Can I disable the 387 temporarily

Can I disable the 387 temporarily

Post by Joel M. Hoffm » Sun, 17 Apr 1994 21:28:48



I'm trying to determine why simple programs keep crashing, and one
possibility is that my '387 chip is buggy.  But, of course, I can't
simply take it off the motherboard without tools I don't own.  Is
there any way to disable it by s/w?  I have emulation compiled into my
kernel, and I want to see if the program will crash Linux even with
the '387 disabled.  Failing that, is there any way to tell GCC not to
include any 387 instructions in it's compiled code?  I suppose this
would involve linking with an emulation library.  Is there one?

Thanks.

-Joel

--
-----------------------------------------------------------------------------
|_|~~ Germany, Europe. 1943.    "The diameter of the bomb was 30 centimeters,
__|~| 16 Million DEAD.           and the diameter of its destruction, about 7
                                meters, and in it four killed and 11 wounded.
 cnc  Bosnia, Europe. 1993.     And around these, in a larger circle of  pain
 cnc  HOW MANY MORE?          and time,  are scattered two  hospitals and one
                          cemetery.   But the young woman who was  buried  in
                    the place from where she came, at a distance of more than
             than 100 kilometers, enlarges the circle considerably.   And the
      lonely man who is mourning her death in a distant  country incorporates
into the circle the whole world.  And I won't speak of the cry of the orphans
that reaches God's chair and from there makes the circle endless and godless."
-----------------------------------------------------------------------------

 
 
 

Can I disable the 387 temporarily

Post by David F » Mon, 18 Apr 1994 01:58:43



] the '387 disabled.  Failing that, is there any way to tell GCC not to
] include any 387 instructions in it's compiled code?  I suppose this
] would involve linking with an emulation library.  Is there one?

Have you looked in the gcc documentation?

 
 
 

Can I disable the 387 temporarily

Post by Gregory McKes » Wed, 20 Apr 1994 10:55:28


    Joel> I'm trying to determine why simple programs keep crashing,
    Joel> and one possibility is that my '387 chip is buggy.  But, of
    Joel> course, I can't simply take it off the motherboard without
    Joel> tools I don't own.  Is there any way to disable it by s/w?
    Joel> I have emulation compiled into my kernel, and I want to see
    Joel> if the program will crash Linux even with the '387 disabled.
    Joel> Failing that, is there any way to tell GCC not to include
    Joel> any 387 instructions in it's compiled code?  I suppose this
    Joel> would involve linking with an emulation library.  Is there
    Joel> one?

You can boot Linux in such a way as to ignore your FPU if you boot
with LILO.  Just pass the no387 at the boot prompt.

ie:

boot: vmlinuz no387

To compile with FPU emulation you can use the gcc option -msoft-float.

Greg.

--
____________________________________________________________________
Gregory McKesey (Software Manager)      Gallium Software Inc.
Tel: (613)721-0902 ext (431)            303 Moodie Dr., Suite 4000
Fax: (613)721-1278                      Nepean, Ontario, Canada.

--------------------------------------------------------------------

 
 
 

Can I disable the 387 temporarily

Post by Beeblebr » Wed, 20 Apr 1994 13:17:21




>] the '387 disabled.  Failing that, is there any way to tell GCC not to
>] include any 387 instructions in it's compiled code?  I suppose this
>] would involve linking with an emulation library.  Is there one?
>Have you looked in the gcc documentation?

You _know_ he hasn't.

To answer all of his questions, he'll need to look at the Lilo & kernel docs
also.
___

C++ consultant and emacs support. Mail me if you have any problems.

 
 
 

Can I disable the 387 temporarily

Post by Joel M. Hoffm » Wed, 20 Apr 1994 22:48:56


Quote:>    Joel> I'm trying to determine why simple programs keep crashing,
>    Joel> and one possibility is that my '387 chip is buggy.  [...]

>[...]
>To compile with FPU emulation you can use the gcc option -msoft-float.

