Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

386/40 or 486sx/20?????????

92 views
Skip to first unread message

ba...@hls.com

unread,
Aug 5, 1991, 2:51:34 PM8/5/91
to
i'm currently looking at either a 386/40 or a 486sx/20. Anyone out there
with info on either chips? The cost difference is about $300, so is worth
the price to have a 486? Are problems with the 386/40 or the 486sx?
Any opinion would be greatly appreciated! Thanks in advance!

Roger L.
Hughes Lans Systems
Mountain View, CA

`Underneath your clothes you are naked!'

kai...@maple.circa.ufl.edu

unread,
Aug 5, 1991, 6:33:53 PM8/5/91
to
You might want to take a look at this month's PC Computing. It has an
article describing the new 486SX chips. I believe the title is something
like "New 486SX gives little reason to switch.." Check it out...

David W Hansquine

unread,
Aug 7, 1991, 4:16:52 PM8/7/91
to
In article <1991Aug5....@hls.com> ba...@hls.com writes:
>i'm currently looking at either a 386/40 or a 486sx/20. Anyone out there
>with info on either chips? The cost difference is about $300, so is worth
>the price to have a 486? Are problems with the 386/40 or the 486sx?
>Any opinion would be greatly appreciated! Thanks in advance!

I'm not too sure about the 486sx, (ie. I'm assuming that's the one
without the math coprocessor), but since it has a 16 bit data bus
rather than a 32 bit data bus, its probably slower than I think.
Anyway, generally for a 386 and 486 at the same speed, the 486 will
run at least twice as fast, since most of the instructions are
optimized to run in a single clock-cycle, while on the 386, they run
in 2 or 3. Those which take more clock cycles are either equivalent,
or longer by a few clock cycles on the 386. The 486 has a few extra
instructions built-in. And you're looking at about the same speed.
You'll either be getting a very fast 386, or a superslow 486. So you
gotta decide what you'd prefer.

SuperDave

-------------------------------------------------------------------
The opinions expressed are not mine, nor anyone's.

Clinton Jeffery

unread,
Aug 7, 1991, 7:07:19 PM8/7/91
to
From article <50...@bnr-rsc.UUCP> hans...@chekov.UU.NET (David W Hansquine ):

> In article <1991Aug5....@hls.com> ba...@hls.com writes:
>>i'm currently looking at either a 386/40 or a 486sx/20. Anyone out there
>>with info on either chips?
> I'm not too sure about the 486sx, (ie. I'm assuming that's the one
> without the math coprocessor), but since it has a 16 bit data bus
> rather than a 32 bit data bus, its probably slower than I think.
> ["correct" 486-techno hype deleted]

Why ask magazine-level questions when you can read the managazine just like
we did? And why answer a magazine question when you can't remember the
article you read anymore?? The 486dx is a real 486 that didn't make it
through 486 boot-camp, but still fondly remembers its 486 chip-school
training, except the trig, which was forcibly removed by sales-droids,
and was too hard anyway. No math coprocessor, yes a full 32 (THIRTY-TWO)
bit data path. Which is better depends on whether you are DOS (386/40)
or UNIX (toss-up, since 486* machines do better with 32-bit instructions).
Of course, all this is just something I infer from what I read in some
magazine, so get out your salt-shakers on this one.

hans li

unread,
Aug 7, 1991, 8:15:03 PM8/7/91
to
Actually a 486SX is a complete 486, just with the math co disabled. It is
still on the chip, but it's turned off. There is NO other difference between
the Sx and the real thing. They are both 32-bit chips. That's why you can
"trade in" your 486SX chip for a 486 and have a complete machine. That's how
the upgrades work with the 486SX chip. The only disadvantage with the SX is
that at present, only 20 MHz chips exist, as compared to the 25-50 MhZ of the
real 486.
1

Todd J Derr

unread,
Aug 8, 1991, 11:32:18 AM8/8/91
to

Well.......

From what i've read, the pinouts on the 486sx are slightly different
than on the 486dx...

Undoubtedly a ploy for Intel to then catch you in the next scam.... The
487sx! The 487sx is actually just a 486dx... plugs into the
'coprocessor' slot on the 486sx boards and actually DISABLES the entire
486sx processor...

If I'm wrong about this, let me know....


But.... it seems like a typical Intel SCAM...


Todd J. Derr
University of Pittsburgh...

Wolf Witt

unread,
Aug 8, 1991, 4:11:37 PM8/8/91
to
In article <162...@unix.cis.pitt.edu> tjd...@unix.cis.pitt.edu

(Todd J Derr) writes:
>From what i've read, the pinouts on the 486sx are slightly different
>than on the 486dx...

Yup, the pin-outs of the 486DX, 486SX and 487SX are all slightly different,
so the chips are not plug compatible.

>Undoubtedly a ploy for Intel to then catch you in the next scam.... The
>487sx! The 487sx is actually just a 486dx... plugs into the
>'coprocessor' slot on the 486sx boards and actually DISABLES the entire
>486sx processor...

