Does a 486DX run Doom faster than an SX?
Dave Gates
a DX. Any opinions? (or hard facts :) ?)
Dave Gates
Saari Anssi
The answer to the question on the subject line is `no.'
:>My friend is buying a computer and doesn't see much reason to buy
:>a DX.
Me neither. Only if he runs software that needs the FPU would it make
sense to pay more for the DX.
Anssi
Neal W. Miller
>My friend is buying a computer and doesn't see much reason to buy
>a DX. Any opinions? (or hard facts :) ?)
An SX is considerably slower than a DX for most processor-intensive
applications and games, including DOOM. A 486-33dx is close to double the
speed of a 486-33sx. On a rating system based on the 486-33sx, the speeds
are LOOSELY as follows:
486sx-33 = 1 486dx2-50 = 2.6
486dx-25 = 1.6 486dx-50 = 2.7
486dx-33 = 1.9 486dx2-66 = 3.3
486dx-40 = 2.2 486dx4-100/P5-60 = 6.0
As I said, these figures are rough. The best "bang for buck" deal
is probably buying an AMD 486dx-40 and overclocking it to 50Mhz. I've been
doing this since November with no problems. My next step-up will probably be
a dx4-100 once I can plug it (somehow) into my 5.5v motherboard, and the chip
comes down to current 486dx prices. DOOM on a 486dx-40 is just fine, and the
improvement at 50Mhz is only noticable in "Slow areas" of DOOM (Mt. Erebus,
and a "grey maze" in Aliens DOOM come to mind).
--
------------------------------------------------------------------------------
Neal Miller | Rensselaer Polytech. | "This Side Up - Not Plummet Please!"
mil...@rpi.edu | Troy, NY 12180 | - the box for my Taiwanese CPU fan
------------------------------------------------------------------------------
Grassl Wolfgang
->dga...@bestsd.sdsu.edu (Dave Gates) writes:
->>My friend is buying a computer and doesn't see much reason to buy
->>a DX. Any opinions? (or hard facts :) ?)
-> An SX is considerably slower than a DX for most processor-intensive
->applications and games, including DOOM. A 486-33dx is close to double the
->speed of a 486-33sx. On a rating system based on the 486-33sx, the speeds
->are LOOSELY as follows:
->486sx-33 = 1 486dx2-50 = 2.6
->486dx-25 = 1.6 486dx-50 = 2.7
->486dx-33 = 1.9 486dx2-66 = 3.3
->486dx-40 = 2.2 486dx4-100/P5-60 = 6.0
-> As I said, these figures are rough. The best "bang for buck" deal
->is probably buying an AMD 486dx-40 and overclocking it to 50Mhz. I've been
->doing this since November with no problems. My next step-up will probably be
->a dx4-100 once I can plug it (somehow) into my 5.5v motherboard, and the chip
->comes down to current 486dx prices. DOOM on a 486dx-40 is just fine, and the
->improvement at 50Mhz is only noticable in "Slow areas" of DOOM (Mt. Erebus,
->and a "grey maze" in Aliens DOOM come to mind).
->--
->------------------------------------------------------------------------------
->Neal Miller | Rensselaer Polytech. | "This Side Up - Not Plummet Please!"
->mil...@rpi.edu | Troy, NY 12180 | - the box for my Taiwanese CPU fan
->------------------------------------------------------------------------------
Thats wrong !
A 486 SX runs Doom with exactly the same speed like a 486 DX
(if you use the same VGA Card and Motherboard).
The 486 SX 25 overclocked to 40 Mhz (sometimes even 50 works fine) is the
cheapest and ideal Doom-Machine !
PS: The only difference between a 486(!) SX and DX is the FPU.
A 486 SX has got 8 KB internal cache.
-----------------------------------------------------------------------
Wolfgang Grassl - gra...@mailer.uni-marburg.de
-----------------------------------------------------------------------
BillyBoB 4
(Dave Gates) writes:
DOOM runs *MUCH* faster on a 486DX/33 than on a 486SX/33. Believe me,
I've seen it running on the 2 different machines in the same room.
摺＃ミ斑
BonesBro
high detail then the SX runs at postage stamp. It is so slow as to be
almost unplayable.
Chad Anson
They are *NOT* any different as far as CPU speed go. Period.
The reason one (DX) is faster must have something to do with
probably the SX has an ISA video card, or no cache, or less
memory.
Doom does not use an FPU, so there will be no slowdown for the SX.
