Does a 487sx shut down the 486sx??
Kenton Shaver
I read it in some magazine, but my ECE
friend is credulous. Will my 487sx
just shut the 486sx down and take over?
Thanks for any replies.
Brian J. Murrell)
>Could anyone clear this up for me?
>I read it in some magazine, but my ECE
>friend is credulous. Will my 487sx
>just shut the 486sx down and take over?
That's what I heard. IMHO the 486sx is a *big scam*.
To get your 486sx you buy a chip which is a functional 486dx with the
co-pro disabled (yes more work than a regular 486dx - yet cheaper!).
At some point in the future you need the co-pro, so you buy a 487sx, which
is just the regular 486dx with a 487sx stamp on it, which shuts down the
486sx crippled chip and takes over the machine running it like the 486dx
it should have been!!
The above is all rumour I've heard from many sources, and read in a trade
rag. It is not the gospel!!
brian
--
Brian J. Murrell br...@ilinx.wimsey.bc.ca
InterLinx Support Services, Inc. uunet!van-bc!ilinx!brian
North Vancouver, B.C.
Platform and Brand Independent UNIX Support - R3.2 - R4 - BSD
Christer Olsson
|> Could anyone clear this up for me?
|> I read it in some magazine, but my ECE
|> friend is credulous. Will my 487sx
|> just shut the 486sx down and take over?
I think so, because some new mothercards simulates 486sx to do
the 487sx 486sx disabling control happy and therefore, the mothercard
only need a 487sx...
Kevin Kramer
>ks...@andrew.cmu.edu (Kenton Shaver) writes:
>
>>Could anyone clear this up for me?
>>I read it in some magazine, but my ECE
>>friend is credulous. Will my 487sx
>>just shut the 486sx down and take over?
>
>That's what I heard. IMHO the 486sx is a *big scam*.
>
>To get your 486sx you buy a chip which is a functional 486dx with the
>co-pro disabled (yes more work than a regular 486dx - yet cheaper!).
>At some point in the future you need the co-pro, so you buy a 487sx, which
>is just the regular 486dx with a 487sx stamp on it, which shuts down the
>486sx crippled chip and takes over the machine running it like the 486dx
>it should have been!!
>
>The above is all rumour I've heard from many sources, and read in a trade
>rag. It is not the gospel!!
>
I don't think it involves more work to make the 486sx chips, they are probably
rejected chips because of a fault in the math coprocessor part. So they just
disable that section of the chip and now they can use what once was a bad
chip. It is a big scam IMHO, I can't see why anyone would buy one of these.
Kevin
m...@wam.umd.edu
>
>>Could anyone clear this up for me?
>>I read it in some magazine, but my ECE
>>friend is credulous. Will my 487sx
>>just shut the 486sx down and take over?
>
>That's what I heard. IMHO the 486sx is a *big scam*.
>
>To get your 486sx you buy a chip which is a functional 486dx with the
>co-pro disabled (yes more work than a regular 486dx - yet cheaper!).
>At some point in the future you need the co-pro, so you buy a 487sx, which
>is just the regular 486dx with a 487sx stamp on it, which shuts down the
>486sx crippled chip and takes over the machine running it like the 486dx
>it should have been!!
>
>The above is all rumour I've heard from many sources, and read in a trade
>rag. It is not the gospel!!
>
>brian
>--
>Brian J. Murrell br...@ilinx.wimsey.bc.ca
487sx. Should plug into the same socket on most so you not only get the full
486DX, you get it at double the speed.
This is also what I have heard. I have only a 386-40, but PC World
had a big comparison of different 486sx systems and found the 486sx 20/25s
with the overdrive chip to be a better value than a 485/50 system.
Dave.
TOBIN MATTHEW CREEK
Yes, the 486SX is a HUGE scam. But there is a way around.
Microlab is selling a 486SX-25 that they call an SX/DX board. You can swap the
SX chip for a DX, change exactly 3 jumpers and a 4-pin chip-looking oscillator
and go clear up to a DX2-66.