Again, this is true. Placing a 487SX into a 486SX machine will in fact
turn off the 486SX. The 487SX does everything. However, note that the
socket is not a *coprocessor* socket. It's supposed to be called "the
performance upgrade socket." (Yeah, yeah, I know -- we're pushing it.)
The idea is that at some later time, there might be something else to
plug into it than just a 487SX.

>If I'm wrong about this, let me know....
>
>But.... it seems like a typical Intel SCAM...

Well, since I work for Intel, I better not comment on whether I think
it's a scam. Although, I do have my opinion about it...

Wolf

=> Wolf Witt, Design Engineer <=> You say there is a bug in this chip? <=
=> i860 Focus Group, Intel Corp. <=> No, no, you don't understand... <=
=> ww...@smdvx1.intel.com <=> This is a feature! <=
=> Intel knows nothing about what I say or do, but they pay me anyway. <=

Ed Moore

unread,
Aug 8, 1991, 1:10:09 PM8/8/91
to
>i'm currently looking at either a 386/40 or a 486sx/20. Anyone out there
>with info on either chips? The cost difference is about $300, so is worth
>the price to have a 486?

This is the subject of the cover story in June 1991 Byte magazine.

Phil Ngai

unread,
Aug 8, 1991, 2:24:23 PM8/8/91
to
hans...@chekov.UU.NET (David W Hansquine ) writes:
>I'm not too sure about the 486sx, (ie. I'm assuming that's the one
>without the math coprocessor), but since it has a 16 bit data bus
>rather than a 32 bit data bus, its probably slower than I think.

Sorry, SuperDave, you don't know what you're talking about.
The 486sx has a full 32-bit bus, what it's missing is the
math coprocessor only. (and it runs at 20 MHz)

For the original question, one important factor is if you
need a math coprocessor and what it will cost. At one point,
Intel was asking about $800 for the 487sx.

--
A place for everything and everything in its place.

Danny Low

unread,
Aug 8, 1991, 7:26:46 PM8/8/91
to
>(David W Hansquine )
>I'm not too sure about the 486sx, (ie. I'm assuming that's the one
>without the math coprocessor), but since it has a 16 bit data bus
>rather than a 32 bit data bus, its probably slower than I think.

The 486SX is a 32 bit machine. You are confusing it with the 386SX.

Danny Low
"Question Authority and the Authorities will question You"
Valley of Hearts Delight, Silicon Valley
HP NPCD dl...@pollux.svale.hp.com

Nur Iskandar Taib

unread,
Aug 13, 1991, 3:03:25 PM8/13/91
to
In article <162...@unix.cis.pitt.edu| tjd...@unix.cis.pitt.edu (Todd J Derr) writes:


You're completely right, of course...

However several makers have come out with a
compromise.. you get what is essentially a
backplane with a plug-in CPU card (though in
most cases the "card" takes up a sizable cor-
ner of the motherboard instead of a vertical
slot. Some examples are ALR (I think) and
IBM's models 90 and 95.

So you unplug the 486SX card and plug in a
486DX card. Problem is these upgrades often
cost more than the $700 or so that a 487SX
costs... the difference is that the clone
makers get the money instead of Intel. And
when the motherboard manufacturer goes under
a couple years later.. 9-)


--
-------------------------------------------------------------------------------
Iskandar Taib | The only thing worse than Peach ala
Internet: NT...@AQUA.UCS.INDIANA.EDU | Frog is Frog ala Peach
Bitnet: NTAIB@IUBACS !
-------------------------------------------------------------------------------

Russ Poffenberger

unread,
Aug 14, 1991, 2:03:46 PM8/14/91
to
In article <1991Aug13.1...@bronze.ucs.indiana.edu> nt...@silver.ucs.indiana.edu (Nur Iskandar Taib) writes:
>In article <162...@unix.cis.pitt.edu| tjd...@unix.cis.pitt.edu (Todd J Derr) writes:
>|In article <li.681...@mrcnext.cso.uiuc.edu| l...@mrcnext.cso.uiuc.edu (hans li) writes:
>||Actually a 486SX is a complete 486, just with the math co disabled. It is
>||still on the chip, but it's turned off. There is NO other difference between
>||the Sx and the real thing. They are both 32-bit chips. That's why you can
>||"trade in" your 486SX chip for a 486 and have a complete machine. That's how
>||the upgrades work with the 486SX chip. The only disadvantage with the SX is
>||that at present, only 20 MHz chips exist, as compared to the 25-50 MhZ of the
>||real 486.
>|
>|Well.......
>|
>|From what i've read, the pinouts on the 486sx are slightly different
>|than on the 486dx...
>|
>|Undoubtedly a ploy for Intel to then catch you in the next scam.... The
>|487sx! The 487sx is actually just a 486dx... plugs into the
>|'coprocessor' slot on the 486sx boards and actually DISABLES the entire
>|486sx processor...

Actually (This info comes from an Intel employee), the 486SX was derived from
market pressure from system manufacturers. Not entirely a unilateral decision
on Intel's part to make the chip.

Russ Poffenberger DOMAIN: pof...@sj.ate.slb.com
Schlumberger Technologies UUCP: {uunet,decwrl,amdahl}!sjsca4!poffen
1601 Technology Drive CIS: 72401,276
San Jose, Ca. 95110 (408)437-5254

0 new messages