[gmx-users] Range checking error

[Yanzi Zhou]

Thank you.
I recompiled the software with gcc, but my job still crashed at about 
170ps. I tried 1 ns simulation for 1PGB in water downloaded from 
http://www.dddc.ac.cn/embo04/practicals/9_14_1.htm, it can be finished. 
But when I tried to do NTP simulation of DPPC from the benchmarks of 
Gromacs, my job died on 6 processes with double precision gromacs.  I 
found the job crashed when I:
do NTP simulation for a large system, and
use PME
on more than two processors with double precision gromacs.


David van der Spoel wrote:
> Yanzi Zhou wrote:
>> Dear Gromacs-Users,
>>
>> Can anybody help me?
>>
>> When I do 1 ns NTP MD simulation using version 3.3.2 on 4 or 6 processes
>> with double precision gromacs compiled by Intel compiler on x86_64
>> GNU/Linux, it was crashing at about 100 ps with a message:
>>
>> -------------------------------------------------------
>>
>> Program mdrun_mpi_d, VERSION 3.3.2
>>
>> Source code file: nsgrid.c, line: 226
>>
>> Range checking error:
>>
>> Explanation: During neighborsearching, we assign each particle to a grid
>>
>> based on its coordinates. If your system contains collisions or 
>> parameter
>>
>> errors that give particles very high velocities you might end up with 
>> some
>>
>> coordinates being +-Infinity or NaN (not-a-number). Obviously, we cannot
>>
>> put these on a grid, so this is usually where we detect those errors.
>>
>> Make sure your system is properly energy-minimized and that the 
>> potential
>>
>> energy seems reasonable before trying again.
>>
>> Variable ci has value 4421. It should have been within [ 0 .. 4212 ]
>>
>> -------------------------------------------------------
>>
>>
>> But it can be finished when I use 1 or 2 processes, or use cut-off for
>> electrostatics instead of PME, or for a NVT system, or with single
>> precision gromacs. I have noticed the bug #167 on Bugzilla, and modified
>> routine stat.c, but it did no help, and actually I didn't write to XTC
>> file. Maybe, there are still bugs for double precision gromacs?
>
> More likely the compiler. Please retry with gcc.
>
>>
>>
>> Thanks for any help.
>>
>> _______________________________________________
>> gmx-users mailing list    gmx-users at gromacs.org
>> http://www.gromacs.org/mailman/listinfo/gmx-users
>> Please search the archive at http://www.gromacs.org/search before 
>> posting!
>> Please don't post (un)subscribe requests to the list. Use the www 
>> interface or send it to gmx-users-request at gromacs.org.
>> Can't post? Read http://www.gromacs.org/mailing_lists/users.php
>
>