home *** CD-ROM | disk | FTP | other *** search
- Path: arlut.utexas.edu!usenet
- From: Bernie Lofaso <lofaso@arlut.utexas.edu>
- Newsgroups: comp.sys.amiga.misc
- Subject: Re: Speed: 68040 vs. 68060
- Date: Wed, 06 Mar 1996 08:55:56 -0600
- Organization: Applied Research Labs
- Message-ID: <313DA77C.22FE@arlut.utexas.edu>
- References: <371.6633T989T2700@horus.co.jyu.fi> <1195.6634T1430T809@teclink.net> <4hfm08$4s@tkhut.sojourn.com>
- NNTP-Posting-Host: gandalf.arlut.utexas.edu
- Mime-Version: 1.0
- Content-Type: text/plain; charset=us-ascii
- Content-Transfer-Encoding: 7bit
- X-Mailer: Mozilla 2.0 (X11; I; HP-UX A.09.05 9000/730)
-
- Matt Harrell wrote:
- >
- > rad (rad@teclink.net) wrote:
- > : Aki Laukkanen <daeron@horus.co.jyu.fi> wrote:
- > : >>According to the Motorolla press release, which I could post here if
- > : >>anyone really gives a hoot, states the the 50MHz 060 is capable of
- > : >>maxing out at 250 MIPS, and averages 80 VAX MIPS.
- >
- > I am curious, though, as to what "VAX MIPS" is. I wondered about that
- > when I first read the press release.
-
- Quoting MIPS for a CPU is a fairly useless measurement. The reason is
- that
- different architectures will require a different number of instructions
- to perform the same task, hence comparing MIPS between different
- architectures tells you nothing. A RISC CPU will execute more
- instructions
- per second that a comparably performing CISC CPU, but will also require
- more instructions to be executed for a given task. Hence, many years
- ago,
- it was decided to select a single popular platform for comparison. The
- VAX (don't recall the exact model) was selected and dubbed a 1 MIP
- machine
- (probably because that's what it executed - 1 million VAX instructions
- per second). Hence a machine that ran twice as fast as this was a 2 MIP
- machine, even if it execute far more or far less native instructions per
- second.
-
- In general, MIPS numbers are still fairly useless, even VAX MIPS. The
- industry has moved to quoting SPEC numbers (specINT, specFLOAT, etc.)
- which compare performance for specific tasks or task mixes.
-
- Bernie Lofaso
- Applied Research Labs
-