From: mlknol@cs.vu.nl (Marcel Knol) Subject: [SUMMARY] GCC Internal Compiler error, the answers Date: Fri, 28 May 1993 19:30:47 GMT
Hello,
Around two weeks ago, I posted a request about an internal compiler
errors, generated by gcc. I asked if this was caused by a software or hardware
failure. I got a really great response and many people asked me to post
a summary on c.o.l. So here it goes:
- First the correct answer to my problem, because all of you are of course
desperate to know the answer :-) :
A lot of people suggested to turn the external cache off. Indeed, when I turned
the external cache off, the problem was gone. However I was not satisfied
that I had to run my expensive 486 without the external cache. Two friends
of mine bought the same computer around the same time, and they didn't have
this problem. So I went back to the shop, and they replaced the mainboard,
memory and cache all in once. From than on, no more problems and I am a
happy Linux User. What a GREAT Operating System !! Thanks to Linus and all
other people who developed it.
- Several people asked me if I was using a recent version of gcc, had
enough memory and the good versions of the libs. These are good remarks
but didn't apply to me after all. Maybe it applies to some other people.
By the way, I use 8 MB RAM and 16 MB swap. This should be enough.
- A few people argued the problem was caused by an old mainboard that
couldn't handle the 486 specific compiler options which were used to
compile the kernel. They suggested I turn the external cache off,
compile the kernel WITHOUT 486 specific options and try again with cache
on. Because the hardware was already replaced, I couldn't try it.
- A lot of people told me the cache-chips were to blame, they would be to
slow or instable. Because of the fact that everything was replaced
together, I can't verify this, but it seems very reasonably to me.
- The temperature of the system could also be of some effect, one person
had troubles during hot days. During my problem, the temperature here
in The Netherlands was around 20 degrees Celcius, that isn't too warm.
Besides I have a Full Tower and the problem appeared also right after
I turned it on. However if anyone has this problems only after some time,
this may be the answer.
- The internal compiler could also be caused by lose SIMM or not enough
wait-states. I have tried to push the SIMM through the board :-), but
this didn't work for me.
I think that everyone who is disabling his/her cache is using a mainboard,
or cache-chip which are defective. Maybe some of the above solutions can
help you out.
I was very pleased with the number of reactions I got. Thank you all very
much. By the way, it is amazing if you compare the quality of the software
for Linux (and Unix of course) and the quality of the software for MS-DOS,
MS-Windows and OS/2. These commercial company can't get it rights it seems,
and some people just doing it for fun, create a wonderful operating system.
Kind regards,
Marcel Knol
-- Marcel Knol, <mlknol@cs.vu.nl> Vrije Universiteit Amsterdam, Department of mathematics and computer science, The Netherlands.