Hi Qinghua!
As you expected, this solution works! Thanks so much!
If the atomic number of the dummy atom is assigned as "-1" in Mg.lib, the atomic number appearing
in the generated topology file would become "2". Actually, any integer other than "1" is OK.
One note is that if the atomic number is set to be "0", the NUMEXTRA (number of extra points found in topology)
at the top of the topology needs to be changed as the number of the atoms of type "EPW".
Best,
Zhicheng
________________________________
From: Qinghua Liao <scorpio.liao.gmail.com>
Sent: Thursday, June 21, 2018 3:42:22 PM
To: amber.ambermd.org
Subject: Re: [AMBER] AMBER18 pmemd.cuda running error for divalent metal ions represented using multisite (dummy) atom model with the 12-6-4 LJ potential
Hello Zhicheng,
Thanks for your interest in our work.
In the topology file, the atomic number was set as 1 for the dummy atom, so
the program treat all the dummy atoms as hydrogen atoms, though the vdW
parameters are set ZERO.
You can set the atomic number of the dummy atom as *-1* in the library
file (*.lib). I think it can fix the problem.
Please let me know if it works or not.
All the best,
Qinghua
On 06/21/2018 11:45 PM, Zuo, Zhicheng wrote:
> Hi,
> The recently released AMBER18 adds GPU support for 12-6-4 ion potentials.
> I did some tests, and found that it did works for the case where an ion is treated
> traditionally as a single point charge with the 12-6-4 LJ potential. However, for the simulation
> of the recent multisite ion (dummy atom) model with the 12-6-4 potential (https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fpubs.acs.org%2Fdoi%2F10.1021%2Facs.jpclett.7b02358&data=01%7C01%7CZhicheng.Zuo%40unthsc.edu%7C6c0c10569258411f091808d5d7c84f29%7C70de199207c6480fa318a1afcba03983%7C0&sdata=Feb2BPyJ1B55nxFamU7RzR%2Bg%2Fb0vHKXtGYaXvX9l7BY%3D&reserved=0),
> I got the following error from pmemd.cuda:
> "Too many hydrogens for a hydrogen network, exiting."
>
> Interestingly, there is no any problem with the cpu version of pmemd for the 12-6-4 multisite ion model above.
>
> It seems that this issue is related to implementation of pmemd.cuda toward such a special divalent ion model.
> Can someone help to figure it out? Thanks!
>
> Best,
> Zhicheng Zuo
> Postdoc research associate from College of Pharmacy,
> UNT Health Science Center, Fort Worth, TX, USA.
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Flists.ambermd.org%2Fmailman%2Flistinfo%2Famber&data=01%7C01%7CZhicheng.Zuo%40unthsc.edu%7C6c0c10569258411f091808d5d7c84f29%7C70de199207c6480fa318a1afcba03983%7C0&sdata=Gyb8uG1%2BNGBIktXpYfT%2FqP1CplSzLVVWIZq7DpCuY3Y%3D&reserved=0
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Flists.ambermd.org%2Fmailman%2Flistinfo%2Famber&data=01%7C01%7CZhicheng.Zuo%40unthsc.edu%7C6c0c10569258411f091808d5d7c84f29%7C70de199207c6480fa318a1afcba03983%7C0&sdata=Gyb8uG1%2BNGBIktXpYfT%2FqP1CplSzLVVWIZq7DpCuY3Y%3D&reserved=0
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Sat Jun 23 2018 - 08:00:02 PDT