Chad Anson
Stephen E. Collins
>dga...@bestsd.sdsu.edu (Dave Gates) writes:
>
>>My friend is buying a computer and doesn't see much reason to buy
>>a DX. Any opinions? (or hard facts :) ?)
>
> An SX is considerably slower than a DX for most processor-intensive
>applications and games, including DOOM. A 486-33dx is close to double the
>speed of a 486-33sx. On a rating system based on the 486-33sx, the speeds
>are LOOSELY as follows:
>
[Meaningless CPU number comparisons and such deleted]
For anyone considering and SX CPU, you should know that this
post is profoundly false and misleading. Doom does not make
*any* specifc use of the math coprocessor, which is the only
difference between an Intel 486 DX and SX chip.
Furthermore, he is very misleading in saying that a 486-33dx
is double the speed of the 486-33sx. Again, the only way a
DX chip is going to outperform an SX chip is in an application
that specifically makes use of the math coprocessor, of which
few exist.
Finally, I just sold my Intel 486-33DX and am using the SX
version until my new chip gets here, and I can verify that
there is no difference whatsoever in the performance of the
two chips for nearly every application I use (including DOOM,
Windows, Word processors, and the like).
Robert Kimmel
>Dave Gates
If your friend is just using the computer to play games, or maybe do a
little word processing, no reason to buy a DX (mainly since you can pick
up Intel's SX chips for less then $100). The only difference between the
two is that the DX has a FPU, and currently, most games don't use it.
Now if he wants to do alot of math work, or graphics work (raytracing, 3d
programs) he'll want the DX because they use the fpu unit.
Robert Kimmel
Besides the obvious fact that the DX is twice as fast as the SX here, you
need to take in Cache size and video cards. They play a large factor in
how well the game plays.
Josh M. Michaels
486 DX's Have a floating point unit [math coprocessor][well my DX2 does, and i think all dx's do] which WILL speed up doom, but not to the extent that it is work more then $100
-josh [still annoyed about doom2.wad]
--
o c t o b e r 3 r d -- i t ' s c o m m i n g -- DOOMSDAY
j o s h m i c h a e l s -- j o s h m @ prairienet. o r g<-------.
e a s y come e a s y go -- l i t t l e high, little l o w |-----.
worldWideWeb programming -- 217 - 367 - 2475/2463 | |
Possible Exploding Thug
: bone...@aol.com (BonesBro) writes:
The diffrence between an sx and a dx is the math coprocessor. I believe
that if you are doing cad work in a windows environment it will improve
your speed. As it should also improve at a dos prompt. It all has to
do with the numbers! It's the numbers man. The higher you get the
faster you go. It's true with amount of money spent on a race car.
It's true of amount of money spent on computer.
babble babble babble.
--
_____________________________________________________________________________
be...@netcom.com be...@aol.com
=`=`=`=`=`=`=`=`=`=``=`=`=`=`=`=`=`=`=`=`=`=`=`=`=`=`=`=`=`=`=`=`=`=`=`=`=`=`=
If I convert all my thought into energy, Shall this make me smarter?
If I convert all my energy into thought, Shall I explode?
Neal W. Miller
>Neal W. Miller (mil...@rebecca.its.rpi.edu) wrote:
>->dga...@bestsd.sdsu.edu (Dave Gates) writes:
>->>My friend is buying a computer and doesn't see much reason to buy
>->>a DX. Any opinions? (or hard facts :) ?)
>-> An SX is considerably slower than a DX for most processor-intensive
>->applications and games, including DOOM. A 486-33dx is close to double the
>->speed of a 486-33sx.
>Thats wrong !
>A 486 SX runs Doom with exactly the same speed like a 486 DX
>(if you use the same VGA Card and Motherboard).
>The 486 SX 25 overclocked to 40 Mhz (sometimes even 50 works fine) is the
>cheapest and ideal Doom-Machine !
>PS: The only difference between a 486(!) SX and DX is the FPU.
> A 486 SX has got 8 KB internal cache.
Er... Have you *tried* DOOM on both machines? I have, on a 486dx-33
and a 486sx-33. They had comparable non-VLB video cards, and eight megs of
memory. The DX was unquestionably superior in DOOM performance to the SX.
I imagine that this comparison would carry over to higher Mhz. The internal
cache makes a major difference.
For the record, a 486sx-25 is about the same speed as a 386dx-40.
--
------------------------------------------------------------------------------
Neal Miller | Rensselaer Polytech. | "This Side Up - Not Plummet Please!"