I have one of these boards. I have heard bad things about Microlab before, but
I had always had good chats with them on the phone. So I decided to try them
anyway. This board is EXCELLENT. It has a 512K AMI BIOS with a full setup
built in. It supports 2 oscillators via 2 sockets for same and a jumper, in
addition to a turbo switch. It is 2/3 baby size, ideal for any installation.
The chipset is ETEQ, and it works with everything I have tried. There are
8 sockets for SIMMS (1M or 4M, 3 or 9 chip). Docs are thorough.
At $299, it gets **** with me.
Ken Germann
The 487SX will shutdown the 486SX.
--
New! ZEOS 486 Upgradable VESA Local Bus Motherboard.
in...@zeos.com
Ken Germann sup...@zeos.com Zeos International, Ltd.
Marcel Wijkstra (AIO)
#In article <1992Aug19.1...@ilinx.wimsey.bc.ca>, br...@ilinx.wimsey.bc.ca (Brian J. Murrell") writes:
#>
#>ks...@andrew.cmu.edu (Kenton Shaver) writes:
#>
#>That's what I heard. IMHO the 486sx is a *big scam*.
#>
#>To get your 486sx you buy a chip which is a functional 486dx with the
#>co-pro disabled (yes more work than a regular 486dx - yet cheaper!).
But of course the math part does not have to be tested.
By the way, the newest 486SX has no co-processor any more.
#>At some point in the future you need the co-pro, so you buy a 487sx, which
#>is just the regular 486dx with a 487sx stamp on it, which shuts down the
#>486sx crippled chip and takes over the machine running it like the 486dx
#>it should have been!!
#Yes, the 486SX is a HUGE scam. But there is a way around.
#
#Microlab is selling a 486SX-25 that they call an SX/DX board. You can swap the
#SX chip for a DX, change exactly 3 jumpers and a 4-pin chip-looking oscillator
#and go clear up to a DX2-66.
The other way around is the new Intel Overdrive, which boosts performance
of an 486SX/20 up to a higher level than a 486DX/33 (if my memory serves me
well - I read an article about it yesterday).
Marcel.
--
X Marcel Wijkstra AIO (wijk...@fwi.uva.nl)
|X| Faculty of Mathematics and Computer Science
X University of Amsterdam The Netherlands
======Life stinks. Fortunately, I've got a cold.========
Andy Glew
>co-pro disabled (yes more work than a regular 486dx - yet cheaper!).
In case y'all haven't heard, the original i486sx's were, indeed, just
the dx with FP disabled.
But the i486sx's now rolling out of the factory actually have the FP
*removed*. This makes the die smaller, providing higher yield and
lower cost. And they are compatible with the original i486sx's.
Can you say "evolutionary strategy"?
--
Andy Glew, gl...@ichips.intel.com
Intel Corp., M/S JF1-19, 5200 NE Elam Young Pkwy,
Hillsboro, Oregon 97124-6497
This is a private posting; it does not indicate opinions or positions
of Intel Corp.
Intel Inside (tm)
str...@uncavx.unca.edu
> I read it in some magazine, but my ECE
> friend is credulous. Will my 487sx
> just shut the 486sx down and take over?
>
This wasn't a good deal until the 486 DX2 came along. The DX2 is a full
486 chip and replaces the 486SX: either by pulling the old chip and inserting
the new one (some systems) or by plugging the DX2 into a free slot. Yes, it
does shut down the SX chip.
The advantage is that you get a 486 system for about $500 cheaper than a DX2
system, and in 2-3 years, when the chip costs about $100, you can upgrade to a
"real" 486. So, all of a sudden, this is not such a bad deal.
**************************************
** Bob Strauss **
** Ramsey Library **
** UNC-Asheville **
** Asheville, NC 28804 **
** Str...@UNCA.BITNET **
** Str...@uncavx.unca.edu **
**************************************
Andrew Haley
No, you've got out of date info. Originally the 486sx _was_ a 486
with the FPU disabled. AFAIK, Intel now have a 486sx die without the
FPU, which is much cheaper to produce because it is much smaller.
Intel get up my nose too, but this isn't really a scam. I wouldn't
buy one, though.
Andrew.
Doug Mohney
>>To get your 486sx you buy a chip which is a functional 486dx with the
>>co-pro disabled (yes more work than a regular 486dx - yet cheaper!).
