26 Feb
2003
26 Feb
'03
8:45 a.m.
Hi all,
I'm trying to track down a 387 math coprocessor IC for an old 386-based
Linux box that's going to be doing a bit of numbercrunching for me. Has
anyone here got an AMD or Intel (AMD preferred) 387 coprocessor rated at
40MHz (-40 part number suffix) that would work correctly with an AMD
Am386DX-40? No, before you ask, the 386DX does *not* have a built in
mathco - the 486DX was (IIRC) the first DX-series chip with a built-in
coprocessor.
Thanks.
--
Phil.
philpem(a)dsl.pipex.com
http://www.philpem.dsl.pipex.com/
26 Feb
26 Feb
9:02 a.m.
At 10:43 PM 2/26/03 -0000, you wrote:
Hi all,
I'm trying to track down a 387 math coprocessor IC for an old 386-based
Linux box that's going to be doing a bit of numbercrunching for me. Has
anyone here got an AMD or Intel (AMD preferred) 387 coprocessor rated at
40MHz
40 MHz? I didn't think they made them over 25 Mhz. I have a couple around here but
I'll have to dig them out. However I'm pretty sure that they're 20 and 25
Mhz.
Joe
(-40 part number suffix) that would work correctly with an AMD
Am386DX-40? No, before you ask, the 386DX does *not*
have a built in
mathco - the 486DX was (IIRC) the first DX-series chip with a built-in
coprocessor.
Thanks.
--
Phil.
philpem(a)dsl.pipex.com
http://www.philpem.dsl.pipex.com/
9:36 a.m.
On Wed, 26 Feb 2003, Joe wrote:
Back in the day, Intel made 386DX chips that ran up to 33MHz. AMD came
out with a clone chip in the 1990-1991 time frame that was a 386DX that
ran at 40MHz. I ran a 33MHz rated ITT 387 clone mathco in mine without
problems. IIRC, the ITT mathco didn't perform quite as well as a real
Intel one, but it was better than not having anything.
I know Intel still has a rather large 80386 embedded processor market,
so I'd not be surprised if they're making much faster versions of it now.
... But apparently they aren't:
<http://www.intel.com/design/intarch/intel386/index.htm>
-brian.
At 10:43 PM 2/26/03 -0000, you wrote:
I'm
trying to track down a 387 math coprocessor IC for an old 386-based
Linux box that's going to be doing a bit of numbercrunching for me. Has
anyone here got an AMD or Intel (AMD preferred) 387 coprocessor rated at
40MHz
40 MHz? I didn't think they made them over 25 Mhz. I have a couple
around here but I'll have to dig them out. However I'm pretty sure
that they're 20 and 25 Mhz.
5:52 p.m.
Brian Chase wrote:
Back in the day, Intel made 386DX chips that ran up to
33MHz. AMD
came out with a clone chip in the 1990-1991 time frame that was a
386DX that ran at 40MHz.
That's the sort of thing I'm after - the 386DX CPU
is currently clocked at
40MHz, I need a math coprocessor for the board that can handle the same
speed. Otherwise I'll have to try and find out what some of the jumpers on
the board do. Which may be a bit difficult considering it has no
manufacturer/model number branding on it.
Later.
--
Phil.
philpem(a)dsl.pipex.com
http://www.philpem.dsl.pipex.com/
27 Feb
27 Feb
2:08 p.m.
On Wed, 26 Feb 2003, Joe wrote:
They most certainly did come in 40MHz. There's one in a machine of mine.
It isn't Intel or AMD though. I'm pretty sure it's made by Cyrix.
Peace... Sridhar
I'm
trying to track down a 387 math coprocessor IC for an old
386-based Linux box that's going to be doing a bit of number crunching
for me. Has anyone here got an AMD or Intel (AMD preferred) 387
coprocessor rated at 40MHz
40 MHz? I didn't think they made them over 25 Mhz. I have a couple
around here but I'll have to dig them out. However I'm pretty sure that
they're 20 and 25 Mhz. 
26 Feb
26 Feb
10:51 a.m.
New subject: AMD or Intel 80387 Math Coprocessor IC
----- Original Message -----
From: "Philip Pemberton" <philpem(a)dsl.pipex.com>
To: <cctalk(a)classiccmp.org>
Sent: Wednesday, February 26, 2003 5:43 PM
Subject: WTD: AMD or Intel 80387 Math Coprocessor IC
Hi all,
I'm trying to track down a 387 math coprocessor IC for an old
386-based
Linux box that's going to be doing a bit of
numbercrunching for me. Has
anyone here got an AMD or Intel (AMD preferred) 387 coprocessor rated at
40MHz (-40 part number suffix) that would work correctly with an AMD
Am386DX-40? No, before you ask, the 386DX does *not* have a built in
mathco - the 486DX was (IIRC) the first DX-series chip with a built-in
coprocessor.