Neal W. Miller
>In article <30r694$p...@usenet.rpi.edu> mil...@rebecca.its.rpi.edu (Neal W. Miller) writes:
>>dga...@bestsd.sdsu.edu (Dave Gates) writes:
>>
>>>My friend is buying a computer and doesn't see much reason to buy
>>>a DX. Any opinions? (or hard facts :) ?)
>>
>> An SX is considerably slower than a DX for most processor-intensive
>>applications and games, including DOOM. A 486-33dx is close to double the
>>speed of a 486-33sx. On a rating system based on the 486-33sx, the speeds
>>are LOOSELY as follows:
>>
>[Meaningless CPU number comparisons and such deleted]
>For anyone considering and SX CPU, you should know that this
>post is profoundly false and misleading. Doom does not make
>*any* specifc use of the math coprocessor, which is the only
>difference between an Intel 486 DX and SX chip.
Can you provide ANY benchmarks to back this up? You CANNOT reduce
the internal cache on a CPU, and tell me that you aren't suffering a
performance loss.
Chad Anson
Josh M. Michaels <jo...@prairienet.org> wrote:
>
>486 DX's Have a floating point unit [math coprocessor][well my DX2 does, and i think all dx's do] which WILL speed up doom, but not to the extent that it is work more then $100
>-josh [still annoyed about doom2.wad]
>--
Doom does not in any shape, form, or function use an FPU. Not at all, not
even a little bit. The only thing that uses an FPU that is even vaguely
connected to Doom are some versions of BSP and DEU (I think those are it, but
there may be more).
Again: Doom does NOT NOT NOT NOT use an FPU (math coprocessor).
Chad Anson
Richard Ward
>
>486 DX's Have a floating point unit [math coprocessor][well my DX2 does, and i think all dx's do] which WILL speed up doom, but not to the extent that it is work more then $100
>-josh [still annoyed about doom2.wad]
>--
Hate to burst your bubble, but DOOM does not useFPU in the DX at all (ask the
guys at ID if you don't belive me). All things being equal (same motherboard,
same memory, cache, video card, etc...) a 486DX-33 will _NOT_ run DOOM any
faster than a 486SX-33. Only programs that use the FPU (and I want to make
this clear) will benefit from the DX.
Richard
PS: Please set you line width on yor terminal to somewhere around 70-80
collomns.
Alan Wen (409)696-8909
>gra...@papin.HRZ.Uni-Marburg.DE (Grassl Wolfgang) writes:
> Er... Have you *tried* DOOM on both machines? I have, on a 486dx-33
>and a 486sx-33. They had comparable non-VLB video cards, and eight megs of
>memory. The DX was unquestionably superior in DOOM performance to the SX.
>I imagine that this comparison would carry over to higher Mhz. The internal
>cache makes a major difference.
You're not using the SAME video cards. You're probably not even using the
same motherboard. Er... What don't you try using the same machine but
removing the CPU and replacing it with each while keeping EVERYTHING else
the same? Try it. You'll learn that you're truly mistaken.
>Neal Miller | Rensselaer Polytech. | "This Side Up - Not Plummet Please!"
alan
Jason Steven O Rourke
> Can you provide ANY benchmarks to back this up? You CANNOT reduce
>the internal cache on a CPU, and tell me that you aren't suffering a
>performance loss.
Not exactly a tribute to your university are you?
Did you somehow confuse the intel/amd SX chips with the Cyrix SLCs, which
have a 1k cache? Or are you thinking of the 16bit IBM SLC with the
larger 16k cache?
Intel suggests that 92% of normal (average) PC use involves integer
calculations. In the case of Doom and other games, it is faster to use
integer approxiamations then to go for more accurate floats. As the fpus
get faster and faster, this may change. For now, the purchase of an SX2
cpu is a good choice for non engineers. The only programs I use that
require the fpu are astro programs, fractal generators, and Falcon.
Neal, take a few more classes and get back to us. Right now you are
doing us a serious disservice by spreading your fabricated benchmarks
(want to site a source for those 'numbers?')
Jason O'Rourke
I&AS Computing Support
rachel ann sloan
> For the record, a 486sx-25 is about the same speed as a 386dx-40.
For the record--check out the pc-mags.
a 486sx25 runs at least 25% faster (more on some things) than a 386dx40.
this is according to the winmarks test and the comparable test used for dos.
I just saw this the other day in computer shopper, so I think I remember it
fairly well.
