[gmx-users] tpbconv

[Erik Lindahl]

Nguyen Hoang Phuong wrote:

>Dear All,
>
>I use the program tpbconv to remove the solvent from my tpr file and rerun
>the simulation without solvent. The simualtion is crashed with the error:
>
>Fatal error: calloc for theta[i] (nelem=-21307068, elsize=8, file pme.c,
>line 699): Cannot allocate memory
>
>Can anyone tell me what is wrong here? Thanks in advance.
>
Yes, as it says in the help text for tpbconv (tpbconv -h) you can not 
run a simulation with a tpr file
created by removing some of the atoms. It is only useful when you have 
written a subset of your system
e.g.  to an xtc file, and need a convenient tpr file to analyze this 
subset of the system.

Just comment out the solvent line in your topology, run trjconv to 
create a gro file with only the protein,
and then grompp to create a new input file. You should probably change 
the center-of-mass motion
removal to remove angular motion if you are simulating in vacuuo.

Cheers,

Erik