[−][src]Function safe_arch::move_mask_m256i
#[must_use]pub fn move_mask_m256i(a: m256i) -> i32
This is supported with target feature
avx2
only.Create an i32
mask of each sign bit in the i8
lanes.
let a = m256i::from([ 0_i8, 11, 2, -13, 4, 15, 6, -17, -8, 19, -20, 21, 22, -23, 24, 127, 0, -1, 3, 4, 5, 1, -2, -4, -8, 12, 13, 14, 29, 30, -31, 32, ]); assert_eq!(0b01000001110000100010010110001000, move_mask_m256i(a));
- Intrinsic:
_mm256_movemask_epi8
- Assembly:
vpmovmskb r32, ymm