Stephen E. Collins
>col...@eng.auburn.edu (Stephen E. Collins) writes:
>
>>In article <30r694$p...@usenet.rpi.edu> mil...@rebecca.its.rpi.edu (Neal W. Miller) writes:
>>>dga...@bestsd.sdsu.edu (Dave Gates) writes:
>>>
>>>>My friend is buying a computer and doesn't see much reason to buy
>>>>a DX. Any opinions? (or hard facts :) ?)
>>>
>>> An SX is considerably slower than a DX for most processor-intensive
>>>applications and games, including DOOM. A 486-33dx is close to double the
>>>speed of a 486-33sx. On a rating system based on the 486-33sx, the speeds
>>>are LOOSELY as follows:
>>>
>
>>[Meaningless CPU number comparisons and such deleted]
>
>>For anyone considering and SX CPU, you should know that this
>>post is profoundly false and misleading. Doom does not make
>>*any* specifc use of the math coprocessor, which is the only
>>difference between an Intel 486 DX and SX chip.
>
> Can you provide ANY benchmarks to back this up? You CANNOT reduce
>the internal cache on a CPU, and tell me that you aren't suffering a
>performance loss.
I don't need any! Ask anyone and they will tell you that the only
diference between the two chips is the math-coprocessor. They *both*
use an 8Kb internal CPU cache as well as my 256Kb secondary cache.
Also, I have tried running both CPU's in the exact same system, and
I can tell from experience that DOOM performance is unaffected.
-Steve
Stephen E. Collins
>-josh [still annoyed about doom2.wad]
Nope, I have tried both a 486-33DX and a 486-33SX in the same machine,
and DOOM performance is *exactly* the same. DOOM does not use the
math coprocessor on a DX chip. In fact, I can not think of any games
offhand that do use it.
Steve
Ron Pritchett
>Nope, I have tried both a 486-33DX and a 486-33SX in the same machine,
>and DOOM performance is *exactly* the same. DOOM does not use the
>math coprocessor on a DX chip. In fact, I can not think of any games
>offhand that do use it.
I've seen DOOM on a 486/25sx, 486/33dx, 486/50dx & 486/66dx2 and it all
looked the same. It plays slower on a 386.
--
---
+============================================================================+
| "They say I'm lazy, but it takes all my time... |
| Life's been good to me so far!" -Joe Walsh |
|----------------------------------------------------------------------------|
| Ron Pritchett Internet: prit...@usceast.cs.scarolina.edu |
| * Team OS/2 * FidoNet: Ron Pritchett (1:376/74.0) |
| * Linux Nut * |
| <<Finger for PGP Public Key>> |
+============================================================================+
David Garry BIGGS - Dave
>From: prit...@usceast.cs.scarolina.edu (Ron Pritchett)
>Subject: Re: Does a 486DX run Doom faster than an SX?
>Date: 24 Jul 1994 10:02:22 -0400
>col...@eng.auburn.edu (Stephen E. Collins) writes:
>>Nope, I have tried both a 486-33DX and a 486-33SX in the same machine,
>>and DOOM performance is *exactly* the same. DOOM does not use the
>>math coprocessor on a DX chip. In fact, I can not think of any games
>>offhand that do use it.
>I've seen DOOM on a 486/25sx, 486/33dx, 486/50dx & 486/66dx2 and it all
>looked the same. It plays slower on a 386.
Absolute crap.....I reackon you need your eyes tested. I can tell the
difference between a 486dx50 and a 486dx266. And I know a 486dx33 is
noticably slower than a 486dx50. There is know way that a 486sx25 is the
same speed as a 486dx266 with doom. I think your absolutley MAD!
Patrick C Miller
: We have a 486SX\25 and a 486DX\50 and the DX 50 runs faster at full screen
: high detail then the SX runs at postage stamp. It is so slow as to be
: almost unplayable.
Then there is something really wrong with your 486SX/25. I have a
486SX/20 and I play Doom with full screen on low detail. It may not be
as smooth as a faster machine, but it is plenty smooth for me. For the
record, I have 8 MB of RAM and a Tseng ET-4000 video card with 1 MB of
VRAM and NO local bus.
============================================
# #
# Patrick Miller #
# Communications Coordinator #
# Energy & Environmental Research Center #
# University of North Dakota #
# #
============================================
Neal W. Miller
>Neal W. Miller <mil...@rebecca.its.rpi.edu> wrote:
>> Can you provide ANY benchmarks to back this up? You CANNOT reduce
>>the internal cache on a CPU, and tell me that you aren't suffering a
>>performance loss.
