Newsgroups: comp.sys.apple2
Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!hobbes.physics.uiowa.edu!news.iastate.edu!ux1.cso.uiuc.edu!uxa.cso.uiuc.edu!jb10320
From: jb10320@uxa.cso.uiuc.edu (Jawaid Bazyar)
Subject: Re: ZIP GS... HOW FAST?
Message-ID: <1991Jun5.024649.2177@ux1.cso.uiuc.edu>
Sender: usenet@ux1.cso.uiuc.edu (News)
Organization: University of Illinois at Urbana
References: <9106050110.AA13274@apple.com>
Date: Wed, 5 Jun 1991 02:46:49 GMT
Lines: 53

ISSLTC@NUSVM.BITNET (Lim Thye Chean) writes:

>Hi. After reading so many replies on my original question 'Zip vs
>Transwarp... which is better', I will settle down for Zip. This is
>because all of you who cares to reply say Zip is better, that is 100%
>agreement. So I assume Transwarp GS is an inferior product (at price/
>performance ratio).

>But two things almost change my mind. The first is the report on that
>disgusting magazine A+/Incider. On the 4 reports they produce on a
>10 MHz Zip with 64K cache, Zip is only slightly (2 seconds in some
>case) faster than the 7MHz Transwarp GS with 8K cache. A similiar
>report is on A2-Central (Title: The need for speed), where Transwarp GS
>beats a 8MHz, 16K cache Zip GSX. So I really wonder, is Zip really
>a faster board? It seems that Transwarp GS performs much faster on
>the same MHz and same cache configuration.

  It depends on what the benchmarks are done on.  Personally, I wouldn't
trust an A+/InCider figure if my life depended on it.  Timing how long it
takes to boot GS/OS is NOT a valid way to determine how 'fast' your applications
will run.
  As Todd Whitesel has pointed out before, the Zip sports a 'latch-on-write'
feature that makes it much faster than the Transwarp when dealing with 
slowdowns (like I/O and screen).

>The second point of view is my friend's experience who original have
>a Transwarp GS, but later bought a Zip GSX (8 MHz, 16K cache). He
>upgrade it to 64K cache, but found that in majority of the applications,
>he only have about 1 second gain. But later I discovered that he
>has replace 16K 70ns cache with 64K 100ns cache, and I am not sure
>whether it will affect the speed. I always believe that if a processor
>(in this case Zip GS) work faster than the memory can handle, the
>computer will crash, so since my friend's GS is working properly, it
>should means that Zip will works OK under 100ns or 70ns. The same
>reason applies for buying 150ns, 100ns, 70ns memory chip for GS, it
>won't increase the speed. But this is just my assumption.

   Putting faster RAMs in a machine won't make the machine faster- the speed
of a machine is determined by a clock crystal.  As for your friend's Zip with
the 100ns rams, it's a wonder it works at all.

>So, any comment? Anybody have actually perform some bencemark test on
>1) Zip vs Zip on different configuration, 2) Zip vs Transwarp?

  For real calculation intensive stuff, the 10MHz-64 Zip beat the 7Mhz-8
transwarp by 30-50%. (this is from inCider, so take it for what it is- but
knowing cache design, there's no way the TransWarp truly could be faster.

--
Jawaid Bazyar               |  "Twenty seven faces- with their eyes turned to
Graduated!/Comp Engineering |    the sky. I have got a camera, and an airtight
bazyar@cs.uiuc.edu          |     alibi.."
   Apple II Forever!        |  I need a job... Be privileged to pay me! :-)