>
>In case y'all haven't heard, the original i486sx's were, indeed, just
>the dx with FP disabled.
>
>But the i486sx's now rolling out of the factory actually have the FP
>*removed*. This makes the die smaller, providing higher yield and
>lower cost. And they are compatible with the original i486sx's.
>
>Can you say "evolutionary strategy"?
Can you say competition from AMD & Cyrix? :-)
Support U.N. military force against Serbia
-- > SYS...@CADLAB.ENG.UMD.EDU < --
Patrick Taylor
>... The DX2 is a full
>486 chip and replaces the 486SX: either by pulling the old chip and inserting
>the new one (some systems) or by plugging the DX2 into a free slot. Yes, it
>does shut down the SX chip.
>The advantage is that you get a 486 system for about $500 cheaper than a DX2
>system, and in 2-3 years, when the chip costs about $100, you can upgrade to a
>"real" 486. So, all of a sudden, this is not such a bad deal.
Actually , our friend from Intel can correct me if I'm wrong, but my
understanding is that the device you are describing is the OverDrive chip.
The DX2 does not go in a "free slot", it is pin compatible with a regular
old DX chip. I have manuals from Intel which flat out state this. Document
numbers available on request.
>**************************************
>** Bob Strauss **
>** Ramsey Library **
>** UNC-Asheville **
>** Asheville, NC 28804 **
>** Str...@UNCA.BITNET **
>** Str...@uncavx.unca.edu **
>**************************************
----------------------------------------------------------------------------
"This must be Thursday. I never could get the hang of Thursdays"
- D Adams
- Patrick Taylor
Ericsson Network Systems
exu...@exu.ericsson.se "Don't let the .se fool you"
alternately, exu...@ZGNews.Lonestar.Org
Parag Patel
> Thats why you save your money and buy an overdrive chip instead of the
>487sx. Should plug into the same socket on most so you not only get the full
>486DX, you get it at double the speed.
And possibly melt your motherboard and/or CPU, which is what someone did
who plugged in the 486DX2 into a notebook designed for a 486SX (I
believe it was one of the nice Toshiba notebooks - an expensive lesson).
Check your power supply's rating, what sort of cooling you have, and
most importantly check with the motherboard's manufacturer. Apparently
some of the lighter notebooks use thinner (mylar?) PC boards that can't
quite provide the necessary power safely without melting. Desktop PCs
should probably be OK but check to be sure.
-- Parag
Sean Eric Fagan
>Can you say competition from AMD & Cyrix? :-)
Not Cyrix: they decided to not get into the 486 business.
Or was it C&T who pulled out? Gah. Brain rot.
--
Sean Eric Fagan | "You can't get lost in in one room, no matter how
s...@kithrup.COM | little effort you make to learn your way around."
-----------------+ -- William E Davidsen (wil...@crd.GE.COM)
Any opinions expressed are my own, and generally unpopular with others.
Doug Mohney
>In article <1992Aug20.1...@eng.umd.edu> sys...@king.eng.umd.edu writes:
>>Can you say competition from AMD & Cyrix? :-)
>
>Not Cyrix: they decided to not get into the 486 business.
>
>Or was it C&T who pulled out? Gah. Brain rot.
Actually, C&T jumped, and Cyrix makes juiced-up '386s which they call '486es :)
Jim Harkins
>Not Cyrix: they decided to not get into the 486 business.
>
>Or was it C&T who pulled out? Gah. Brain rot.
Cyrix has the '386 that they call a 486. C&T pulled out of the 486 market.
jim
--
Lawyers Creed: What's mine is mine, what's yours is negotiable.
--------------------------------------------------------------
Jim Harkins [ucsd|uunet]!pacdata!jim
Pacific Data Products j...@pacdata.com
--------------------------------------------------------------
David Willmore
>Actually , our friend from Intel can correct me if I'm wrong, but my
>understanding is that the device you are describing is the OverDrive chip.
>The DX2 does not go in a "free slot", it is pin compatible with a regular
>old DX chip. I have manuals from Intel which flat out state this. Document
>numbers available on request.