>Neal, take a few more classes and get back to us. Right now you are
>doing us a serious disservice by spreading your fabricated benchmarks
>(want to site a source for those 'numbers?')
I cited the source very clearly... I said I that I estimated them
based on my experience with the machines. I never claimed to be an
authority, and I never claimed to have any academic degrees in programming
or hardware design. I have plenty of experience on literally scores of
different PCs with process that run the full gamut, and all of my claims are
based on that experience. Take it or leave it.
Neal W. Miller
Which issue? I didn't think CS had benchmarked 386 systems in years.
Stephen E. Collins
>
>>>Nope, I have tried both a 486-33DX and a 486-33SX in the same machine,
>>>and DOOM performance is *exactly* the same. DOOM does not use the
>>>math coprocessor on a DX chip. In fact, I can not think of any games
>>>offhand that do use it.
>
>>I've seen DOOM on a 486/25sx, 486/33dx, 486/50dx & 486/66dx2 and it all
>>looked the same. It plays slower on a 386.
>
>Absolute crap.....I reackon you need your eyes tested. I can tell the
>difference between a 486dx50 and a 486dx266. And I know a 486dx33 is
>noticably slower than a 486dx50. There is know way that a 486sx25 is the
>same speed as a 486dx266 with doom. I think your absolutley MAD!
Agreed. CPU speed does make a big difference on levels where a
lot is going on at the same time (E3-M6 comes to mind).
I should have also mentioned before that I ran the 3DBench test
on my machine with the 33-DX and 33-SX chips, and got the exact
same score of 28.1 each time.
Trust me, DOOM does not make use of a 486 math co-processor!
-Steve
David Hawks
Looks like some people out there should spend a couple months reading
the comp.sys.ibm.pc.hardware.* groups. As many have noted there is no
difference between an SX and a DX processor at the same clock speed when
it comes to DOOM. As a simple example, my 486SX-33 blows away several
DX-33's around the block when it comes to smooth graphics. How can that be?
Well, I would think the two biggest things would be the cpu/memory
throughput I get on my motherboard and the use of an S3 based VLB graphics
card. What happens when your friend has a VLB graphics card that is
half the speed of yours? What happens when you see 12 wait states on his
MB when there should only be between 3 and 4. Your BIOS is optimized
with all the wait states at there lowest. He can't seem to get the
lowest as his MB wasn't put together that well or his memory is too
slow. And onwards blah blah, techno-garble, blah blah. So in short,
there is just alot more to compare then just the CPU itself.
As someone mentioned, what games use a FPU? Well, actually, there are
some flight simulators that use a FPU. I have seen one that has 4
settings, 3 for integer ops and then one that hits the FPU. And of course,
the one that uses the FPU gives you the most detail. But hey, being that
I am not designing space ships to Mars with the latest CAD software and
calculating my costs with the blah-blah 2.0 spreedsheet, I don't need to
blow the extra cash on a DX processor. Save your money and buy a fast
VLB graphics card or even better a faster processor.
-- David
--
*-----------------------------------------------------------------------------*
| ha...@seq.cms.uncwil.edu | That which doesn't kill | W:(910)350-7803 9-5 EST|
| hawk...@wl.corning.com | me makes me stronger. | H:(910)799-2898 |
*-----------------------------------------------------------------------------*
Neal W. Miller
>col...@eng.auburn.edu (Stephen E. Collins) writes:
>>>>col...@eng.auburn.edu (Stephen E. Collins) writes:
(much deleted)
>Looks like some people out there should spend a couple months reading
>the comp.sys.ibm.pc.hardware.* groups. As many have noted there is no
>difference between an SX and a DX processor at the same clock speed when
>it comes to DOOM. As a simple example ...
Not to be picky but does your statement cover ALL SX and DX chips?
I may be wrong as far as the 486dx/sx business goes, but I'm quite certain
that neither the 386sx nor the 386dx have math-cos, and the DX, in many
applications is faster than the SX.
Why is there a speed diff. in DOOM between the 386sx and the 386dx
then, seeing as the math-co point is moot?
David Hawks
>ha...@sol.cms.uncwil.edu (David Hawks) writes:
>>col...@eng.auburn.edu (Stephen E. Collins) writes:
>>>>>col...@eng.auburn.edu (Stephen E. Collins) writes:
>(much deleted)
>>Looks like some people out there should spend a couple months reading
>>the comp.sys.ibm.pc.hardware.* groups. As many have noted there is no
>>difference between an SX and a DX processor at the same clock speed when
>>it comes to DOOM. As a simple example ...
