# f32v4tof16 $aDst0:Dst0+1 $aSrc0:Src0+3

*Single-precision* 4-element vector to *f16* vector