> - D Adams
> - Patrick Taylor
Ok, let me see if I have this right. The 486DX has the cpu core with
an fpu and an 8K I/D cache on one chip. The 486SX used to be 486DXs
which had fpu's which failed to pass test, but are now a die that is
specific only to them. The 486DX and 486SX are not pin compatible.
The 486DX2 is an internally clock doubled processor which is otherwise
just like a 486DX of half the processor speed.
Next comes the RapidCad which (I guess) is a 486DX in a 386DX disguise.
My guess is that there is a modified 486DX in the 386DX's place and
some kind of dongle in the 387 socket. I'm not sure on this one at all.
Finally there is the OverDrive chip, which, as I understand (read guess),
is a DX2 in a special socket that is supposed to be included on 486SX
motherboards? Of this last part, I am least sure.
Does anybody know for sure what all these chips are? Someone from Intel?
Dennis, maybe?
--David Willmore
will...@iastate.edu
John F. Woods
>>To get your 486sx you buy a chip which is a functional 486dx with the
>>co-pro disabled (yes more work than a regular 486dx - yet cheaper!).
>In case y'all haven't heard, the original i486sx's were, indeed, just
>the dx with FP disabled.
And, to keep people from arguing over it again (ha), the disabled FP unit
was not tested, increasing yield and decreasing testing time, both factors
making the SX chips somewhat cheaper even from the start.
Brrrrr. Me defending Intel. What is the world coming to? ;-)
Keith Smith
>ka...@cray.com (Kevin Kramer) writes:
>: In article <1992Aug19.1...@ilinx.wimsey.bc.ca> br...@ilinx.wimsey.bc.ca (Brian J. Murrell") writes:
>: >
>: >That's what I heard. IMHO the 486sx is a *big scam*.
>: >
>:
[ 486sx a faulty dx ... ]
>No, you've got out of date info. Originally the 486sx _was_ a 486
>with the FPU disabled. AFAIK, Intel now have a 486sx die without the
>FPU, which is much cheaper to produce because it is much smaller.
>Intel get up my nose too, but this isn't really a scam. I wouldn't
>buy one, though.
>
>Andrew.
Ah, acurate. SCAM, no, not really, you are buying the technology, not
the plastic & silicon. It CAN make good sense. Instead of paying $300
for a full blown 486DX now, buy the $90 SX part, then plug in the 487
when the prices for the 487 come down to a more reasonable level, say
$50 or so in a couple of years. If you don't really need the
Coprocessor horepower now, why not reap the benifits of on-chip cache
and improved instruction execution time vs a similarly priced 386?
Heck, I'm still running a 386. I can't seem to justify the purchase of
a 486xX when I'd never use it.
--
Keith Smith uunet!ksmith!keith 5719 Archer Rd.
Digital Designs BBS 1-919-423-4216 Hope Mills, NC 28348-2201
Somewhere in the Styx of North Carolina ...
Joseph Sarkes
: what made Intel think a new pinout was remotely a good idea, they could
: have just left off the FPU.
:
: DX2 chips use less power than the DX chips of the same (internal) speed,
: because they don't drive the bus as fast. How much less power is
: something someone with the spec sheets will have to provide, but unless
: my memory is completely wrong it's quite a bit.
:
From the 486SX / 487SX intel manual
The 486SX is available in the 196 pin PQTP package which is a surface
mount package(cheaper to make the package and mount to the board I suppose.
The 486 SX is also available in the 168 pin PGA ceramic package like the
regular 486DX, and I suppose that the DX will fill the same socket without
any problem if the PGA socket is used.
The 487SX is available in a 169 pin PGA.
The 486SX has a no connect on the FERR# pin, MP# pin, IGNNE# pin, all of
which are for the numeric coprocessor.
The 487SX has a KEY pin that prevents the chip form being inserted in
the wrong orientation.
From what I can gather, when the 487SX only causes the 486SX to float
the bus and give it to the 487SX instead. Thus the 486SX would still
draw current. However, this does allow a cheap mass produced product
with surface mount technology, and also allow the ability to upgrade
by inserting a new CPU chip that is keyed and has minimum potential
for an unknowledgeable consumer to screw up royally by rotating the
chip and shorting out $$$ worth of his computer.