Oooops I was refering to just the 486 class of CPU's.
> Not to be picky but does your statement cover ALL SX and DX chips?
>I may be wrong as far as the 486dx/sx business goes, but I'm quite certain
>that neither the 386sx nor the 386dx have math-cos, and the DX, in many
>applications is faster than the SX.
The 386DX is faster than the 386SX right? That is what you are saying?
And you would be right of course since the external data path is 32 bits
for the 386DX and 16 bits for the 386SX.
> Why is there a speed diff. in DOOM between the 386sx and the 386dx
>then, seeing as the math-co point is moot?
Oh. See the above comment. The 386DX CPU has a 32 bit data path twice that
of the 386SX. In otherwords, the CPU can grab twice as much data from
memory at a time.
Michael Rice
wrote:
>Absolute crap.....I reackon you need your eyes tested.
>I can tell the difference between a 486dx50 and a
>486dx266. And I know a 486dx33 is noticably slower than
>a 486dx50. There is know way that a 486sx25 is the same
>speed as a 486dx266 with doom. I think your absolutley
>MAD!
Well, my 486DX50 ran Doom slower than my 486SX33 - the SX33 had
an old paradise SVGA, the DX50 had a Diamond Stealth VRAM.
Stuck a new Stealth 32/VLB in Saturday and now the DX50 runs
much faster than the SX33.
So, the lesson here is that video card speed is more important
than CPU speed (at least in the 486 class of chip...)
MR
Moocher/FairLight
: >col...@eng.auburn.edu (Stephen E. Collins) writes:
: >>>>col...@eng.auburn.edu (Stephen E. Collins) writes:
: (much deleted)
: >Looks like some people out there should spend a couple months reading
: >the comp.sys.ibm.pc.hardware.* groups. As many have noted there is no
: >difference between an SX and a DX processor at the same clock speed when
: >it comes to DOOM. As a simple example ...
: Not to be picky but does your statement cover ALL SX and DX chips?
: I may be wrong as far as the 486dx/sx business goes, but I'm quite certain
: that neither the 386sx nor the 386dx have math-cos, and the DX, in many
: applications is faster than the SX.
: Why is there a speed diff. in DOOM between the 386sx and the 386dx
: then, seeing as the math-co point is moot?
There is a BIG difference between a 386dx and a 386sx! The sx has only got
a 16 bits external bus, while the dx has 32. But we are talking about the
difference between a 486dx and a 486sx here.. The only difference there is
that the sx has not got a math co-processor.
If you take two computers with the same amout of memmory, the same bus,
and the same vidocard, only the processor is different (sx and dx 486:ers)
DOOM will run at the same speed on both machines. period.
Richard Ward
>ha...@sol.cms.uncwil.edu (David Hawks) writes:
>
>>col...@eng.auburn.edu (Stephen E. Collins) writes:
>
>>>>>col...@eng.auburn.edu (Stephen E. Collins) writes:
>
>(much deleted)
>
>>Looks like some people out there should spend a couple months reading
>>the comp.sys.ibm.pc.hardware.* groups. As many have noted there is no
>>difference between an SX and a DX processor at the same clock speed when
>>it comes to DOOM. As a simple example ...
>
> Not to be picky but does your statement cover ALL SX and DX chips?
>I may be wrong as far as the 486dx/sx business goes, but I'm quite certain
>that neither the 386sx nor the 386dx have math-cos, and the DX, in many
>applications is faster than the SX.
>
> Why is there a speed diff. in DOOM between the 386sx and the 386dx
>then, seeing as the math-co point is moot?
There _is_ a speed difference between the 80_3_86SX and the 80386DX because of
the differences in the way the two chips handle external memory (system RAM)
read/writes. The 386DX uses a 32-bit bus to the memory, and the 386SX uses a
16-bvit bus to memory (Intel used to clain that SX meant sixteen, now they
claim it means "inexpensive"). The _ONLY_ (and I repeat _ONLY_) difference
between the 80_4_86SX and 40_4_86DX is the existence (or lack thereof) of an
FPU. Both chips use the same memory bus width, so they both access system RAM
(or cache RAM) at the same speed.
