486sx VS 486dx Differences needed please...
x92dod...@gw.wmich.edu
I'm trying to convince may boss to get me a PC and I said that I would
like to have an 486DX compared to 486SX, then came the fatel question
"What are the differences between an SX and a DX...? How is it going
to help you do your work more efficiently...can the cost difference
be justified....?"
eeeeeekkkkk !!! I screamed, I was able to give him some answers on the
spur of the moment such as
1. the way memory is accessed...
2. Speed
3. .......started to choke......
Can any kind soul please give some pros and cons (I'll keep the cons away
from my boss...;-)) between a Intel 80486sx and a 80486dx...
thanks,
SC
please mail responses to med...@emunix.emich.edu
James J Wu
The 486SX and 486DX are identical chips (access memory the same way, 8K
internal cache, etc.) except for the fact that the 486SX is "crippled".
Explanation: The math co-processor on the 486SX is disabled.
That's it. If you do work that requires a lot of number crunching, then you'll
see a noticeable difference in speed (if the programs use the math
co-processor). Otherwise, the difference is minimal. Graphics will be
slightly slower.
There was a big fuss over the disabled FPU when the 486SX was released. Some
people are still upset about it. The 486SX is still faster than most 386's,
but the big advantage of the 486DX is the integrated math co-processor.
Without the FPU, an i486 is nothing more than a more efficient 386 with an
internal cache.
FYI, the AMD 386-40 MHz chip is faster than a 486SX chip.
Hope this helps.
james
--
Eye on the fuelgauge \
Westchester Thruway \ jj...@magnus.acs.ohio-state.edu
that triple octane \__________________________
won't contain the empty feeling in Dolby's heart... \
--T. Dolby \
\_____________
Tor Slettnes
| Without the FPU, an i486 is nothing more than a more efficient 386 with an
| internal cache.
Wrong. The 486 has some extra instructions; as well as some
more advanced stuff that I forgot. Among other things,
NeXTStep/486 requires a 486 to run.
+------
| FYI, the AMD 386-40 MHz chip is faster than a 486SX chip.
+------
Also wrong.
The 486sx comes in three (four?) speeds: 20, 25, 33 (40?) MHz.
A 486sx/25 is slightly faster than a 386dx/40 for most
operations.
(The most common instructions have reduced the cycle count
from 2 to 1).
_______________________________________________________________
If you think 10ยข is much to brag about, someone stole your bits
Dwight Joe
writes:
>Hi fellow Netters,
> I'm trying to convince may boss to get me a PC and I said that I would
>like to have an 486DX compared to 486SX, then came the fatel question
> "What are the differences between an SX and a DX...? How is it going
> to help you do your work more efficiently...can the cost difference
> be justified....?"
The only difference between a 486DX and a 486SX is that the DX has an
active FPU and that the SX has an disable FPU. The SX should be cheaper
because the yield from the same fabrication process (that produces both
the SX and the DX) is higher for the SX than for the DX. (If you can't reason
out why, then send me e-mail.)
Is the higher cost of the DX justified? That depends on what type of work
that you are doing.
Given that you are posting from a university, I'll make the possibly
unwarranted assumption that you are employed there. If so, then I'm sure
that there is a SunSparcstation or DecStation for the real number crunching.
If you are using an x86 for number crunching, you are making a mistake.
The Sparc is about the speed of a 486 for integer operations but is lots
faster for FP. (Many benchmarks confirm this result.)
Since you probably don't need number-crunching power, the 486 SX is fine!
Why. I read in the local San Francisco Chronicle that CompUSA offers
a Compaq Prolinea with 486SX and 3 years of warranty for only $1199.
(It also comes loaded with MS-DOS 6.0, Windows 3.1, 4MB of memory,
2 FDD, 120 MB HD, mouse etc.) That price undercuts even no-name clones; one
such clonemaker comes to mind... It's a great buy if ever I saw one.
DISCLAIMER: (1) These opinions are _exclusively_ my opinions and do not, in
any way, reflect the opinions of those with whom I may be
affiliated. (2) My place of employment is the Toy Factory, Inc.
on Uranus. (3) In other words, get off my back!
jyf...@hkucc.hku.hk
or even 33, while the price gap is not much indeed.
Guy Dawson
In article <1993Aug9.1...@leland.Stanford.EDU>, unde...@leland.Stanford.EDU (Dwight Joe) writes:
> In article <1993Aug8.2...@gw.wmich.edu> x92dod...@gw.wmich.edu
> writes:
> >Hi fellow Netters,
> > I'm trying to convince may boss to get me a PC and I said that I would
> >like to have an 486DX compared to 486SX, then came the fatel question
> > "What are the differences between an SX and a DX...? How is it going
> > to help you do your work more efficiently...can the cost difference
> > be justified....?"
>
> The only difference between a 486DX and a 486SX is that the DX has an
> active FPU and that the SX has an disable FPU. The SX should be cheaper
> because the yield from the same fabrication process (that produces both
> the SX and the DX) is higher for the SX than for the DX. (If you can't reason
> out why, then send me e-mail.)
Bzzzzzt!