--
My opinions are of course my own, but are indeed for sale
(to the highest bidder). Don't shoot!! I know secrets!!
Joseph Sarkes (junkyard!joe)
exu...@exu.ericsson.se
>Ok, let me see if I have this right. The 486DX has the cpu core with
>an fpu and an 8K I/D cache on one chip. The 486SX used to be 486DXs
>which had fpu's which failed to pass test, but are now a die that is
>specific only to them. The 486DX and 486SX are not pin compatible.
>The 486DX2 is an internally clock doubled processor which is otherwise
>just like a 486DX of half the processor speed.
>Does anybody know for sure what all these chips are? Someone from Intel?
>Dennis, maybe?
>--David Willmore
>will...@iastate.edu
Like I said, I have the book. Order from Intel (they will send free):
241245-001
"Intel486 DX2 Mircoprocessor Data Book"
This is the one that states
" * Binary Compatible with Large Software Base
...
* High Integration Enables On-Chip
- 8 Kbyte Code and Data Cache
- Floating Point Unit
- Paged, Virtual Memory Management
* 168-Pin Grid Array Package
- Pin Compatible with Intel486 DX Microprocessor
* IEEE 1149.1 Boundary Scan Compatibility
* High Performance Design
- 50 MHz/66 MHz Core Speed Using 25Mhz/33 Mhz Bus Clocks
- RISC Integer Core with Frequent Instructions Executing in
One Core Clock
- 80, 106 Mbyte/sec Burst Bus
- Dynamic Bus Sizing for 8-, 16-, and 32-Bit busses
- Complete 32-Bit Architecture
* Multiprocessor Support
- Cache Consistancy Protocols
- Support for Second Level Cache
"
Wish you could see the block diagram :-)
241257-001
Little sales brochure on DX2's, showing the 50 on the cover.
241254-002
"Intel486 DX2 Microprocessor Performance Brief" <it is not brief>
----------------------------------------------------------------------------
"This must be Thursday. I never could get the hang of Thursdays"
- D Adams
- Patrick Taylor
exu...@exu.ericsson.se
>From: davi...@ariel.crd.GE.COM (william E Davidsen)
>Subject: Re: Does a 487sx shut down the 486sx??
>Date: 24 Aug 92 20:22:05 GMT
>In article <26#n0bc....@netcom.com>, pa...@netcom.com (Parag Patel) writes:
>| m...@wam.umd.edu writes:
>|
>| > Thats why you save your money and buy an overdrive chip instead of the
>| >487sx. Should plug into the same socket on most so you not only get the full
>| >486DX, you get it at double the speed.
>|
>| And possibly melt your motherboard and/or CPU, which is what someone did
>| who plugged in the 486DX2 into a notebook designed for a 486SX (I
>The SX and DX2 are not pin compatible, that may help explain the
>problem. I believe the Overdrive chip /is/ SX pinout. I have no idea
>what made Intel think a new pinout was remotely a good idea, they could
>have just left off the FPU.
It is not a new pinout. The DX2 shares pinouts with the DX, which makes
sense. Think of the Overdrive as an SX2, as far as pinout is concerned.
>DX2 chips use less power than the DX chips of the same (internal) speed,
>because they don't drive the bus as fast. How much less power is
>something someone with the spec sheets will have to provide, but unless
>my memory is completely wrong it's quite a bit.
According to Intel's 241245-001, the Power Supply current is 1200 mA for the
66 MHz DX2 and 950 mA for the 50 MHz DX2.
It's also marked "PRELIMINARY" all over the book.
Also of note is the 50 mA rating for supply current in "Power Down Mode".
It further states this is a target value and has not been tested.
Steve Larsen
Because right now, they are 1/3 the price of DX chips, and if you look
around, you can find motherboards with the extra pin socket so that upgrading
to the DX down the road is as simple as changing the CPU (and possibly the
oscillator, depending on whether or not you change MHz, and whether your
motherboard supports multiple oscillator frequencies).
Steve Larsen
: ks...@andrew.cmu.edu (Kenton Shaver) writes:
:
: >Could anyone clear this up for me?
: >I read it in some magazine, but my ECE
: >friend is credulous. Will my 487sx
: >just shut the 486sx down and take over?
