RE: AMBER: SANDER BOMB in subroutine nonbond_list

From: Ross Walker <ross.rosswalker.co.uk>
Date: Wed, 30 Jan 2008 08:20:51 -0800

Hi Pankaj,
 
You get all these warnings in your output file:
 
QMMM: WARNING!
QMMM: Unable to achieve self consistency to the tolerances specified
QMMM: No convergence in SCF after 1000 steps.
QMMM: Job will continue with unconverged SCF. Warning energies
QMMM: and forces for this step will not be accurate.
QMMM: E = -0.2250E+07 DeltaE = 0.2276E-04 DeltaP = 0.1519E-07
QMMM: Smallest DeltaE = 0.1215E-06 DeltaP = 0.7503E-07 Step = 702

vlimit exceeded for step 3; vmax = 22.6432
vlimit exceeded for step 4; vmax = 45.1148
vlimit exceeded for step 5; vmax = 55.3639
vlimit exceeded for step 6; vmax = 95.1641
vlimit exceeded for step 7; vmax = 319.9325
vlimit exceeded for step 8; vmax = 87.9556
vlimit exceeded for step 9; vmax = 287.1400
vlimit exceeded for step 10; vmax = 90.3142
vlimit exceeded for step 11; vmax = 1194.8515

And yet you only get concerned about the message the code gives when it
crashes?
 
These warnings should at least make you think that something might be wrong.
Essentially what is happening here is that the QM SCF is not converging,
most probably because your QM region does not make chemical sense. This
leads to high forces and your system is blowing up. There are numerous
problems with your input file and you will need to start thinking more
carefully about how you run these simulations.
 
Firstly you have qmcut=7 -----> Never ever reduce the cut off value for
anything in AMBER below 8.0, the PME assumes 8.0 is the minimum value for
the cut off, anything less than this will cause significant problems. If you
get errors about the cut off being too big, even at 8.0 then make your water
box bigger, do not reduce the cut off.
 
Then if we look at your QM region you have significant problems. E.g. just
look at the list of link atoms:
 
QMMM: nlink = 6 Link Coords Resp Charges
QMMM: MM(typ) - QM(typ) X Y Z MM QM
QMMM: 1825 CT 1827 CT -4.835 -2.508 4.581 0.007
-0.048
QMMM: 3036 CT 3038 CT -5.105 0.187 -0.601 0.007
-0.048
QMMM: 3048 CT 3050 CT -5.508 1.353 3.533 0.007
-0.048
QMMM: 7027 OS 7028 P 1.340 5.480 4.677 -0.523
1.166
QMMM: 7040 N* 7038 CT -0.240 5.231 -1.028 0.058
0.036
QMMM: 7092 n 7090 c3 2.363 1.637 -4.321 -0.067
0.378
QMMM:
------------------------------------------------------------------------

 
You are cutting an OS - P bond which has probably the biggest dipole of any
bond in your entire system. Do you really think that xxx-P-H is a good
approximation to xxx-P-O-xxx ? I would suggest taking a look through the
literature to see the sorts of places people cut the QM region and the sorts
of issues they see with link atoms etc. Essentially you should be aiming to
cut at bonds that have the smallest possible dipole moment and are purely
single bond in character - thus the CT-CT bonds are good candidates.
 
Try setting writepdb=1 so that sander writes you a pdb file of your QM
region that you can then open in something like VMD and visualize to make
sure your QM region makes sense. Also note that with phosphates in the QM
region it is possible that the charge of the QM region is not zero hence you
may be running the simulation with more electrons that should really be
there. I.e. it could have a charge of +2 - you will need to look at the
system you pick for the QM region and use some chemical intuition to try to
work out what the integer charge should be.
 
Good luck,
 
Ross
 

/\
\/
|\oss Walker

| Assistant Research Professor |
| San Diego Supercomputer Center |
| Tel: +1 858 822 0854 | EMail:- ross.rosswalker.co.uk |
| http://www.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
Pankaj R. Daga
Sent: Wednesday, January 30, 2008 07:14
To: amber.scripps.edu
Subject: AMBER: SANDER BOMB in subroutine nonbond_list



Dear Amber Community

 

Thanks for your suggestions about setting up the input file for the QM/MM
simulation. As suggested by Dr. Case, I have decreased the number of atoms
in QM region from 110 to 67. The input file works fine.

 

I have carried out minimization, heating and initial equilibration using MM
method. Now, I intend to carry out equilibration using QM/MM integration. I
am facing new error in the simulation. The job starts and gets aborted after
20 minutes with the error message,

 

"SANDER BOMB in subroutine nonbond_list

  volume of ucell too big, too many subcells

  list grid memory needs to be reallocated, restart sander"

 

I have searched in the archives for similar error, I could not succeed.
Could anyone please suggest me the reason for this error? I would appreciate
if you please can suggest me a way out. I am attaching my input and output
file along with.

 

Thanks and Regards

 

Pankaj

 

===============================================================
If your ship doesn't come in, swim out to it...Jonathan Winters
===============================================================
Pankaj R. Daga |
Dept. of Medicinal Chemistry | e-mail: pdaga_at_olemiss.edu
417 Faser Hall, | fax: +1-662-915-5638
University of Mississippi | phone: +1-662-915-1853
University, MS, 38677-1848 |
===============================================================

 


-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Sun Feb 03 2008 - 06:07:07 PST
Custom Search