Path: utzoo!attcan!uunet!aplcen!uakari.primate.wisc.edu!zaphod.mps.ohio-state.edu!mips!dalek!keith
From: keith@mips.COM (Keith Garrett)
Newsgroups: comp.arch
Subject: Re: int x int -> long for * (or is it 32x32->64)
Keywords: arithmetic,arbitrary precision,benchmark,modular arithmetic
Message-ID: <41501@mips.mips.COM>
Date: 14 Sep 90 16:34:29 GMT
References: <3984@bingvaxu.cc.binghamton.edu> <41425@mips.mips.COM> <353@kaos.MATH.UCLA.EDU> <2118@charon.cwi.nl>
Sender: news@mips.COM
Reply-To: keith@mips.COM (Keith Garrett)
Organization: MIPS Computer Systems, Inc.
Lines: 18

In article <2118@charon.cwi.nl> dik@cwi.nl (Dik T. Winter) writes:
>[multipy comparisions deleted]
>MIPS:	Multiply is 10 cycles (I found this somewhere in Kanes book, but
>	can not find it now).  Divide for small numbers (i.e. 32/32) is a
>	single instruction, for which I cannot find the timing now.  Larger
>	numbers require more cycles.  My code uses 9 cycles per bit, which
>	can be reduced to 6 if only the remainder is required.  Using hardware
>	divide might give some benefits here.  Note that hardware multiply and
>	hardware divide are free running, i.e. other operations can be
>	performed, but the benefit would be very limited in this case.
integer multiply is actually 12 cycles (32*32->64)...integer divide is 35
cycles (32/32->quot,rem) there are addition cycles required to move the
result from the special mult/div result registers to gp registers (MFHI/MFLO).
>[...]
-- 
Keith Garrett        "This is *MY* opinion, OBVIOUSLY"
      Mips Computer Systems, 930 Arques Ave, Sunnyvale, Ca. 94086
      (408) 524-8110     keith@mips.com  or  {ames,decwrl,prls}!mips!keith