:
: That's what I heard. IMHO the 486sx is a *big scam*.
:
: To get your 486sx you buy a chip which is a functional 486dx with the
: co-pro disabled (yes more work than a regular 486dx - yet cheaper!).
Ah, but there is not more work, actually less. The reason that they
offer this chip for less is that they don't have to spend the man-hours
testing the coprocessor. It is also artificially deflated in price
because Intel wishes to make it the entry level processor and knock
AMD and others out of the water with their 386 chips. However, Intel
has acknowledged at least part of this to the public.
william E Davidsen
| >DX2 chips use less power than the DX chips of the same (internal) speed,
| >because they don't drive the bus as fast. How much less power is
| >something someone with the spec sheets will have to provide, but unless
| >my memory is completely wrong it's quite a bit.
|
| According to Intel's 241245-001, the Power Supply current is 1200 mA for the
| 66 MHz DX2 and 950 mA for the 50 MHz DX2.
Does someone have the figure for 50DX? I believe the 66 draws more
than the 50, DX2 or not, the question in my original post is if the
50DX2 really draws much less than the 50DX.
--
bill davidsen, GE Corp. R&D Center; Box 8; Schenectady NY 12345
I admit that when I was in school I wrote COBOL. But I didn't compile.
Tom Barrett
>
>Ah, but there is not more work, actually less. The reason that they
>offer this chip for less is that they don't have to spend the man-hours
>testing the coprocessor.
Yeah... like maybe 1/10 cent per chip (test vectors are
cheap)
>It is also artificially deflated in price
>because Intel wishes to make it the entry level processor and knock
>AMD and others out of the water with their 386 chips.
*THAT'S* more like it.
--
Tom Barrett (TDBear) tdb...@tandon.com voice 805-378-6207
Tandon Corporation tdb...@p6.f1006.n102.z1.fidonet.org fax 805-529-8895
Sr. HW Design Engineer "War is Peace, No is Yes, And We're All Free!"
[NOTE: The feed to tdb...@tandon.com may be interrupted in Sept. <<<===--]
Andy Glew
>offer this chip for less is that they don't have to spend the man-hours
>testing the coprocessor.
Yeah... like maybe 1/10 cent per chip (test vectors are cheap)
I don't really know if the first poster was correct - the price
difference was probably as much anticipation of the chips that had the
FPU removed, which are cheaper both in yield, area, and testing - but
I would like to correct the second statement.
Test vectors are cheap. But tester time isn't. Testers are often the
manufacturing bottleneck. When this is true, every time unit of tester
time is a time unit lost on the expensive machinery in the rest of the
fab - and that machinery costs out at one hell of a lot more than 1/10
cent per chip.
"Buy more testers" is only part of the answer.
(NB. by "tester" I mean test machinery, not human operators.)
Raul Izahi Lopez Hernandez
> >Ah, but there is not more work, actually less. The reason that they
> >offer this chip for less is that they don't have to spend the man-hours
> >testing the coprocessor.
>
> Yeah... like maybe 1/10 cent per chip (test vectors are cheap)
>
>Test vectors are cheap. But tester time isn't. Testers are often the
>manufacturing bottleneck. When this is true, every time unit of tester
>time is a time unit lost on the expensive machinery in the rest of the
>fab - and that machinery costs out at one hell of a lot more than 1/10
>cent per chip.
>
>"Buy more testers" is only part of the answer.
>
>(NB. by "tester" I mean test machinery, not human operators.)
Good test vectors are NOT cheap, and they allow for shorter tester
usage.
Random, Pseudo-Random, and God forbid, sequential testing of all combinations
might be cheap but consume a lot of tester time.
--
-----> All opinions expressed here are my own, not IBM's <-----
Raul Izahi Lopez Izahi Engineering
iz...@bsc.no IBM Bergen Environmental Sciences and Solutions Centre
Thormoehlensgate 55, 5008 Bergen, NORWAY (47-5)54-4653
Parag Patel
the following. (Please let me know if I'm wrong.)
A 25/50MHz 486DX2 chip (25MHz off-chip clock, 50MHz on-chip) will draw
less power than a 50Mhz 486DX chip, but will drow MORE than a 25MHz
486DX chip. (This applies for the faster DX2 chips as well.)
