecConvert {
static inline VectorizedN<dst_t, dst_n> apply(
const VectorizedN<src_t, src_n>& src) {
constexpr int count = std::min(
VectorizedN<src_t, src_n>::size(), VectorizedN<dst_t, dst_n>::size());
__at_align__ src_t src_buf[VectorizedN<src_t, src_n>::size()];
src.store(src_buf);
__at_align__ dst_t dst_buf[VectorizedN<dst_t, dst_n>::size()];
for (int i = 0; i < count; i++) {
dst_buf[i] = static_cast<dst_t>(src_buf[i]);
}
return VectorizedN<dst_t, dst_n>::loadu(dst_buf, count);
}
};
Source
Analyze Your Own Codebase
Get architecture documentation, dependency graphs, and domain analysis for your codebase in minutes.
Try Supermodel Free