Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!gem.mps.ohio-state.edu!uakari.primate.wisc.edu!aplcen!haven!uvaarpa!hudson!astsun9.astro.Virginia.EDU!gl8f
From: gl8f@astsun9.astro.Virginia.EDU (Greg Lindahl)
Newsgroups: comp.sys.atari.st
Subject: Re: What does the TT Buy me??
Message-ID: <2248@hudson.acc.virginia.edu>
Date: 10 Nov 89 19:43:36 GMT
References: <46bcb82f.14a1f@force.UUCP> <12430004@acf5.NYU.EDU>
Sender: news@hudson.acc.virginia.edu
Reply-To: gl8f@astsun9.astro.Virginia.EDU (Greg Lindahl)
Organization: Department of Astronomy, University of Virginia
Lines: 33

In article <12430004@acf5.NYU.EDU> mitsolid@acf5.NYU.EDU (Thanasis Mitsolides) writes:
>Let us see some facts. From BYTE magazines:
>
>		IIcx	IIci	ALR/25	SIA/32
>
>Matrix		16.2	10.5	 2.62	 2.10
>Sieve		31.4	19.8	14.06	11.02
>Sort		29.7	19.4	10.52	 8.26
>
>
>Mac IIcx, 16Mhz, 120ns Dram, No cach, No Burst fill mode.
>Mac IIci, 25Mhz, 80ns  Dram, No cash, Burst fill mode.
>
>ALR 386/25, 25Mhz, 80ns Dram, Cash, interleaved memory.
>SIA 386/32, 32Mhz, 80ns Dram, Cash, interleaved memory.

You left off a few details. OS for the 80x86 boxes? What fraction, if
any, of the IIci memory is running with burst fill mode? And was the
on-chip cache on the 68030 enabled?

From the speedup between the x and i, I would think that the IIci wasn't
using burst fill at all.

Finally, benchmarking is a black art. I prefer benchmarks which are
similar to the programs which I run, which are all large-memory-model
programs that are big. The Byte benchmarks aren't in this category.

The FACTS will arrive when you can run your *application* on a TT and
on other machines and compare the results.

------
Greg Lindahl
gl8f@virginia.edu                                             I'm not the NRA.