* `f16v4sub $aDst0:Dst0+1, $aSrc0:Src0+1, $aSrc1:Src1+1`
* `f16v4sub $aDst0:Dst0+1, $aSrc0:BL, $aSrc1:Src1+1`
* `f16v4sub $aDst0:Dst0+1, $aSrc0:BU, $aSrc1:Src1+1`

*Half-precision* floating-point 4-element vector subtraction