So plugging in a 25/50MHz DX2 into a board designed for a 25Mhz DX chip
could cause a meltdown, especially in laptops which are much more
sensitive to this sort of thing. (Unless your board is designed for
taking the clock-doubler chips, I wouldn't try plugging one in without
first checking with the manufacturer.)
As far as the 487/Overdrive series of chips go, they'll draw as much
current as the equivalent 486DX, since that's what they really are on
the inside. However, when one of these is plugged in, the 486SX it
overrides is also still plugged in and drawing power. This will waste
power (which will shorten the battery life on portables), but at least
the board is designed for it so it should not be damaged.
-- Parag Patel <pa...@Netcom.COM>
Tom Barrett
>manufacturing bottleneck.
Andy, since you are in a good position to find out... could
you post how long it takes to test the 486DX and how long it
takes for the 486SX? Then could you provide an approximate
number of chips which go through one particular tester each
day (on the average). I know I am curious about this, and I
am sure others are, too.
I truely cannot believe that the 486sx PGA (which probably
still has the numeric) would not be tested thoroughly to
ensure that all of the silicon is working (even though it
won't be used) since there is a slight chance a damaged
numerics area might cause contamination to the rest of the
chip if (say) something overheats.
With the time and chip counts, we can come up with a
reasonable estimate once and for all :)
--
Tom Barrett (TDBear) tdb...@tandon.com voice: 805-378-6207
Tandon Corporation 609 Science Drive Moorpark, CA 93021
Sr. HW Design Engineer "War is Peace, No is Yes, And We're All Free!"
-=>NOTE: The feed to tdb...@tandon.com may be interrupted in Sept.<=-
Russ Schnapp
|> In article <1992Sep4.1...@tandon.com> tdb...@tandon.com (Tom Barrett) writes:
|> >I truely cannot believe that the 486sx PGA (which probably
|> >still has the numeric) would not be tested thoroughly to
|> >ensure that all of the silicon is working (even though it
|> >won't be used) since there is a slight chance a damaged
|> >numerics area might cause contamination to the rest of the
|> >chip if (say) something overheats.
|>
|> the numeric section, therefore it will be untested.
I'm _certain_ that I have read (and re-read, and re-re-read) here that
the 486SX die has not included the numeric coprocessor for many, many
moons. It was there initially, but not anymore!
--
...Russ Schnapp
BIX: rschnapp Email: uunet!metaflow!rschnapp or rsch...@metaflow.com
Metaflow Technologies Voice: 619/452-6608x230; FAX: 619/452-0401
La Jolla, California Unless otw specified, I`m speaking only for myself!
Stanley T.H. Chow
>>Test vectors are cheap. But tester time isn't. Testers are often the
>>manufacturing bottleneck.
>
>Andy, since you are in a good position to find out... could
>you post how long it takes to test the 486DX and how long it
>takes for the 486SX? Then could you provide an approximate
>number of chips which go through one particular tester each
>day (on the average). I know I am curious about this, and I
>am sure others are, too.
This would indeed be interesting. However, I expect the chance
of Andy being allowed to post such numbers is only very slightly
better than the proverbial snowball in hell.
>I truely cannot believe that the 486sx PGA (which probably
>still has the numeric) would not be tested thoroughly to
>ensure that all of the silicon is working (even though it
>won't be used) since there is a slight chance a damaged
>numerics area might cause contamination to the rest of the
>chip if (say) something overheats.
I assume the 486sx die would have at least disabled or disconnected
the numeric section, therefore it will be untested.
--
Stanley Chow InterNet: sc...@BNR.CA
Bell Northern Research UUCP: ..!uunet!bnrgate!bqneh3!schow
(613) 763-2831
Me? Represent other people? Don't make them laugh so hard.
Iskandar Taib
>The SX and DX2 are not pin compatible, that may help explain the
>problem. I believe the Overdrive chip /is/ SX pinout. I have no idea
>what made Intel think a new pinout was remotely a good idea, they could
>have just left off the FPU.
I was my impression that the Overdrive chip (NOT the OEM DX/2) is pin
compatible with the _387SX_, not the 386SX. I think the OEM DX/2 is pin
compatible with the DX.