The current SX die does not have the FPU on it at all. It's not disabled,
it's not there! The yield for SXs is higher because the number of defects
on a wafer is related to wafer size and chip technology. Make the chips
smaller ( remove the FPU ) and you can get more on the wafer getting a
higher yield...
[ stuff deleted ]
Guy
--
-- -----------------------------------------------------------------------------
Guy Dawson - Hoskyns Group Plc.
gu...@hoskyns.co.uk Tel Hoskyns UK - 71 251 2128
gu...@austin.ibm.com Tel IBM Austin USA - 512 838 2334
Dwight Joe
unde...@leland.Stanford.EDU (Dwight Joe) writes:
>The only difference between a 486DX and a 486SX is that the DX has an
>active FPU and that the SX has an disable FPU. The SX should be cheaper
>because the yield from the same fabrication process (that produces both
>the SX and the DX) is higher for the SX than for the DX. (If you can't reason
>out why, then send me e-mail.)
I seem to have generated some e-mail and newsgroup flames on this comment of
mine. The early production runs of 486SX's did come from the same fab
line as the 486DX's.
Yes. The later production of 486SX's come from a different line, and
the FPU's no longer exist.
Brad.Wright
James> said: NNTP-Posting-Host: magnusug.magnus.acs.ohio-state.edu
James> FYI, the AMD 386-40 MHz chip is faster than a 486SX chip.
James> Hope this helps.
James> james
I keep hearing this 386-40 is faster than an 486SX and it JUST NOT TRUE.
Maybe a 40 is faster than the ol 20Mhz SX, but you hardly see them anymore
and a 25Mhz is FASTER that a 386-40. Go test them yourself if you don't
believe it....I did!!
--
___________________________________________________________________________
"It's a small mind who can think of only one way to spell a word" - Twain
----Yes, these opinions are mine, not those of NCR......bla bla bla bla---
Brad Wright Brad....@ftcollins.ncr.com
Jeremy Phillips
[stuff deleted]
>>
>> The only difference between a 486DX and a 486SX is that the DX has an
>> active FPU and that the SX has an disable FPU. The SX should be cheaper
>> because the yield from the same fabrication process (that produces both
>> the SX and the DX) is higher for the SX than for the DX. (If you can't reason
>> out why, then send me e-mail.)
>Bzzzzzt!
>The current SX die does not have the FPU on it at all. It's not disabled,
>it's not there! The yield for SXs is higher because the number of defects
>on a wafer is related to wafer size and chip technology. Make the chips
>smaller ( remove the FPU ) and you can get more on the wafer getting a
>higher yield...
I beg to differ. Upon running a program called Landmark on my 486SX-25,
the program detected what it 'thought' was an FPU, running at an
unbelievably _slow_ rate (I believe it was somewhere in the kHz). I have
also heard that some programs (in particular the older versions of Lotus
1-2-3) auto-detect for an FPU on a system and if they find one, send
some of the calculations its way. When it gets nothing back from the
pseudo-FPU, the program just substitutes 0's it its place.
It could just be some form of anomaly of the 486SX chip, but I'm pretty
sure it's indicative of a disabled FPU.
Jeremy Phillips.
aka: The Big J
Email: jp...@tartarus.uwa.edu.au
Patrick Hens 44223
>>>>>> On 9 Aug 93 03:36:42 GMT, jj...@magnus.acs.ohio-state.edu (James J Wu)
> James> said: NNTP-Posting-Host: magnusug.magnus.acs.ohio-state.edu
> James> FYI, the AMD 386-40 MHz chip is faster than a 486SX chip.
>
> James> Hope this helps.
>
> James> james
>
>I keep hearing this 386-40 is faster than an 486SX and it JUST NOT TRUE.
>Maybe a 40 is faster than the ol 20Mhz SX, but you hardly see them anymore
>and a 25Mhz is FASTER that a 386-40. Go test them yourself if you don't
>believe it....I did!!
My motherboard holds a 486DLC 33MHz processor. Is this the same processor
as a 486SX or 486DX. Does it have a co-processor?
Patrick Hens.
James J Wu
>> James> FYI, the AMD 386-40 MHz chip is faster than a 486SX chip.
>>
>>I keep hearing this 386-40 is faster than an 486SX and it JUST NOT TRUE.
>>Maybe a 40 is faster than the ol 20Mhz SX, but you hardly see them anymore
>>and a 25Mhz is FASTER that a 386-40. Go test them yourself if you don't
>>believe it....I did!!
I've received e-mail and read several posted comments about this fact. I
apologize for posting somewhat out of date information. :)
I still find it both amusing and appalling that the initial release of the
486SX chip (20 MHz) was slower than the AMD 386DX-40 MHz chip.
>My motherboard holds a 486DLC 33MHz processor. Is this the same processor
>as a 486SX or 486DX. Does it have a co-processor?
I am assuming (yes, I know the saying...) that this is a Cyrix CPU. To
paraphrase a couple magazines: the Cyrix 486DLC/33 is faster than an AMD
386DX/40 and slower than an Intel 486SX/25. It does have a co-processor of
sorts which does arithmetic computations. A floating-point unit (FPU), which
many people call a math co-processor, is not included internally but can be
purchased separately.
