pub fn f32x4_sub(a: v128, b: v128) -> v128
simd128
Subtracts pairwise lanes of two 128-bit vectors interpreted as four 32-bit floating point numbers.