--
-------------------------------------------------------------------------------
Iskandar Taib | The only thing worse than Peach ala
Internet: NT...@SILVER.UCS.INDIANA.EDU | Frog is Frog ala Peach
Bitnet: NTAIB@IUBACS !
Andy Glew
>Test vectors are cheap. But tester time isn't. Testers are often the
>manufacturing bottleneck.
Andy, since you are in a good position to find out... could
you post how long it takes to test the 486DX and how long it
takes for the 486SX? Then could you provide an approximate
number of chips which go through one particular tester each
day (on the average). I know I am curious about this, and I
am sure others are, too.
You're testing me, right?
Seriously, I very much doubt that I could ever post such information.
The microarchitecture of the chip that I am working on is moderately
interesting and relatively secret --- but details about process, time
per stage, and yields are EXTREMELY secret. Process details would
allow a competitor to estimate cost of manufacture, and hence
anticipate our pricing strategies. I think that you will observe that
all papers on such manufacturing details have graphs expressed in
relative numbers, not absolute values.
I truely cannot believe that the 486sx PGA (which probably
still has the numeric) would not be tested thoroughly to
ensure that all of the silicon is working (even though it
won't be used) since there is a slight chance a damaged
numerics area might cause contamination to the rest of the
chip if (say) something overheats.
I am not 100% certain, but I am reasonably sure that the original
i486sx was *not* a CPU that passed integer tests but failed FP tests.
The present i486sx, of course, doesn't need to pass FP tests because
it has no FP on chip.
There would be far too much risk involved in chips that failed any
part of their functionality tests. Imagine that FP tests failed
because of an intermittent short in the FP side - a short that would
eventually close, even with integer only use...
william E Davidsen
| I am not 100% certain, but I am reasonably sure that the original
| i486sx was *not* a CPU that passed integer tests but failed FP tests.
|
| The present i486sx, of course, doesn't need to pass FP tests because
| it has no FP on chip.
|
| There would be far too much risk involved in chips that failed any
| part of their functionality tests. Imagine that FP tests failed
| because of an intermittent short in the FP side - a short that would
| eventually close, even with integer only use...
This brings up a good technical point, how do you properly test a chip
with a castrated FPU? If you test with the FPU enabled and then disable
it you might cause a problem in the disable process (unless you do it by
mounting with a pin tied high of some such). If you disable and then
test, you might miss a problem such as you describe above.
If you test before and after the cost of manufacture goes up on the
chip you seel for less (that wouldn't astonish me from intel).
REE...@maine.maine.edu
are even sawn, using probe stations. Devices that only fail FPU can have
selected interconnects laser-vaporized and will become SX's. Those that
pass everything will become DX's and those that fail everything will become
keychains!
I am not certain that Intel uses wafer probe, but I would be amazed if
they wasted the effort of packaging and bonding bad die! Once the faulty
subsection was disconnected (electrically) who cares what goes on within it?
It would not be connected to any power or signal traces, in effect it is
just dead space.
Jeff Andle
Richard Lee
If I under stood it correctly. Intel used the faultly DX chips for SX in the
begining. Then they started producing straight SX chips... but who knows
what intel does...really.
--
| Rich Lee | Richard Lee | |
| My Power is | 1711 Whipple Dr. | rl...@csugrad.cs.vt.edu |
| Beyond Your | Apt 11 | or |
| Understanding | Blacksburg,Va 24060 | rl...@earth.tjhsst.edu |
Mark Gilbert
Perhaps I diverge, taking this thread back a step, but I think that you are
making a bad assumption that the cost of a product depends primarily on the
cost of manufacturing. Companies routinely adjust prices for a variety of
reasons, including: relationship to other products in a product line,
perceived value of the product by the consumer, and the products and prices
of competitors.
In the case of the 486SX vs. 486DX, Intel sets the DX price high and the SX
price low to establish a product line. The DX is the premium product and
commands a premium price, producing a higher margin. The lack of
competition for the DX makes this possible. Competition from the 386 keeps
SX prices low, and Intel wants to compete successfully against the 386 to
promote the perceived value of the 486.
Mark Gilbert