Addressing another inaccuracy in my previous posting:
-the 486SX no longer contains an FPU
james
--
!
...Down with the landing gear !
up with the useless prayer... ! jj...@magnus.acs.ohio-state.edu
--T. Dolby !
!
fred j mccall 575-3539
>gu...@austin.ibm.com (Guy Dawson) writes:
>[stuff deleted]
>>Bzzzzzt!
Only if you have a very, very old SX is this even possible, and even
then I wouldn't consider it likely since on those the internal
connections to the FPU were physically cut so you weren't going to see
anything back from it anyway.
How do you expect Landmark to be able to tell if the math is
physically there or being emulated? The only way to tell is by the
machine configuration. There is no FPU on any 486SX that has been
manufactured in quite some time. Don't try to 'reason' it out -- just
look at the frigging die!
--
"Insisting on perfect safety is for people who don't have the balls to live
in the real world." -- Mary Shafer, NASA Ames Dryden
------------------------------------------------------------------------------
Fred....@dseg.ti.com - I don't speak for others and they don't speak for me.
fred j mccall 575-3539
Neither. If this is the Cyrix part with that designation, it is a
chip with the i486 instruction set, a 1k internal cache (vice 8k), no
floating point unit, and the 386DX pinout. It is, for all intents and
purposes, a 486SX with a 386DX pinout (plus one extra pin for the
cache enable line).
fred j mccall 575-3539
>Hi fellow Netters,
> I'm trying to convince may boss to get me a PC and I said that I would
>like to have an 486DX compared to 486SX, then came the fatel question
> "What are the differences between an SX and a DX...? How is it going
> to help you do your work more efficiently...can the cost difference
> be justified....?"
> eeeeeekkkkk !!! I screamed, I was able to give him some answers on the
>spur of the moment such as
> 1. the way memory is accessed...
No difference.
> 2. Speed
Only for floating point math intensive things, unless you are talking
about being able to get DX chips that run at faster clocks (and SX
goes up to 33 MHz these days, I think).
> 3. .......started to choke......
Only if you try to swallow them without cutting the little legs off.
;-)
>Can any kind soul please give some pros and cons (I'll keep the cons away
>from my boss...;-)) between a Intel 80486sx and a 80486dx...
The only real difference you are going to see is if you are doing
things that involve a lot of floating point math (CAD, spreadsheets,
etc.). For word processeing and such, there really isn't much
difference.
fred j mccall 575-3539
>In article <1993Aug8.2...@gw.wmich.edu> x92dod...@gw.wmich.edu
>writes:
>>Hi fellow Netters,
>> I'm trying to convince may boss to get me a PC and I said that I would
>>like to have an 486DX compared to 486SX, then came the fatel question
>> "What are the differences between an SX and a DX...? How is it going
>> to help you do your work more efficiently...can the cost difference
>> be justified....?"
>The only difference between a 486DX and a 486SX is that the DX has an
>active FPU and that the SX has an disable FPU. The SX should be cheaper
>because the yield from the same fabrication process (that produces both
>the SX and the DX) is higher for the SX than for the DX. (If you can't reason
>out why, then send me e-mail.)
They don't use the same process or packaging, and the masks for the SX
don't even include the FPU (i.e., it's not disabled, it's GONE).
Hamish Moffatt
> The 486SX and 486DX are identical chips (access memory the same way, 8K
> internal cache, etc.) except for the fact that the 486SX is "crippled".
>
> Explanation: The math co-processor on the 486SX is disabled.
I was under the impression that in many cases now the 486SX actually
does not contain a co-processor. Anyone?
> FYI, the AMD 386-40 MHz chip is faster than a 486SX chip.
That's good. (I'm getting my second 386-40 system soon. :-)
hamish
--
Hamish Moffatt, APANA Map co-ordinator
ham...@cloud.apana.org.au For APANA info mail in...@apana.org.au
h.mo...@apana.org.au Cloud Nine BBS: +61-3-803-6954
Melbourne, Australia 3:635/552@fidonet, 58:4100/43@intlnet
Tanvir Hafiz
Jj J
> go with ryzen 7 3700x. it smokes sx, dx all other intel processor for last 20 generations out of water
Jj J
> ... almost 20 years old thread
Scratch that. I meant _more_ than 20 years old thread.
vladimir...@gmail.com
Robert Riebisch
--
Robert Riebisch
vladimir...@gmail.com
JJ
>
> I think, you're late. By 27 years only. *g*
first place, became the boss himself, or no longer work at that company.
Either way, the reality have already branched. Answering it now won't create
another branch of reality.
IOTW, no use on shooting the footprints of a deer, in hopes that it ran the
other way.
Tom Pestak
R.Wieser
Does the term "necrobumping" mean anything to you ? That message you
responded to is almost *thirty years* old. Heck, there is a possibility
its older than you. :-)
Regards,
Rudy Wieser
Tom Pestak
JJ
Kerr-Mudd, John
many on-topic posts here these days!)
--
Bah, and indeed Humbug.