NetNews Usenet Archive 1992 #19/NN_1992_19.iso/spool/comp/sys/ibm/pc/hardware/23103

Internet Message Format | 1992-08-31 | 3.5 KB

Path: sparky!uunet!cs.utexas.edu!uwm.edu!ogicse!news.u.washington.edu!gibdo!tvp
From: [e-mail]
Newsgroups: comp.sys.ibm.pc.hardware
Subject: Re: 486 DX vs DX2: What's the difference ?
Message-ID: <[e-mail]>
Date: 1 Sep 92 05:24:48 GMT
Article-I.D.: u.1992Sep1.052448.6162
References: <[e-mail]> <[e-mail]> <[e-mail]>
Sender: [e-mail] (USENET News System)
Organization: clearer than blir
Lines: 61

In article <[e-mail]> [e-mail] writes:
><[e-mail]> writes:
>
>>  So, what is the performance ratio?  Rather depends on what you do.  Most
>>PC systems are I/O bound anyways... no loss there.  Some applications are
>>CPU bound...  no loss there.  So, unless your application is memory bound,
>>you might never know whether it was a DX or DX2!  Benchmarks run anywhere
>>from 50% to 90% performance ratio of a DX2 to a DX...
>
>So, which is better: a 50Mhz DX2 or a 33Mhz DX?  (Yeah, I know, it depends.)
>And this is NOT an academic question, these kinds of systems are showing up
>in the discount stores for roughly the same $$.  What you are saying would
>seem to say that most of the time the 50Mhz DX2 will beat the 33Mhz DX,
>right?
>---
>Michael Nolan, [e-mail]      "Freedom of the press is still alive in
>Tailored Software Services, Inc.    America, at the U. S. Mint"  (Gallagher)
>Lincoln, Nebraska [phone]

This is what PC magazine says in its review of 63 DX2/50 systems
in a box entitled "DX2 vs. OverDrive: Do You Want Your Speed Boost
Now or Later?"

   "Does the OverDrive chip give true 50-MHz 486DX performance?
    No, but it's close. Intel claims 1.8 times the performance
    of a 25-MHz 486. This coincides with the performance of
    factory installed DX2s seen in this roundup, which performed
    at 96 [sic: they meant 92] percent the speed of a true 50-MHz
    486DX on PC Magazine Labs' processor benchmark test."

This 92 percent figure roughly matched what I figured for my own
486DX2/50 from GW2000. The performance turns out to be BETTER
than any of us were expecting and luckily the prices won't go
back *up* because of it. The trick seems to be in the cache size.
Under "What the Numbers Mean: 486 DX2 PCs" they state:

   "Considering only those DX2s that feature at least a 128K
    cache we found that the average DX2/50 performed at 96
    percent of the speed of the average DX/50 (505 versus 525
    operations per second)...

    The average memory performance for all the DX2s was 6,314
    KBps (compared with 8,017 KBps for the DX/50s)--78% of the
    performance of true 50-MHz systems....

    The reasonable conclusion: Under most applications, PCs
    utilizing the clock-doubler technology should perform
    nearly on the same level as a true 50-MHz 486 when the
    DX2 chip used with an appropriately designed external
    cache. Memory intensive work will suffer, since memory
    accesses leave that fast 50-MHz world inside the CPU
    and venture onto the memory bus at 25MHz."

So there you have it. Hope this helps. (BTW: Occidental's 486DX2/50
(with 8MB RAM, 240MB 12ms IDE hard drive, one (of eight) local bus
slots, and SAME DAY SHIPPING) looks pretty good at $2495 depending on
the video card they're using (The ad doesn't say what it is.).

-- Tad Perry   Internet:    [e-mail]
               CompuServe:  70402,3020
               NIFTY-Serve: GBG01266