Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sun-barr!ccut!wnoc-tyo-news!toumon!wucc!ytsuji
From: ytsuji@wucc.waseda.ac.jp (Y.Tsuji)
Newsgroups: comp.sys.atari.st
Subject: Re: Atari Mortis
Summary: real performance of a powerful CPU
Keywords: history, cheerleading, admonitions
Message-ID: <5754@wucc.waseda.ac.jp>
Date: 29 May 91 01:20:24 GMT
References: <4528@bnr-rsc.UUCP> <1991May19.035413.14005@chinet.chi.il.us> <1991May28.180143.4644@colorado.edu>
Organization: The Centre for Informatics, WASEDA Univ.
Lines: 15

It is really off the point talking about what is the _maximum_ MIPs of a 030
machine. TT is likely to be less than 2.0 MIPs. Even if the CPU got a vast
cache memory, the access time to the actual DRAM is often the decisive factor.
Good CPU boards usually have 80-ns-4-MEGAbit DRAMs and when successive 
addresses are accessed the Column Address Strobe is repeated, reducing the
access time dramatically. If these modern techniques are missing, the
instructions already on the cache memory may be executed very fast, but the
overall performance will be deadly slow.
I once used a 020 machine at 10 MHz but it was actually much slower than our
ST because it had to be swapping memory to/from disks as it was a virtual
memory machine!
The actual performance can be measured only by using a typical application
program on them: I think my ST at 8 MHz is fast enough for my need.
Cheers,
Tsuji