NetNews Usenet Archive 1992 #16/NN_1992_16.iso/spool/comp/arch/8194

Internet Message Format | 1992-07-22 | 3.0 KB

Path: sparky!uunet!iWarp.intel.com|inews!mkahn
From: [e-mail] (Mitch Kahn)
Newsgroups: comp.arch
Subject: Re: Request Info Re Bus Trends & Intel 486s
Message-ID: <[e-mail]>
Date: 22 Jul 92 16:10:03 GMT
References: <[e-mail]>
Sender: [e-mail]
Organization: /eng/eng3/mkahn/.organization
Lines: 59
In-reply-to: [e-mail]'s message of 15 Jul 92 09:16:13 GMT

[e-mail] (Trilogy Business Systems) writes:

 Summary: I need some info re alternative Bus designs & Diffs Between 486s
] Keywords: Intel 486 Bus
] Posted: Wed Jul 15 04:16:13 1992
] Nntp-Posting-Host: extro.ucc.su.oz.au
]
] I am seeking some information concerning 2 different subjects.
]
] The first is cocerning 486 Chips.
]
SImply stated:

A 486DX50 is faster than a 486DX2-50. The DX50 has a 50MHz bus and 50MHz core.
The DX250 has a 50MHz core and a 25MHz bus. SOunds like the salesman is
either ignorant or out to screw you. Actual performance numbers were posted
by the x86 guys a few weeks ago in comp.sys.intel.

]
] Secondly are the 486SX chips just 486DX chips that fail the co-processor
] certification (and 487's that same that fail the CPU test), in the same
] way that the old single sided disks (remember them ??) were double sided
] disks that failed certification on one side ?


The 486SX used to have a dead (or disabled FPU). They now have no FPU at all.
When the production ramp warranted, the FPU was removed to save on die cost.
This is an all pervasive technique in the semicon industry.

]
] Thirdly what in user terms are the major advantages of the P5 over the current
] 486 chip, will the 486 go the way of the 386, or will they become seperate
] product lines ?


The P5 will have very significant performance advantages over the 486 family.
I would expect a similar 486->P5 migration as we are seeing now with the
386->486. But it won't happen until about 1994. (My opinion)

]
] My second major question concerns bus structures. It seems to me that the EISA
] bus is too slow for any of the major changes happening to desktop computing,
] all the companies I know of making multi-processor machines are using a
] proprietry bus due to the relatively slow speed of EISA and Microchannel.
] HP seem to have implemented a multi-proceesor machine over it PA bus, but
] I don't see this as a major desktop bus architecture in the years to come.
] DEC have Turbo-channel, but I heard that even that has a limited life. I
] have heard that IEEE are in the process of defining something called FutureBus
] but at the rate that concensus seems to take, this will be

There are two local bus standards currently in the wings. VESA and PCI.
PCI (from Intel) has not been formally announced but has been in the
press quite a bit lately (check PC Magazine this month).

]
--
|  Mitch Kahn              | "I hope you know that this will go down on |
|  [e-mail]  |  your permanent record..."                 |
*************************MY OWN OPINIONS*********************************