AMBER Archive (2006)

Subject: RE: AMBER: Amber9 serial installation -tests- severe (174): SIGSEGV, segmentation fault occurred

From: Ross Walker (ross_at_rosswalker.co.uk)
Date: Tue Oct 24 2006 - 10:33:45 CDT


Hi Mina,

> the problem is
> still showing up then it needs looking into."
>
> Well, I did as you suggested but I get the same results. When
> I use the gcc and g++ compilers though (instead of the icc
> and icpc that I was replacing manually in the config.h) the
> differences in the TEST_DIFF files are only roundoff
> errors..So, I would guess that the problem is again the
> compiler, or not?

Ah yes, this would explain why we haven't seen this problem before. Almost
everybody uses gcc to compile the c code in Amber on Linux boxes and so this
problem hasn't cropped up before (or it did and the person running the tests
didn't notice [or didn't run the tests]). We should probably try to track
this problem down and see where it is coming from. Perhaps Junmei can try it
out and see if he can find the problem.

For the time being I think the advice should be to stick with the default
gcc compiler that the configure script gives you.
 
> PS:At some point I thought that the problem might be caused
> because in the $AMBERHOME/src/netcdf/src there is are config
> files that use gcc, while in the $AMBERHOME/src configure
> file I had made the replacemnets from gcc to icc and g++ to
> icpc. So I made the replacements there (although there is a
> possibility I was not that accurate) but this didn't work either.

This is all created by netcdf's configure script that is run automatically
when you run Amber's configure. It tracks down what is the default compiler
on your system and uses that. If it works I wouldn't touch it. Besides
changing this co icc and icpc should have only affected the runs that access
binary files, so ptraj and sander. And then only a couple of test cases.

I would just use the default options that configure gives you and stick with
that. If that all works now (after the bugfix) then you are good to go.

I will keep my eye on the antechamber problem though as it is always
possible that a specific compiler is uncovering a bug in the code.

All the best
Ross

/\
\/
|\oss Walker

| HPC Consultant and Staff Scientist |
| San Diego Supercomputer Center |
| Tel: +1 858 822 0854 | EMail:- ross_at_rosswalker.co.uk |
| http://www.rosswalker.co.uk | PGP Key available on request |

Note: Electronic Mail is not secure, has no guarantee of delivery, may not
be read every day, and should not be used for urgent or sensitive issues.

-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu