Forums before death by AOL, social media and spammers... "We can't have nice things"
|    comp.lang.asm.x86    |    Ahh, the lost art of x86 assembly    |    4,675 messages    |
[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]
|    Message 3,369 of 4,675    |
|    Terje Mathisen to wolfgang kern    |
|    Re: Reciprocal MUL LUT    |
|    25 Apr 18 09:46:50    |
      From: terje.mathisen@nospicedham.tmsw.no              wolfgang kern wrote:       >       > Terje asked:       >       >> Interesting ideas, I've got to ask though, what is the need/intention       >> here?       >       >> I.e. why do you need the prime reciprocals, what do you need them for,       >       > speed up my 512 bit calculator for 1/P 1/N with minor rounding issues.       > ["black peep hole watchers" and a few "metoo scientists" asked for]              Again, what is it used for?              1/P == reciprocal prime, 1/N any reciprocal which can then be syntesized       by first factoring N and then combining the prime reciprocals?              >       >> why just 512 of them,       >       > not limited to 512 primes, but I want 512 bits precise 1/N.              Anything special with 512?              What is the final requirement?              >       >> and is size more or less important than speed of access?       >       > I search for both but see bloat as speed-eater (load-time&cache-burst).       > __              OK              Terje              --       - |
[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]
(c) 1994, bbs@darkrealms.ca