They come up on epay all the time, pick yer flavor. I just got a 40MHz
Cyrix Fasmath chip for $2 (includes $1 shipping) a week ago. Supposedly
clock for clock it's the fastest of the '387 FPUs, but IDK. Popped right in
to to my P70.
I'm still looking for a set of eight 2MB 72 pin IBM SIMMs with presence
detect, and a Cyrix 486Drx2, which is a 486 with a '386 pinout, if anyone
has a spare.
Thanks.
--
Phil.
Bob
12:41 p.m.
New subject: AMD or Intel 80387 Math Coprocessor IC
There are more then 1 variety of 40Mhz math coprocessors. Back when my main
computer was a 386 in the early 90's I purchased the Cyrix version for my
AMD 386/40DX and it crunched nubers very well (funny how cyrix went from the
fastest FPU unit in the 386/486 days to the WORST during the pentium period
when it mattered).
I currently have a 386dx/40 and a math coprosessor but I dont think its a
cyrix (using it for a dos game rig) Did IIT make them also?
----- Original Message -----
From: "Robert F. Schaefer" <rschaefe(a)gcfn.org>
To: <cctalk(a)classiccmp.org>
Sent: Wednesday, February 26, 2003 7:48 PM
Subject: Re: AMD or Intel 80387 Math Coprocessor IC
----- Original Message -----
From: "Philip Pemberton" <philpem(a)dsl.pipex.com>
To: <cctalk(a)classiccmp.org>
Sent: Wednesday, February 26, 2003 5:43 PM
Subject: WTD: AMD or Intel 80387 Math Coprocessor IC
in
Hi all,
I'm trying to track down a 387 math coprocessor IC for an old
386-based
Linux box that's going to be doing a bit of
numbercrunching for me. Has
anyone here got an AMD or Intel (AMD preferred) 387 coprocessor rated at
40MHz (-40 part number suffix) that would work correctly with an AMD
Am386DX-40? No, before you ask, the 386DX does *not* have a built in
mathco - the 486DX was (IIRC) the first DX-series chip with a built-in
coprocessor.
They come up on epay all the time, pick yer flavor. I just got a 40MHz
Cyrix Fasmath chip for $2 (includes $1 shipping) a week ago. Supposedly
clock for clock it's the fastest of the '387 FPUs, but IDK. Popped right
to to my P70.
I'm still looking for a set of eight 2MB 72 pin IBM SIMMs with presence
detect, and a Cyrix 486Drx2, which is a 486 with a '386 pinout, if anyone
has a spare.
Thanks.
--
Phil.
Bob 
12:43 p.m.
New subject: AMD or Intel 80387 Math Coprocessor IC
I'm
trying to track down a 387 math coprocessor IC for an old
386-based
Linux box that's going to be doing a bit of
numbercrunching for me. Has
anyone here got an AMD or Intel (AMD preferred) 387 coprocessor rated at
40MHz (-40 part number suffix) that would work correctly with an AMD
Am386DX-40? No, before you ask, the 386DX does *not* have a built in
mathco - the 486DX was (IIRC) the first DX-series chip with a built-in
coprocessor.
They come up on epay all the time, pick yer flavor. I just got a 40MHz
Cyrix Fasmath chip for $2 (includes $1 shipping) a week ago. Supposedly
clock for clock it's the fastest of the '387 FPUs, but IDK. Popped right in
to to my P70. I'm still looking for a set of eight 2MB 72 pin
IBM SIMMs with presence
detect, and a Cyrix 486Drx2, which is a 486 with a '386 pinout, if anyone
has a spare.
It's not quite a 486. It actually is a double clocked 386 with a small
internal L1 cache. It doesn't share most of the instructions unique to the 486.
I don't currently have a DRX2, but I do have a DRX, which is essentially the
same thing without the clock doubled core. The L1 cache made it about 20%
faster than an equivalently clocked 386. You had to explicitly turn on the
cache, and set regions as uncachable. (Caching the video card usually made
for problems.)
Eric
27 Feb
27 Feb
9:54 p.m.
New subject: AMD or Intel 80387 Math Coprocessor IC
----- Original Message -----
From: "Eric J. Korpela" <korpela(a)ssl.berkeley.edu>
To: <cctech(a)classiccmp.org>
Cc: <cctalk(a)classiccmp.org>
Sent: Wednesday, February 26, 2003 9:38 PM
Subject: Re: AMD or Intel 80387 Math Coprocessor IC
instruction
the
486.
