Path: utzoo!utgpu!news-server.csri.toronto.edu!mailrus!uwm.edu!zaphod.mps.ohio-state.edu!julius.cs.uiuc.edu!apple!snorkelwacker!bloom-beacon!eru!hagbard!sunic!news.funet.fi!funic!polaris.utu.fi!polaris!jita From: jita@polaris.utu.fi (Marko Katajisto) Newsgroups: comp.sys.amiga Subject: Re: Amiga competitiveness Message-ID: Date: 30 Sep 90 17:31:15 GMT References: <90271.190320UH2@psuvm.psu.edu> <420@caslon.cs.arizona.edu> <1091@ucsvc.ucs.unimelb.edu.au> Sender: jita@utu.fi (Marko Katajisto) Organization: University of Turku, Finland. Lines: 49 In-Reply-To: U3364521@ucsvc.ucs.unimelb.edu.au's message of 30 Sep 90 05:27:27 GMT In article <1091@ucsvc.ucs.unimelb.edu.au> U3364521@ucsvc.ucs.unimelb.edu.au (Lou Cavallo) writes: >CJ> In article <420@caslon.cs.arizona.edu>, cjeffery@cs.arizona.edu >CJ> (Clinton Jeffery) writes: > >CJ> Example #2: an Amiga's hard disk, even with FFS, is *ridiculously* slow >CJ> compared to the _identical_ hard disk running DOS or UNIX/386. This >CJ> is based on empirical observation doing heavy C compiles on a 2000 >CJ> with 1MB ram and a 40 Meg ST-251-1. Sorry. I wish it were untrue. I'd like to hear more about these empirical observations... First, how can you say that the limitation of the speed is harddisk/filesystem when comparing two cpu/compiler/controller combinations where nothing else than the disk is same? How you measure what's the weakest link, or is it just that you think 68000 at 7.x MHz should compete with x.x MHz 386 SX when system's harddisk is same....? >What was the controller on the DOS or UNIX system? >What was the controller on the Amiga system? Obviously DOS/UNIX system's controller is somekind of ST502/504 controller and I assume Amiga's controller is same type, otherwise there's no sense in this comparasion... >I have seen a Quantum 40 Meg equipped 386 DX apparently out perform (2 Mbits >per sec benchmark) quoted A2000 Quantum 40 Meg throughputs but the 386 had a >32 bit buss... {which qualitatively explains its approx 2x throughput over a >16 bit A2000 buss}. I'd really like to see that Quantum/386 combination myself. How can it be possible that you suck 2 MB/sec (I hope that 2 MBits is a typo) on any system from a harddisk that just can't perform that fast??? It's just that Amiga uses no caching with it's disk, DOS and UNIX are usually forced to utilize somekind of a cache, DOS because of those lousy ST-drives :) and UNIX for some other reasons. The bus width (In this case) has actually not some much to do with harddisk's speed because that's not the weakest links. BTW: dperf claims that Quatum 105S is roughly twice as fast when connected to Sun 3.x/80's SCSI-bus than when connected to Amiga's SCSI (HardFrame) bus, and that's absolute bs as we all know... ;) I can get about 800KB/sec out from my 105S (A2000b/A2620) and if I want to get significantly faster I must buy a Wren or something. So, anybody got a Wren IV connected to his Amiga and is willing to run diskperf of DiskSpeed...? >CJ> Clint Jeffery, U. of Arizona Dept. of Computer Science >CJ> cjeffery@cs.arizona.edu -or- {noao allegra}!arizona!cjeffery > >Yours truly, >Lou Cavallo. Jita - Marko Katajisto University of Turku Department of Computer Science Internet: jita@utu.fi Bitnet: jita@firien.bitnet Tel: 358 21 325910