bigmul

big multiplication in C
git clone git://git.rr3.xyz/bigmul
Log | Files | Refs | README | LICENSE

README (556B)


      1 This is an implementation of arbitrary width multiplication in C. The main goal
      2 here was to learn what CPU hardware features might benefit arbitrary width
      3 multiplication (similar to how an ADDC (add with carry) instruction makes
      4 arbitrary width addition efficient). I think the main takeaway is that an
      5 FMAA (fused multiply-add-add) instruction would be useful, especially
      6 considering Dadda multipliers (and probably other hardware multiplier designs)
      7 can be extended with a relatively small amount of circuitry to support FMAA
      8 without increasing latency.