Hi David,
> Le 16 déc. 2019 à 12:40, David Case <david.case.rutgers.edu> a écrit :
>
> On Mon, Dec 16, 2019, Jerome Eberhardt wrote:
>>
> (Why are you setting up
>> periodic condition when I tell you to not do it?!)
>
> The original reason for setting up a "virtual box" when igb=0 and ntb=0
> was to re-use the nonbonded list code that was already present for
> periodic systems.
I am more familiar with the CHARMM environnement, and I know that if I want to do a minimization with periodic conditions, I have to do it manually. So I was surprised that I couldn’t do the same with sander.
>
> This "works" for many systems, but not for ones where atoms can move a
> lot from their initial positions. Setting igb=6 was the recommended
> workaround for this problem. But the curse of "backwards
> compatibility", aka "dusty deck syndrome" means that we (collectively)
> never turned off the original option.
>
>>
>> ### Output imin = 1 ###
>> NSTEP ENERGY RMS GMAX NAME NUMBER
>> 50 -3.9771E+03 5.8922E+00 1.0709E+02 H1 4340
>>
>> BOND = 546.2850 ANGLE = 923.4480 DIHED = 2521.1605
>> VDWAALS = 2749.6095 EEL = -21571.2602 HBOND = 0.0000
>> 1-4 VDW = 872.4954 1-4 EEL = 9640.5688 RESTRAINT = 340.6237
>> EAMBER = -4317.6932
>>
>> ### Output imin = 5 ###
>> NSTEP ENERGY RMS GMAX NAME NUMBER
>> 50 -5.6607E+03 5.5423E+00 7.8322E+01 O 3388
>>
>> BOND = 541.7428 ANGLE = 769.3931 DIHED = 2506.6400
>> VDWAALS = 2494.5050 EEL = -22821.0985 HBOND = 0.0000
>> 1-4 VDW = 825.6802 1-4 EEL = 9594.1609 RESTRAINT = 428.2861
>> EAMBER = -6088.9764
>
> Do the energies match a step 1?
At step 1 they both match.
>
> If you use igb=6, be sure to also use a large cutoff, ideally some value
> larger than the size of your system. There is really no other good way
> to handle long-range electrostatics for vacuum systems.
THIS does the trick! While it is not working on one frame (see previous email), it does work on the whole trajectory.
Thanks again!
Jerome.
>
> ....dac
>
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Mon Dec 16 2019 - 17:00:02 PST
