* `f16v4cmac $aSrc0:Src0+1, $aSrc1:Src1+1`

*Half-precision* floating-point vector element-wise multiply with
*single-precision* 2x2 lateral sum and accumulate.