# f16v4sub $aDst0:Dst0+1 $aSrc0:Src0+1 $aSrc1:Src1+1 (and 2 more)

*Half-precision* vector subtraction