Richard
John Reinhold
> In article <30un5n$k...@usenet.rpi.edu> mil...@rebecca.its.rpi.edu (Neal W. Miller) writes:
> >>Looks like some people out there should spend a couple months reading
> >>the comp.sys.ibm.pc.hardware.* groups. As many have noted there is no
> >>difference between an SX and a DX processor at the same clock speed when
> >>it comes to DOOM. As a simple example ...
Ok, this may be a little off the mark... But if the FPU is enabled in the
BIOS, won't most math be sent to the co-pro?
Don't flame me, I don't know much about hardware... :-)
Also, in the lab that I work at we have
Gateway 2000 (had lots of trouble with them).
We have 486DX/33 's and 486SX/33's. The SX are WAY (about 15 points
less on 3dbench) slower than the DX. Did gateway screw us? They look
the same, the motherboard looks the same, they both have WDCaviar drives,
etc.. etc..
--
----------------------------<>---------------------------
S.P.E.R.M.
Society for the Preservation of Equal Rights for Men
~ ~ ~ ~ ~
===============================
John Reinhold Portales, New Mexico (USA)
Gun Control is Un American.
Jason Steven O Rourke
>then, seeing as the math-co point is moot?
Let's review for a moment:
386 DX Standard 386 chip. 32bit internal, 32 bit external
SX Sux. 32 bit internal, 16 external. Does 2 data fetches for each DX
Since the SX takes twice as long to do IO, it runs about half the speed.
486 DX Standard 486 chip. 32/32, 8k internal cache, internal fpu
SX no fpu (disable fpu on early chips)
Since Doom is an integer game, these chips run at same speed. Note that
the 486 line executes 32bit code much faster than the 386. That and
the 486 optimizations of doom mean that the 486sx25 should smoke a
AMD386-40.
Anymore posts by you and we'll have to send you the FAQ for the
comp.sys.hardware.ibm.pc
Jason O'Rourke
Inaccurate Decay
>col...@eng.auburn.edu (Stephen E. Collins) writes:
>>Nope, I have tried both a 486-33DX and a 486-33SX in the same machine,
>>and DOOM performance is *exactly* the same. DOOM does not use the
>>math coprocessor on a DX chip. In fact, I can not think of any games
>>offhand that do use it.
>I've seen DOOM on a 486/25sx, 486/33dx, 486/50dx & 486/66dx2 and it all
>looked the same. It plays slower on a 386.
you're saying a dx2/66 shows NO better performance than an sx/25, full screen
hi-res?
Inaccurate Decay
---
"When people are asleep, we must all become alarm clocks" - Jello Biafra
male: ri...@onramp.net --- "If you love your fun, Die for it!!!" - LARD
GO d-(---) -p+ c++++ l u--- e m* s+/++ n--- h--@ f-- g+ w+(+++) t--- r(+) y**
Dave Ahn
>In article <30ts9e$8...@poplar.cs.scarolina.edu> prit...@usceast.cs.scarolina.edu (Ron Pritchett) writes:
>>col...@eng.auburn.edu (Stephen E. Collins) writes:
>>>Nope, I have tried both a 486-33DX and a 486-33SX in the same machine,
>>>and DOOM performance is *exactly* the same. DOOM does not use the
>>>math coprocessor on a DX chip. In fact, I can not think of any games
>>>offhand that do use it.
>>I've seen DOOM on a 486/25sx, 486/33dx, 486/50dx & 486/66dx2 and it all
>>looked the same. It plays slower on a 386.
>you're saying a dx2/66 shows NO better performance than an sx/25, full screen
>hi-res?
I think we need to take things into perspective. You have to take a lot
of things into account aside from just the CPU and its speed.
Assumption: only 486 32 bit CPU systems are being considred, none of the
486-SLC/DLC 16 bit external 32 bit external crap. And, I'm assuming that
the 486's are stricly Intels, because AMD or Cyrix's use different internal
cache (and chip) techonology to boost performance.
Doom is currently the informal "benchmark" and "compatibility" testers
for a lot of PC vendors right now (go see SIGGRAPH or any EXPO). Why?
Because Doom pushes the capabilities of a 486 box as a whole. Every
relevant component must be well designed and fitted together with every
other component in order for Doom to work well. If you have a Pentium 90
CPU with a cheap $50 VLB video card and 4MB of 80ns RAM, you're not going
to see any better performance than a 486-25SX with a top-of-the line
32/64 bit VLB accelerator board and 16MB of 60ns RAM. Because the crappy
components in the Pentium 90 box cause a bottle neck.
