Thanks for your quick response. I set ntwx = 100 and I was able to get an
output file after running the simulation for a couple of minutes without
encountering the problem of disk space, or a lack thereof. Increasing the
ntwx really made a difference, especially for longer runs. Thanks!!
On 7/25/07, Ross Walker <ross.rosswalker.co.uk> wrote:
>
> Hi Lili,
>
> You are writing to the trajectory file on every MD step ntwx=1 - really
> you only want to do this for debugging over short <1000 steps runs. When you
> switch to running nanosecond long MD sims you should set NTWX to something
> like 250 or even 500 to 1000 depending on how often you want to save the
> coordinates. Note trajectory files can get very very large. You don't say
> how big your system is but consider a system of 50,000 atoms. Writing a
> formatted trajectory file which is uncompressed needs 8 bytes per coordinate
> so 24 bytes per atom per frame. So for 50,000 atoms this is approx 1.1 Meg
> per frame. If you run a 100,000 step simulation writing on every step to the
> mdcrd file then your mdcrd file will end up 112 GB!
>
> Even if you dump only every 1000 steps you will still get a 100MB
> trajectory file so make sure you have sufficient disk space available to run
> on. If you are running on a cluster you likely have a quota on your home
> directory so will need to write to a scratch directory that has more space
> available. Check with the admin of the machine to find out where very large
> files should be written during a run.
>
> All the best
> Ross
>
> /\
> \/
> |\oss Walker
>
> | HPC Consultant and Staff Scientist |
> | San Diego Supercomputer Center |
> | Tel: +1 858 822 0854 | EMail:- ross.rosswalker.co.uk |
> | http://www.rosswalker.co.uk | PGP Key available on request |
>
> Note: Electronic Mail is not secure, has no guarantee of delivery, may not
> be read every day, and should not be used for urgent or sensitive issues.
>
>
> ------------------------------
> *From:* owner-amber.scripps.edu [mailto:owner-amber.scripps.edu] *On
> Behalf Of *Lili Peng
> *Sent:* Wednesday, July 25, 2007 10:38
> *To:* amber.scripps.edu
> *Subject:* Re: AMBER: MD simulation error
>
> Thank you Drs. Case and Walker for your insightful comments. The only
> reason why I have a large timestep is that I initially tried it with dt=
> 0.001, but I get the error:
>
> "forrt1: Disk quota exceeded.
> forrt1: severe (38) error during write, unit 12, file
> /users/lpeng/md_classical10000.mdcrd"
>
> I wonder if this has anything to do with Amber (my guess is a no). I'm
> accessing AMBER via SDSC's Teragrid machine, and I presume that it's because
> of my disk space on the SDSC server. Any thoughts?
>
> Thanks again,
> Lili
>
> On 7/24/07, Ross Walker <ross.rosswalker.co.uk> wrote:
> >
> > Hi Lili,
> >
> > Most probably it is to do with time step - you are suing a 5fs time step
> > and while you fix all bonds (ntf=3, ntc=3) this is still likely too large -
> > take a look at your trajectory file and see what it looks like. Note I don't
> > think ntf>2 is used much so you may want to check that it actually works
> > properly for your system - is there a reason why you want to keep all the
> > bonds fixed? You should note that most things have not been parameterized to
> > run in this mode so I would be very cautious of your results - make sure you
> > can prove that your choice to keep the bonds fixed is justified - I.e.
> > perhaps try this out on some model systems and show that you can recover the
> > same results for the entity you are interested in whether you fix all bonds
> > or just bonds to hydrogen.
> >
> > You also have an extremely large gamma_ln value.
> >
> > All the best
> > Ross
> >
> > /\
> > \/
> > |\oss Walker
> >
> > | HPC Consultant and Staff Scientist |
> > | San Diego Supercomputer Center |
> > | Tel: +1 858 822 0854 | EMail:- ross.rosswalker.co.uk |
> > | http://www.rosswalker.co.uk | PGP Key available on request |
> >
> > Note: Electronic Mail is not secure, has no guarantee of delivery, may
> > not be read every day, and should not be used for urgent or sensitive
> > issues.
> >
> >
> > ------------------------------
> > *From:* owner-amber.scripps.edu [mailto:owner-amber.scripps.edu] *On
> > Behalf Of *Lili Peng
> > *Sent:* Tuesday, July 24, 2007 16:54
> > *To:* amber.scripps.edu
> > *Subject:* AMBER: MD simulation error
> >
> > Dear all,
> >
> > I am getting the following error for my MD simulation:
> >
> > Coordinate resetting (SHAKE) cannot be accomplished, deviation
> > is too large
> >
> >
> > NITER, NIT, LL, I and J are : 0 1 7 16 28
> > Note: This is usually a symptom of some deeper
> > problem with the energetics of the system.
> >
> >
> > My input file is the following:
> >
> > 310K constant temp MD
> > &cntrl
> > imin=0, ntb=1,
> > cut=8.0, ntc=3, ntf=3,
> > tempi=310.0, temp0=310.0,
> > ntt=3, gamma_ln=50,
> > nstlim=10000, dt= 0.005,
> > ntpr=1, ntwx=1,
> > /
> >
> > Does anyone have any idea why this occurs? Does it have anything to do
> > with the timestep?
> >
> > Thank you for your time,
> > Lili
> >
> >
>
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Sun Jul 29 2007 - 06:07:14 PDT
