Path: utzoo!attcan!uunet!husc6!bloom-beacon!gatech!udel!princeton!phoenix!pupthy2!lgy From: lgy@pupthy2.PRINCETON.EDU (Larry Yaffe) Newsgroups: comp.arch Subject: Re: integer multiplies (Was: architecture/implementation -- MIPS) Message-ID: <3006@phoenix.Princeton.EDU> Date: 24 May 88 15:44:35 GMT References: <2231@gumby.mips.COM> Sender: news@phoenix.Princeton.EDU Reply-To: lgy@pupthy2.PRINCETON.EDU (Larry Yaffe) Organization: Physics Dept, Princeton Univ Lines: 41 In article <2231@gumby.mips.COM> earl@mips.COM (Earl Killian) writes: [[ Much information about the R/3000 including: ]] >> Integer Multiply/Divide > >How is multiply is implemented [software, multiply step, hardware]? hardware >How many cycles to perform 32x32->32 multiply? 13 ^^ > >> Floating Point > >What are the floating point operation latency/issue/repeats? > > 32-bit 64-bit 80-bit > mul 4/ 1/ 4 5/ 1/ 5 n.a. ^^ >UUCP: {ames,decwrl,prls,pyramid}!mips!earl >USPS: MIPS Computer Systems, 930 Arques Ave, Sunnyvale CA, 94086 I'm curious and a little surprised about the relative speed of integer and floating point multiplies. Are integer multiplies substantially less frequent than floating point mults (in some selection of "real" programs) that integer multiply being three times slower than f.p. is not a significant bottleneck? What fraction of integer multiplies are not multiplications by small compile-time constants (in that same set of programs)? Since MIPS Co. is reputed to have studied real programs when considering design trade-off's, I'm hoping someone there can offer some insight. Also, despite having looked through the MIPS architecture book quite carefully, I can not find any mention of the latency of integer multiply or divide operations. Am I just getting sloppy, or is the information really missing? ------------------------------------------------------------------------ Laurence G. Yaffe Internet: lgy@pupthy.princeton.edu Department of Physics Bitnet: lgy@pucc Princeton University UUCP: ...!princeton!pupthy!lgy PO Box 708, Princeton NJ 08544