Date: Thu Nov 17 2016 - 03:38:59 CST

Hi JC,

Is the bug being fixed in VMD for WIN32, version 1.9.3 beta 4 (2016-10-28) platform? Could you please help to elaborate more details about the bug?

Thank you.


From: Ern Ong
Sent: Tuesday, 15 November 2016 4:33 PM
To: 'JC Gumbart';
Subject: RE: vmd-l: ffTK: How to determine the optimized bond and angle parameters

I am using the VMD for WIN32, version 1.9.2 (December 29, 2014).


From: JC Gumbart []
Sent: Tuesday, 15 November 2016 4:20 PM
To: Ern Ong;<>
Subject: Re: vmd-l: ffTK: How to determine the optimized bond and angle parameters

Can you first say precisely which version of vmd you are using? We found a bug for repeated runs of the bond /angle optimization routine last winter.

Also, lower objective values are better.


On Sun, Nov 13, 2016, 4:34 AM Ern Ong <<>> wrote:

This is my first time using ffTK plugin to perform bond and angle optimization. I have been going through the ffTK tutorial and forums but I still can’t grasp the way to determine whether the obtained bond and angle parameters are optimized or not. For charge optimization, I adopted the SA-DH method with multiple rounds of SA and followed by DH. However in bond and angle optimization, I found that the values change a lot with each round of either SA or DH. The following is the questions.

1. Is it okay for me to run a long SA (probably T=25, Tsteps=100 and Iter=100 as one did and shared in this vmd forum) and assumed that the outcome values are optimized?

2. Actually how does the final objective value work out? Does it mean that the higher final objective value the more optimized it is?

Thanks for the help.