Function _mm512_castps256_ps512

#[target_feature(enable = "avx512f")]
pub const fn _mm512_castps256_ps512(a: __m256) -> __m512

Cast vector of type __m256 to type __m512; the upper 256 bits of the result are indeterminate.

In the Intel documentation, the upper bits are declared to be "undefined". This is not equivalent to mem::MaybeUninit; instead, these bits are non-deterministically set to some valid value. In practice, this is typically equivalent to mem::zeroed.

This intrinsic is only used for compilation and does not generate any instructions, thus it has zero latency.

Intel's documentation