pub unsafe fn _mm256_set_ps(
    a: f32,
    b: f32,
    c: f32,
    d: f32,
    e: f32,
    f: f32,
    g: f32,
    h: f32
) -> __m256
This is supported on x86-64 and target feature avx only.
Expand description

Sets packed single-precision (32-bit) floating-point elements in returned vector with the supplied values.

Intel’s documentation