VMD-L Mailing List

From: Roberto Gaspari (roby.gaspari_at_gmail.com)
Date: Thu Nov 08 2012 - 05:11:19 CST

Dear all,

I found what makes the trr not readable in my case.
I did an equilibration run on my local machine using some
setting stored in a eq.tpr file.

For the production run I copied the final checkpoint file on the remote
machine together with a new run.tpr file which I produced with some new
settings
(e.g. a different time interval for output writing).

By restarting the simulation on the remote computer I got a warning
message in the log file

"Gromacs binary or parallel settings not identical to previous run.
Continuation is exact, but is not guaranteed to be binary identical."

Now I tried again and load the eq.tpr file instead of the run.tpr in the
remote
machine. By restarting the simulation I did not get the warning anymore
and I could read the trr file.

Hope this helps also someone else

Roberto

2012/11/7 Bogdan Costescu <bcostescu_at_gmail.com>

> On Wed, Nov 7, 2012 at 9:05 AM, Roberto Gaspari <roby.gaspari_at_gmail.com>
> wrote:
> > This does not always happen. I tried for example to open another trr file
> > produced by gromacs on another machine and it went all right.
>
> You are changing here both the trr file and the machine, so the
> conclusion doesn't have any relation to the failed try.
>
> > If I try to load on another machine.
>
> It's not clear to me... have you tried opening the same trr with same
> version of VMD on another machine ? If so, what was the result ? Did
> this other machine have more memory ?
>
> > So it seems there's a problem in handling those libraries.
>
> It could also be a problem of limited memory, if the trr file contains
> many atoms and/or many frames. Another source of memory corruption
> could be a malformed trr file; 'gmxcheck' which is part of GROMACS can
> tell you whether the file is OK.
>
> Cheers,
> Bogdan
>