AMBER Archive (2004)

Subject: AMBER: question about delphi and UHBD

From: xhu1_at_memphis.edu
Date: Tue Sep 28 2004 - 11:19:32 CDT

Dear amber users,

I try to use delphi and UHBD to do MM-PBSA. For the same system, I get
the same coulombic energy, but reaction field energies are totally
different. Input parameters(grid size/dimension, focusing,
internal/solvent dielectric, temperature) are same except convergence
threshold. In UHBD, this value is 1 X 10E-6 (it is default value. I
don't know how to change it), and convergence is achieved at 200 steps
or so:
####################################################################
 ITERATION VALUES
   Maximum iterations (maxi #) : 400
   Convergence criteria (conv #.#) :0.1000E-05
   Linear equation solver flag (solv #) : 1
     - incomplete Cholesky preconditioned conjugate gradient
 PERFORMING SETUP AND CALCULATION
 ITERATION SUMMARY
   Iterations required : 189
   Norm of the constant vector :0.2526E+04
   Norm of the residual vector :0.2445E-02
   Convergence achieved :0.9680E-06
 ITERATION SUMMARY
   Iterations required : 189
   Norm of the constant vector :0.2526E+04
   Norm of the residual vector :0.2445E-02
   Convergence achieved :0.9680E-06
######################################################################

In delphi, I choose statement "RMSC" with the same value and
convergence can not be achieved at the required value in 1200 steps:

######################################################################
 rms-change max change #iterations
   1.0950725E+09 980.2277 at 10 iterations
   3.9263824E+08 273.5977 at 20 iterations
   1.4039014E+08 148.4565 at 30 iterations
   4.9958216E+07 86.13183 at 40 iterations
   .............. ........ .. .. ..........

   6.6439105E-05 2.4414063E-04 at 1120 iterations
   6.6107619E-05 2.4414063E-04 at 1130 iterations
   6.5878099E-05 2.4414063E-04 at 1140 iterations
   6.5984161E-05 2.4414063E-04 at 1150 iterations
   6.5964879E-05 2.4414063E-04 at 1160 iterations
   6.5717104E-05 2.4414063E-04 at 1170 iterations
   6.6142711E-05 2.4414063E-04 at 1180 iterations
   6.6114844E-05 2.4414063E-04 at 1190 iterations
   6.5725071E-05 2.4414063E-04 at 1200 iterations
finished qdiffx linear iterations

######################################################################

The final solvation energies for ligand, receptor and complex are
-1466, -1529 and -2730 kcal/mol in UHBD and -1035, -1056, -1224
kcal/mol in delphi.

Does anybody have any suggestion to solve this problem? Thank you so
much!

All the best

Shawn HU
Chemistry Department
University of Memphis

-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu