I'm working with the OMAP3530's C64x DSP and I'm investigating its multiply capabilities. From the data sheet, I see the following:
· Two Multipliers Support Four 16 x 16-Bit Multiplies (32-Bit Results) per ClockCycle or Eight 8 x 8-Bit Multiplies (16-Bit Results) per Clock Cycle
Is this performance achieved only by using the multiply instrinsics or does this not matter?
What performance do I get if I do 32x32-bit multiplies? Do I get one of those multiplies per cycle? And will my result be a 64-bit or 40-bit result?
Is the C64x considered a 16-bit or 32-bit processor?