AMBER Archive (2003)

Subject: Re: MD steps limit in sander (Amber 6)

From: Marco Preto (mcpreto_at_fc.up.pt)
Date: Wed Oct 30 2002 - 21:41:01 CST

Thanks you for your quick response!

No, it was not the first time it was writting something. The Dynamic was
setup for a 1ns run but it only run for allmost 500 ps. We are currently
using Amber 6.0, on a monoprocessor P-IV desktop PC, without any queuing
system - I'm basicaly the only person using it!

About the pressure questions, can anyone show me some light at the end
of the tunnel?

Thanks for all

Marco A. C. Preto (Ph.D student)
Theoretical Chemistry group
Chemistry Department - FCUP
Porto, Portugal

Carlos Simmerling wrote:

> compare the step # that they died at to all of your
> writing frequencies, such as NTPR, etc. see if it
> on the first write of anything.
>
> Also, let us know what AMBER version you have and
> more info about when it died. Often no output goes
> to the AMBER output file but instead it goes to
> some std output- did you run these under a queuing
> system, etc.
>
> Was this an MPI job? Often error messages from non-master
> process can also go to the std output (such as SHAKE
> failures). This may vary based on AMBER version so make
> sure to tell us that too.
>
> If you can, rerun the job explicity sending the std output to
> a file so you can look at that as well as the AMBER output.
> run.job >& runout &
>
> Carlos
>
> ----- Original Message -----
> From: "Marco Preto" <mcpreto_at_fc.up.pt>
> To: "Amber Mailing list" <amber_at_heimdal.compchem.ucsf.edu>
> Sent: Wednesday, October 30, 2002 10:09 PM
> Subject: MD steps limit in sander (Amber 6)
>
>
>> Dear Amber users:
>>
>> I have been runing some MD runs of small peptides (8 aa) in a water box,
>> using periodic boundery condictions. I've tryed to run for a 1ns period
>> with a time step of 1fs (using shake for bonds involving H atoms, NPT
>> ensamble simulation with privious mdrun in NVT ensamble). I was trying
>> these simulations at diferent cutoffs and both of mdruns "died" at
>> exactly the same step! There was none error message in the Sander output
>> file! My question is, is there any limit to the number of steps on a MD
>> in sander ?
>>
>> Another thing: while looking at the mden file (energy data over the
>> trajectory) from these runs, we've noticed there was three diferent
>> types of pressure: pressure; Pres_scal_solu and Pres_scal_solv (I'm
>> using Berendsen barostact). My guess is that pressure is refering to the
>> instant pressure, but I haven't still found any reference to confirm
>> this! I used pres0=1.0 and Pres_scal_solv in around that value during
>> the simulation run! Is this the value of pressure under Berendsen
>> Pressure control? What do Pres_scal_solv and Pres_scal_solu refer to?
>>
>> Sorry for these naive questions and thanks you for all your attention
>> and replies!
>>
>> Sincerely yours
>>
>> Marco A. C. Preto (Ph.D student)
>> Theoretical Chemistry group
>> Chemistry Department - FCUP
>> Porto, Portugal
>>
>>