Ok. I thought it was a 486.
> They come up on epay all the time, pick yer
flavor. I just got a 40MHz
> Cyrix Fasmath chip for $2 (includes $1 shipping) a week ago. Supposedly
> clock for clock it's the fastest of the '387 FPUs, but IDK. Popped
right in
to to my P70.
It was a tiny bit faster than the AMD and Intel parts depending upon mix.
Ah. Was it a 100% clone, or were there extra added instructions?
> I'm still looking for a set of eight 2MB 72 pin IBM SIMMs with presence
> detect, and a Cyrix 486Drx2, which is a 486 with a '386 pinout, if
anyone
has a spare.
It's not quite a 486. It actually is a double clocked 386 with a small
internal L1 cache. It doesn't share most of the instructions unique to
I don't currently have a DRX2, but I do have a DRX, which is essentially
the
same thing without the clock doubled core. The L1
cache made it about 20%
faster than an equivalently clocked 386. You had to explicitly turn on
the
cache, and set regions as uncachable. (Caching the
video card usually
made
for problems.)
Guess I want the DRX2 then. No sense in spending all that time on something
just to have it be less than optimal. I'll have to dig up reference sheets
on 'em.
BTW, what do you run on the DRX? I'm upgrading an IBM P70 luggable, the
target OSes are AIX-PS/2 and Solaris in particular, and then as many others
as will work.
Eric
Bob
6 Mar
6 Mar
12:10 p.m.
New subject: AMD or Intel 80387 Math Coprocessor IC
I don't
currently have a DRX2, but I do have a DRX, which is essentially
the
same thing without the clock doubled core. The
L1 cache made it about 20%
faster than an equivalently clocked 386. You had to explicitly turn on
the
cache, and set regions as uncachable. (Caching
the video card usually
made
for problems.)
Guess I want the DRX2 then. No sense in spending all that time on something
just to have it be less than optimal. I'll have to dig up reference sheets
on 'em.
BTW, what do you run on the DRX? I'm upgrading an IBM P70 luggable, the
target OSes are AIX-PS/2 and Solaris in particular, and then as many others
as will work.
12:38 p.m.
New subject: AMD or Intel 80387 Math Coprocessor IC
----- Original Message -----
From: "Eric J. Korpela" <korpela(a)ssl.berkeley.edu>
To: <cctech(a)classiccmp.org>
Sent: Thursday, March 06, 2003 9:07 PM
Subject: Re: AMD or Intel 80387 Math Coprocessor IC
Sorry for the delayed response. I've used Linux,
OS/2 Warp and Win3.1 on
them.
Right now it's an OpenDOS machine.
Ok. I was mainly looking to see if it was only kind-of compatable, but it
sounds like any quirks are minor or non-intrusive.
Turning on internal cache requires ring 0 access. OS/2 required a device
driver
at
startup. Win 3.1 and DOS run a program from
autoexec.bat. ISTR that the
linux
kernel had support built in for the DRX. I don't
know about AIX and
Solaris.
Probably need a kernel patch to tweak the bits, same as linux. I wonder if
NetBSD supports the cache...
There are also later model BIOS versions that turn on the cache at boot.
I
don't know if the P70 would have that l turn on
the cache at boot. I
don't
know if a P70 would have an appropriate BIOS.
Probably not, but if my googling has been any indication, there are a goodly
number of drx/drx2 chips installed in P70s-- it's about the fastest chip you
can install without major hardware voodoo or adding a CPU card.
Eric
Bob
26 Feb
26 Feb
5:18 p.m.
Hi Phil,
There may well be a working AMD DX-40 with 387 machine in my Parent's shed -
where abouts are you based?
Tim.
I'm trying to track down a 387 math coprocessor
IC for an old 386-based
Linux box that's going to be doing a bit of numbercrunching for me. Has
anyone here got an AMD or Intel (AMD preferred) 387 coprocessor rated at
40MHz (-40 part number suffix) that would work correctly with an AMD
Am386DX-40? No, before you ask, the 386DX does *not* have a built in
mathco - the 486DX was (IIRC) the first DX-series chip with a built-in
coprocessor.
Sent using the Entourage X Test Drive.
7962
days inactive
7970
days old
11 comments
8 participants
participants (8)
-
korpela＠ssl.berkeley.edu -
philpem＠dsl.pipex.com -
rigdonj＠cfl.rr.com -
rschaefe＠gcfn.org -
teoz＠neo.rr.com -
tim.myers＠sunplan.com -
vance＠neurotica.com -
vaxzilla＠jarai.org