[gmx-users] error while running mdrun

[Justin Lemkul]

On 12/17/12 8:47 PM, Christopher Samuel wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> On 17/12/12 21:57, ananyachatterjee wrote:
>
>> [ganga:04205] *** Process received signal *** [ganga:04205] Signal:
>> Segmentation fault (11) [ganga:04205] Signal code: Address not
>> mapped (1) [ganga:04205] Failing at address: 0x201a01c20
>
> Segmentation faults are usually bugs in the program where it is trying
> to access memory it shouldn't be.  Very occasionally they can be
> caused by hardware issues corrupting pointers, but usually it's just a
> bug.
>
> If, as Justin suggests, it's a problem in your simulation then I would
> suggest that the program should detect it and tell you, not got off
> into the weeds and trigger a condition that causes the OS kernel to
> kill it with a SEGV.
>
> Might be worth reporting a bug and seeing what the maintainers say.
>

The presence of the printed warning indicates it's not a bug:

"t = 239.934 ps: Water molecule starting at atom 75561 can not be settled.
Check for bad contacts and/or reduce the timestep.
Wrote pdb files with previous and current coordinates"

Whenever this happens, the integration is failing.  Perhaps the seg fault isn't 
elegant, but I don't know if it's unavoidable or not.  When the constraints 
fail, typically molecules are careening across the system with (approaching) 
infinite velocity and/or coordinates.  There are various environment variables 
to allow more leeway when this happens and/or suppress coordinate dumping, but 
they're rarely used because this is not something one should usually try to 
blindly circumvent.

In 99.9999% of cases (rough guess ;), this is a failure due to instability 
caused by insufficient minimization/equilibration, bad topology, or unstable run 
settings and not a bug.

-Justin

-- 
========================================

Justin A. Lemkul, Ph.D.
Research Scientist
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

========================================