So, given two identical systems, a faster CPU will normally result in
better Doom performance so long as the other components (in particular
the memory, bus speed, video and HD/controller). This is not always
true, though. If you compare two identical systems with a DX2-66 and
a DX-50, the DX-50 might be faster. Why? because the CPU can communicate
faster with other components because of the faster bus speed. (Go
read the comp.sys.ibm.pc.hardware.* FAQ if you don't know the difference
between DX's and DX2's and DX4's)
I have an SX-25, Diamond Stealth 32 VLB 2MB, a good MB, 8MB of 60ns RAM
and a cached SCSI ISA controller. It performs better than a friend's
not so well equipped DX-40 (AMD), and almost as well as a 486DX2-50 with
ATI Mach32 VLB, 8MB 70ns RAM and VLB IDE. No choppiness at full
screen except when you have a _lot_ of monsters/actions (Nightmare!)
As far as the SX/DX is concerned, any program (including Doom) that
does not use the floating point coproc will perform equally well when
run under SX or DX of the same speed. So, an SX2-66 will give identical
performance to a DX2-66 and faster performance vs. a DX2-50.
So, when you argue or claim that my XX computer with YY CPU is faster
than your AA computer with BB CPU, please consider that you have to
take other things into account. If your video card is crappy and
is limiting the performance of Doom, upgrading your CPU and/or MB
and adding 64MB of RAM isn't going to help you much if at all.
Dave.
>Inaccurate Decay
>---
> "When people are asleep, we must all become alarm clocks" - Jello Biafra
> male: ri...@onramp.net --- "If you love your fun, Die for it!!!" - LARD
>GO d-(---) -p+ c++++ l u--- e m* s+/++ n--- h--@ f-- g+ w+(+++) t--- r(+) y**
--
Dave Ahn Internet: a...@hbar.phy.wfu.edu, a...@wfu.edu
#include <stdisclaimer.h>
"When you were born you cried, and the world rejoiced. Try to live your life
so that when you die you will rejoice, and the world will cry." -1/2 jj^2
Richard Ward
>Richard Ward (rrw...@netcom.com) wrote:
>> In article <30un5n$k...@usenet.rpi.edu> mil...@rebecca.its.rpi.edu (Neal W. Miller) writes:
>> >>Looks like some people out there should spend a couple months reading
>> >>the comp.sys.ibm.pc.hardware.* groups. As many have noted there is no
>> >>difference between an SX and a DX processor at the same clock speed when
>> >>it comes to DOOM. As a simple example ...
>
>
>Ok, this may be a little off the mark... But if the FPU is enabled in the
>BIOS, won't most math be sent to the co-pro?
their performance. 486 systems usually don't have external (a seperate chip)
co-proc, so the point is moot.
>Don't flame me, I don't know much about hardware... :-)
I never flame someone for asking an honets question (I used to do tech support
for a living).
>Also, in the lab that I work at we have
>Gateway 2000 (had lots of trouble with them).
I can belive that!
>We have 486DX/33 's and 486SX/33's. The SX are WAY (about 15 points
>less on 3dbench) slower than the DX. Did gateway screw us? They look
>the same, the motherboard looks the same, they both have WDCaviar drives,
>etc.. etc..
The thing you want to look at is the video cards. If both machines are
running at the same clock speed with integer only software (which 3dbench is),
then the only thing that could be causing the speed difference is the quality
of the video cards. Being that the SX machines probably cost a LOT less than
the DX machines, Gateway probably stuffed cheaper video cards into them. If
the video cards arte the same, then there is something wrong in the BIOS
settup of the SX.
Richard
Michael Lowe
: In <23JUL199...@bestsd.sdsu.edu> dga...@bestsd.sdsu.edu (Dave Gates) writes:
: The answer to the question on the subject line is `no.'
: :>My friend is buying a computer and doesn't see much reason to buy
: :>a DX.
: Me neither. Only if he runs software that needs the FPU would it make
: sense to pay more for the DX.
Having upgraded my own systems, several times, and having gone through
486SX, DX and DX2 CPUs, both at home and at work, I can say that without
a doubt, the ONLY performance difference between the 486 SX and DX CPUs
is the FPU.
DOOM ran just as fast on my machine when it was an SX as when it was a DX.
The numbers that were quoted earlier were numbers based on FPU tests, or
on un-equal motherboards.
Try LANDMARK's SPPED on a motherboard with an SX CPU, then remove it and
put in a DX of the same speed. Landmark will report that the CPU speed
is the same, and the video speed is the same. The only difference will
be found in the FPU speed.