[comp.sys.ibm.pc] Extended Precision 2's Complement Multiply?

J0S@PSUVM.BITNET (07/27/89)

I need to preform a complex times complex 2's complement integer multiply.
The machine architecture has both a 16-bit signed multiply and a 16-bit
unsigned multiply.  The real and imaginary parts of the multiplicand are
both 32-bit 2's complement integers.  The real and imaginary parts of the
multiplier are both 16-bit 2's complement integers.  The complex magnitude
of the multiplier is unity.  The real and imaginary parts of the complex
product are to be 32-bit 2's complement integers containing the most
significant bits of the intermediate products.

Do you have an efficient algorithm for this process or could point me
towards a reference?

                       Thanks, Jack W. Sharer, JWS@PSUARLC.BITNET