Path: utzoo!attcan!uunet!lll-winken!brutus.cs.uiuc.edu!unix.cis.pitt.edu!zaphod.mps.ohio-state.edu!samsung!emory!auc!rar From: rar@auc.UUCP (Rodney Ricks) Newsgroups: comp.sys.amiga Subject: Re: Amiga 3000 Summary: Some qualification of 68040 / SPARC speed comparisons. Message-ID: <32389@auc.UUCP> Date: 13 Mar 90 20:24:10 GMT References: <13479@baldrick.udel.EDU> <9884@batcomputer.tn.cornell.edu> <6857@cps3xx.UUCP> <3944@nmtsun.nmt.edu> <6360@sbcs.sunysb.edu> <3951@nmtsun.nmt.edu> <6398@sbcs.sunysb.edu> Reply-To: rar@auc.UUCP (Rodney Ricks) Organization: Atlanta University Center, Atlanta, Ga. Lines: 78 I don't have the BYTE article here with me that gives the information, so I'm going by memory. In article <6398@sbcs.sunysb.edu> root@sbcs.sunysb.edu (Systems Staff) writes: >In article <3951@nmtsun.nmt.edu> dksnsr@nmtsun.nmt.edu (Dr. Mosh) writes: >>In article <6360@sbcs.sunysb.edu> root@sbcs.sunysb.edu (Systems Staff) writes: >> >>> Unqualified statements such as yours are meaningless. Which, >>> exactly, 68040 system are you referring to? Are you talking compute, >> >>Meaningless?? I think you are missing the point here... The original >>question was whether to build an Amiga SPARC or 68040, based on pure >>processing performance, the 68040 IS Faster... > ^^^^^^^^^^^^^^^ > > Faster than *what*? A Sparc clocked at equivalent rate? From what I remember of the article, the 68040 is faster than a SPARC of the SAME clock rate (25 MHz). >>And you don't think Motorola could meet their own "reported" performance >>in the FINAL stages of release? Have they ever NOT met their reported >>figures?? > > Yes, they can meet their "reported" performance if they only > state it to be 20 mips. Without qualifying how they measure > a MIP then it could be 20 mips of NOP or 20 MIPS of FP in the > extremes. 20 MIPS of a "typical instruction mix" (whatever that means...). Using NOP's as the "instructions" is an Intel trick (according to a developer on the net). >> ...if in fact it does run as "reported", it would >>in most aspects beat out a current SPARC... That should, of course, be "a current SPARC with the same clock rate". > Oh, you mean faster than the Soulborne machine? Seriously, you > haven't been around that long if you think every statement a company > makes, especially about performance, is always met. Not being a hardware expert, I can only go by what I've heard. From what I've heard, Motorola doesn't usually go around blowing up the performance claims about their CPU's. If someone out there has any valid reason to believe otherwise, I invite them to share it with us. Of course, on usenet, such an invitation is totally unnecessary! > Until there > is a 68040 that can be SPECMark'ed by someone other than Motorola, > then the '040 clocks 0 mips in my book. Amazingly, companies don't ALWAYS lie. Some of them are reasonably truthful, especially about making performance claims that, a couple of months later, people can verify as being true or untrue. Anyway, going back to what started all of this, even if the 68040 is slower than the SPARC chips (of the same clock rate), the Amiga should still use a 68040 chip, for compatibility purposes. Unless the SPARC chip is much, much faster than the 68040, the SPARC emulating a 680x0 would be much slower than the 68040 running 680x0 software. Along the lines of what another person said, *IF* the Amiga 3000 is released late this year, if it is based on a 68030, it will be VERY anticlimatic. >>> Rick Spanbauer >>> State U of NY/Stony Brook Rodney Ricks -- "We may have come over here in different ships, but we're all in the same boat now." -- Jesse Jackson // \\ // Rodney Ricks, Morehouse College \/