home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.sys.super
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!sol.ctr.columbia.edu!eff!ssd.intel.com!hays
- From: hays@ssd.intel.com (Kirk Hays)
- Subject: Re: World's Most Powerful Computing Sites
- Message-ID: <C19r20.2wx@SSD.intel.com>
- Sender: usenet@SSD.intel.com
- Nntp-Posting-Host: taos
- Organization: Intel Supercomputer Systems Division
- References: <1993Jan20.232809.29241@nas.nasa.gov> <1993Jan21.165159.10149@meiko.com> <1993Jan22.015827.26653@nas.nasa.gov>
- Date: Fri, 22 Jan 1993 18:51:35 GMT
- Lines: 24
-
- In article <1993Jan22.015827.26653@nas.nasa.gov>, fineberg@wilbur.nas.nasa.gov (Samuel A. Fineberg) writes:
- |> In article <1993Jan21.165159.10149@meiko.com>, richard@meiko.com (Richard Cownie) writes:
- |> |> fineberg@wilbur.nas.nasa.gov (Samuel A. Fineberg) writes:
-
- |> |> : Its certainly no less realistic than those for the i860.
- |> |>
- |> |> I have to disagree with you there. I know of *some* applications where
- |> |> the i860 can achieve a good fraction of claimed peak speed, e.g. on
- |> |> a double-precision matrix multiply you can do over 35MFLOPS, against
- |> |> a peak rate claimed as 40MFLOPS (or sometimes 60MFLOPS, because you can do
- |> |> 2 adds for each multiply). In any case, it's well over 50% of peak.
- |>
- |> I don't know too many people that write assembly code, and that is what you
- |> need to do to get 35 MFLOPs.
-
- False. Using the PGI compilers on C and FORTRAN source code, selected codes can
- achieve > 35 MFLOPS.
-
- As you point out, most codes get less.
-
- --
- Kirk Hays - NRA Life, seventh generation.
- "The only thing necessary for the triumph of evil is for good men to
- do nothing." -- Edmund Burke (1729-1797)
-