Okay, for you to have a restart that went into this thing without box
information, you must have done a run with ntb = 0; at least for pmemd that
is the only way this would happen, and pmemd tries to mimic sander on such
things (translation - I looked at pmemd to guess what sander would probably
do too). So there are almost certainly some mixups between
periodic/nonperiodic runs involved here - not in the prmtop itself perhaps,
but there are ways you could have done a nonperiodic run in a stream with a
prmtop intended for use under periodic boundary conditions, I believe. You
need to review all your mdin's leading up to the simulation that failed and
see if you set ntb = 0 in any of them. Another thing to check - in the
mdout's, in periodic boundary condition simulations with pme, you will
always get box dimensions listed as "box x =", "box y =", "box z =" in the
ewald parameters section. These should obviously never be 0.0. It does
take a lot of attention to detail to set up a new simulation.
Regards - Bob Duke
----- Original Message -----
From: "Arturas Ziemys" <arturas.ziemys.uth.tmc.edu>
To: <amber.scripps.edu>
Sent: Tuesday, July 01, 2008 1:17 PM
Subject: Re: AMBER: "radius = 0.000"
> Hi,
> So, here is the answers:
>
> 2 Bob Duke:
> A. I have done input in "true" float for cut-off - the same problem.
>
> Q. One simple thing to check - is your compiler indeed converting the
> integer input value for "cut" to float correctly (should be apparent from
> the listing of used values in the mdout file). I actually don't know what
> the spec says should happen a namelist on this, but the intel compiler
> will do the correct conversion for you; most folks I believe would bother
> to specify 12.0 in the mdin for cut. This may not be it at all; other
> worries would be if the input prmtop/inpcrd from leap are actually what
> you think they are (for a solvated system).
>
>
> 2 Thomas Steinbrecher:
> A. My prmtop contains "1" - so systems setup done as periodic:
> 0 0 0 0 0 0 0 1 26
> 0
> However: minimization restart (output) file contains no dimension info.
> The DNA turorial examples does the same, but runs w/ no problem.
>
>
> 2 Carlos Simmerling
> A. I have run just DNA tutor case - it was OK. AMBER after compilation
> passed all tests (parallel and serial, g95 compiler, AMBER 9, Xeon,
> Redhat, gcc 3.x).
> Q. have you run all of the test cases?
> Q. did they work without error?
>
> A. My system is simple protein with standard amino acids, no
> modifications, no extra topologies. System contains that protein, water
> (TIP3P) and one Sodium ion. No restrains, no fixed particles = just
> "plain" MD.
> Q. is there anything unusual about your system- just standard protein
> and water?
>
> A. Here is my essential build steps:
> parm99 = loadamberparams parm99.dat
> pdb = loadpdb "1T29.pdb"
> addions pdb Na+ 0
> solvatebox pdb TIP3PBOX 10.0
> saveamberparm pdb BRCT.prmtop BRCT.inpcrd
> During the build there was no errors, just warnings including a
> long 3.2A distance between atoms on backbone, reconstructed missing atom
> in residues.
> Q. you might give us some info from leap on how you built it, and
>
> A. The restart file taken from minimization, where the input of job
> setup was taken from DNA tutorial. That restart file contains no box
> dimensions and velocities ( as DNA tutorial does - the same).
> Q. also what kind of simulation gave you the restart file.
>
>
> Thank you for your input and patience. It seems that my migration to AMBER
> needs some time ...
>
> Arturas
>
>
> Some update:
> After minimization it seems there are no velocities and box size (the last
> line) in a restart file. However, say DNA test in AMBER tutor worked fine
> and those lines are present.
>
> Carlos Simmerling wrote:
>> a few things to check first:
>> have you run all of the test cases?
>> did they work without error?
>> is there anything unusual about your system- just standard protein and
>> water?
>> you might give us some info from leap on how you built it, and
>> also what kind of simulation gave you the restart file.
>>
>> On Tue, Jul 1, 2008 at 12:08 PM, Arturas Ziemys
>> <arturas.ziemys.uth.tmc.edu> wrote:
>>
>>> Hi,
>>>
>>> I have minimized my pretty simple protein+water system. Supplied
>>> minimized
>>> structure (from restart output) to equilibration. My job was hanging
>>> forever. Using '-o stdout' I saw that my job stalled at after that line
>>> is
>>> printed: "sphere to fit in unit cell has radius = 0.000".
>>>
>>> I have tried to find some answer on internet but without success - just
>>> one
>>> similar question. My feeling says that job does not know my systems
>>> dimensions ? What is the remedy ?
>>>
>>> job setup:
>>> &cntrl
>>> imin = 0,
>>> ntx = 1,
>>> irest = 0,
>>> ntt = 3,
>>> gamma_ln = 1.0,
>>> temp0 = 300.0,
>>> tempi = 300.0,
>>> ntp = 1,
>>> ntb = 2,
>>> ntf = 2,
>>> ntc = 2,
>>> cut = 12,
>>> dt = 0.001,
>>> nstlim = 100
>>> /
>>>
>>>
>>> --
>>>
>>> Arturas Ziemys, PhD
>>> School of Health Information Sciences
>>> University of Texas Health Science Center at Houston
>>> 7000 Fannin, Suit 880
>>> Houston, TX 77030
>>> Phone: (713) 500-3975
>>> Fax: (713) 500-3929
>>> -----------------------------------------------------------------------
>>> The AMBER Mail Reflector
>>> To post, send mail to amber.scripps.edu
>>> To unsubscribe, send "unsubscribe amber" (in the *body* of the email)
>>> to majordomo.scripps.edu
>>>
>>>
>> -----------------------------------------------------------------------
>> The AMBER Mail Reflector
>> To post, send mail to amber.scripps.edu
>> To unsubscribe, send "unsubscribe amber" (in the *body* of the email)
>> to majordomo.scripps.edu
>>
>>
>
>
> --
>
> Arturas Ziemys, PhD
> School of Health Information Sciences
> University of Texas Health Science Center at Houston
> 7000 Fannin, Suit 880
> Houston, TX 77030
> Phone: (713) 500-3975
> Fax: (713) 500-3929
> -----------------------------------------------------------------------
> The AMBER Mail Reflector
> To post, send mail to amber.scripps.edu
> To unsubscribe, send "unsubscribe amber" (in the *body* of the email)
> to majordomo.scripps.edu
>
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" (in the *body* of the email)
to majordomo.scripps.edu
Received on Wed Jul 02 2008 - 06:07:45 PDT