First off, thanks to all who replied about LILO, and the no387 option.
It was exactly what I needed, and proved that my '387 has been causing
the crashes that have plagued me for so long.  I have two questions:
first, is using "-msoft-float" faster than kernel emulation?  I would
think it is.

Secondly:  I'm going to try to return my '387, but I know I'm going to
have a hard time doing so.  Does anyone have any experience with this?
The '387 has "ULSI" on the chip.  I guess that's the manufacturer.
I'm going to try returning the chip to the store where I bought it.
If that doesn't work (and I doubt it will) does anyone know how to
contact ULSI?

The approach I'm going to try is claiming that the store and ULSI
violated truth in advertising laws when they claimed the chip was
compatible with the Intel '387.  Anyone have any better suggestions?
While we're at it, does anyone have any further documentation that
ULSI chips are buggy?

Thanks!

-Joel

--
-----------------------------------------------------------------------------
|_|~~ Germany, Europe. 1943.    "The diameter of the bomb was 30 centimeters,
__|~| 16 Million DEAD.           and the diameter of its destruction, about 7
                                meters, and in it four killed and 11 wounded.
 cnc  Bosnia, Europe. 1993.     And around these, in a larger circle of  pain
 cnc  HOW MANY MORE?          and time,  are scattered two  hospitals and one
                          cemetery.   But the young woman who was  buried  in
                    the place from where she came, at a distance of more than
             than 100 kilometers, enlarges the circle considerably.   And the
      lonely man who is mourning her death in a distant  country incorporates
into the circle the whole world.  And I won't speak of the cry of the orphans
that reaches God's chair and from there makes the circle endless and godless."
-----------------------------------------------------------------------------

 
 
 

Can I disable the 387 temporarily

Post by dglid.. » Fri, 22 Apr 1994 01:45:18



: violated truth in advertising laws when they claimed the chip was
: compatible with the Intel '387.  Anyone have any better suggestions?
: While we're at it, does anyone have any further documentation that
: ULSI chips are buggy?

FWIW - a friend of mine has been running Linux on his 386/40 w/ULSI 387/40
for several months now, doing all manner of "goofy" stuff that one would
typically do under Linux when you play around with everything (i.e.
recompile kernels, run X, ghostscript, xv, lots 'o communication stuff
via term & etc, Mosaic....) with no problems whatsoever.

IMHO you've just got a bad FPU, rather than ULSI having serious problems.

--


GAT c++(++++) -d+(?) e+(*) f+ g+(+++) h+ l++ m+/---
    n-- -p+ r s+/- t++ u++ w+++

 
 
 

Can I disable the 387 temporarily

Post by David Soul » Fri, 22 Apr 1994 23:39:06


I also have a 386sx-40 with a ULSI 387/40.

Not a problem, and Im running UMSDOS 1.0 with an old NCR 7920 serial terminal as
a secondary from ttyS0.

If anyone needs help configuring UMSDOS, I have successfully re-installed it
about 9 times now (grin) and have become very familiar with it (out of
necessity! (Grin))

 
 
 

1. Can I disable 387 at lilo prompt on install?

I believe that problems I am haveing installing linux can be
traced to my math coprocessor. I have an Intel RapidCad chipset.

Is there a way to disable the math coprocessor at the lilo boot
prompt? It has been suggested that I use:

        ramdisk no387

This does not work and I still get an error that 386/387 coupling
failed.

Thanks

2. Reading is much more interesting than TV (0106/1708)

3. 387 problem

4. Matrox driver for Linux

5. Linux finds a nonexistant 387

6. ipchains and SMTP help

7. Anyone using a Cyrix 387 clone out there?

8. Telephone dialer for Linux?

9. 386/387 coupling ...

10. IIT 387 problems

11. PROBLEM WITH COPROCESSOR 387 AND IRQ13.....HELP!

12. Cyrix 387 error reporting broken

13. pl14 thinks I have a 387