Ah, I just read the rest of the email, so it appears that HPC support staff
have built the latest version with the newest bug fixes. It may be helpful
to install your own version to help hunt this down, so you can compile as
you need to (i.e. with debug symbols, -g)
Good luck!
Jason
On Tue, Nov 9, 2010 at 6:06 PM, Jason Swails <jason.swails.gmail.com> wrote:
> I would suggest recompiling after applying all of the bug fixes if you
> haven't already. This may be an issue that was fixed already. If you patch
> amber10 and recompile (don't forget to make clean before trying to
> recompile) and the issue remains, it may be worth running through a debugger
> (i.e. compile with -g and use gdb/idb). divide-by-zero errors are typically
> relatively easy to find using debuggers, in my experience.
>
> If none of this helps, the best option would be to create a reproducible
> failure that developers can attack, but until the problem can be reproduced
> by others, it can't really be addressed by them.
>
> Hope this helps,
> Jason
>
>
> On Tue, Nov 9, 2010 at 5:15 PM, Xingqing Xiao <xxiao3.ncsu.edu> wrote:
>
>> Dear,
>> My jobs are no problems to be running in the parallel verison of
>> AMBER10 when using sander.MPI. But it fails in the serial version when
>> using sander.
>> Error message is as follows:
>> *************************** current error ***********************
>> forrtl: severe (71): integer divide by zero
>> Image PC Routine Line Source
>> libvml.so 00002B9652B431DA Unknown Unknown Unknown
>> *****************************************************************
>> I can use it all the time. But these days ago, it failed. I don't know
>> what is wrong with AMBER10. Could you please give me some advices?
>> Attached please find our dicussions about AMBER problems. Maybe you
>> can find some useful information. Thanks a billion!
>> Best,
>> Xiao
>>
>>
>>
>>
>> I'm not sure what library files you are referring to?
>> Only library Amber linked to other than code supplied
>> by Amber is the Intel Math Kernel Library (MKL).
>>
>> MKL is used by quite a few applications and
>> individual's codes and we have not had other
>> complaints related to it.
>>
>> I expect that contacting Amber will be a much more
>> authoritative source of information than searching
>> the internet.
>>
>> -Eric
>>
>> On Nov 9, 2010, at 3:43 PM, Xingqing Xiao wrote:
>>
>> > Dear Dr. Sills,
>> > No, i haven't. Because i am not sure where are the problems, or Amber
>> > itself, or the library file in the hpc? I will figure it out again on
>> > the internet. Any progress about it, i will share with you. Thanks
>> > very much!
>> > Best,
>> > Xiao
>> >
>> >
>> >> The latest patch from Amber has been applied (last week).
>> >>
>> >> Have you reported the problem to Amber developers?
>>
>> _______________________________________________
>> AMBER mailing list
>> AMBER.ambermd.org
>> http://lists.ambermd.org/mailman/listinfo/amber
>>
>
>
>
> --
> Jason M. Swails
> Quantum Theory Project,
> University of Florida
> Ph.D. Graduate Student
> 352-392-4032
>
--
Jason M. Swails
Quantum Theory Project,
University of Florida
Ph.D. Graduate Student
352-392-4032
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Nov 09 2010 - 15:30:06 PST
