home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!sol.ctr.columbia.edu!emory!ogicse!das-news.harvard.edu!cantaloupe.srv.cs.cmu.edu!netnews-2.srv.cs.cmu.edu!moss
- From: moss@cs.cmu.edu (Eliot Moss)
- Newsgroups: comp.benchmarks
- Subject: Re: DEC ALPHA Performance Claims
- Message-ID: <MOSS.92Nov17085951@CRAFTY.cs.cmu.edu>
- Date: 17 Nov 92 13:59:51 GMT
- Article-I.D.: CRAFTY.MOSS.92Nov17085951
- References: <BxH7s7.5Cv@inews.Intel.COM> <4248@bcstec.ca.boeing.com>
- Sender: news@cs.cmu.edu (Usenet News System)
- Reply-To: moss@cs.cmu.edu
- Followup-To: comp.benchmarks
- Organization: Dept of Comp and Info Sci, Univ of Mass (Amherst)
- Lines: 28
- In-Reply-To: silverm@bcstec.ca.boeing.com's message of 16 Nov 92 06:36:50 GMT
- Nntp-Posting-Host: crafty.fox.cs.cmu.edu
-
- >>>>> On 16 Nov 92 06:36:50 GMT, silverm@bcstec.ca.boeing.com (Jeff Silverman)
- >>>>> said:
-
- Jeff> There is a lot of discussion on this thread about hardware vs. software
- Jeff> speed up. I suspect that the question is moot. One of the features,
- Jeff> attributes, whatever of the Alpha is that the compiler is supposed to
- Jeff> sort instructions so that the hardware pipeline works optimally. This
- Jeff> means that, for all intents and purposes, the compiler and the CPU are
- Jeff> closely coupled in a new and unique way. It also has some rather
- Jeff> interesting side effects:
-
- I disagree. This is not new to the Alpha, but came into its own with RISC
- chips, and was really there in some ways all along. Instruction scheduling and
- register allocation (and good compilers in general) are very important in
- obtaining good performance from modern processors. What may be new in the last
- ten years or so is the extent to which the hardware designers are now relying
- on good software to use the hardware to best advantage, rather than simply
- looking at static and dynamic instruction usage distributions to see what
- should be sped up more, etc.
- --
-
- J. Eliot B. Moss, Associate Professor Visiting Associate Professor
- Department of Computer Science School of Computer Science
- Lederle Graduate Research Center Carnegie Mellon University
- University of Massachusetts 5000 Forbes Avenue
- Amherst, MA 01003 Pittsburgh, PA 15213-3891
- (413) 545-4206, 545-1249 (fax) (412) 268-6767, 681-5739 (fax)
- Moss@cs.umass.edu Moss@cs.cmu.edu
-