Register-based complex number processing

A data processing system stores vectors of complex numbers in vector registers. The complex numbers are stored with the real and imaginary parts of the same number in adjacent vector entries. A signed doubling multiply add long bottom top instruction multiplies the lower of a pair of vector entries from a first source register with the upper entry in the corresponding pair in a second source register. The result is then added to the corresponding entry in a destination entry, the entries of which are twice as long as those in the source registers. Further instructions allow other combinations of the entries to be multiplied and allow the result to be subtracted from the destination register. The instruction may specify a transposition operation to be carried out on one of the vectors before the multiplication. The result may be stored in the destination register entries rather than being added.