AMBER Archive (2006)

Subject: AMBER: Did these differences come from so-called "different machines" ?

From: Zhihong Yu (computation_at_mail.nankai.edu.cn)
Date: Sun Apr 30 2006 - 03:47:28 CDT

Dear All,

    Now I'm learning Amber8 by performing Workshop Tutorial 1(http://www.rosswalker.co.uk/tutorials/amber_workshop/Tutorial_one/index.htm) step by step, but when I perform minimization before MD, I got different results compared with theirs. I examined the two out files carefully and find that: the results of steepest descent minimizations are completely same, but after 50 conjugate gradient minimizations the results begin to be different bit by bit, the final results are :

My result:

   NSTEP ENERGY RMS GMAX NAME NUMBER
   1000 -3.9091E+04 7.9419E-01 3.5390E+01 Na+ 641

 BOND = 2859.5296 ANGLE = 435.5573 DIHED = 452.7539
 VDWAALS = 6985.1288 EEL = -50065.8360 HBOND = 0.0000
 1-4 VDW = 248.0616 1-4 EEL = -294.5216 RESTRAINT = 288.7350

Their result:

   NSTEP ENERGY RMS GMAX NAME NUMBER
   1000 -3.8989E+04 5.9208E-01 2.7355E+01 P 446

 BOND = 2829.9826 ANGLE = 433.7389 DIHED = 452.8278
 VDWAALS = 6893.6489 EEL = -49839.8168 HBOND = 0.0000
 1-4 VDW = 248.3104 1-4 EEL = -294.6786 RESTRAINT = 286.5245
 EAMBER = -39275.9867

    Now my question is : Are these differences in conjugate gradient minimization in error range? Did these differences come from so-called "different machines" in manual Page 10?

    Any advice will be greatly appreciated!

           sincerely yours,Zhihong Yu

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