[gmx-users] Re: range_check ci fatal error

[Giorgos Karvounis]

> Dear all,
>
> after 107480 steps of a simulation without T/P coupling (equilibrated
> with weak coupling for 20000 steps of 2 fs before the full run) i got
> the following range_check error:
>
> Fatal error: ci = -214748364
> 8 should be in 0 .. 139 [FILE nsgrid.c, LINE 210]
>
> Do i have to change the range of my box or is it something more
> profound?

have you checked energies and density, temperature and pressure?

are you running on an overheated Athlon?

Dear users,

after checking the Energy and flog files, it seems that the temperature
increases something like 10 k on every step and at the moment that i get
my first LINCS warning the temp is around 1000 k ! Do i have to use PME
? I am not familiar (yet)
and also i am not using ions , i just want to switch off the T/P
coupling in order to be able to reproduce the same run.
By the way, do you consider implementing this in any of the following
releases?

Thanks

Giorgos