[RMCProfile-users] [EXTERNAL] SWAP_MULTI

[Isaac Abrahams]

Dear Aron

thanks for the advice.

Isaac


On 15/09/2022 23:06, SUMMER, ARON (PGR) via rmcprofile-users wrote:
> Hi Issac,
>
> I've come across this previously, and in fact it's still a warning which prints on my latest RMC runs.
>
> In my experience it will always print that warning to the cmd window as long as "SWAP ::" is not included in the .dat file. I've been able to safely ignore it, I suspect it's a message which wasn't updated when 'SWAP_MULTI ::' became a function.
>
> Here is an example of a working 'SWAP_MULTI ::' section of my .dat files using RMCprofile v6.7.8. I know I'm running a version behind but I can't imagine this feature has changed.
> -----------------------------------------
> SWAP_MULTI ::
>    > SWAP_ATOMS :: 1 2 0.15
>    > SWAP_ATOMS :: 3 4 0.15
> ------------------------------------------
>
> Your setup looks like about right but I am worried about the two 0.5 weightings, in my understanding that leaves no space for translations to occur leaving all your atom sites fixed, as the total probability of a swap being chosen is 1.0 with atom swaps of 1 - 2 and 3 - 4 both having a 50 % chance to occur, unless you're doing something which requires no translations, I'd recommend lowering those values.
>
> I find keeping atom swaps at below a total of 0.4 allows the model to converge in a reasonable timeframe, and atom swaps of below 0.1 get rejected a lot resulting in very few functional swaps.
>
> My advice would be to run with lower swap probabilities and see if any swaps get reported in the 'Refinement summary' part of the cmd output, below is an example from a 1-minute run of mine where the warning of no swaps was printed. See that 221 swaps were actually accepted, but at only half the desired probability.
> -------------------------------------------------------------------------
>   Refinement summary
>
>   Move type  n(acc)     mean DX^2     p(acc)     p(gen)
>
>   Trans.       1318       90102.0743        0.8564        0.7000
>   Spin            0           0.0000             0.0000        0.0000
>   Swap          221       58781.0370        0.1436        0.3000
> -------------------------------------------------------------------------
>
> Hope this helps,
> Aron
>
> ________________________________
> From: rmcprofile-users <rmcprofile-users-bounces at shadow.nd.rl.ac.uk> on behalf of Zhang, Yuanpeng via rmcprofile-users <rmcprofile-users at shadow.nd.rl.ac.uk>
> Sent: 15 September 2022 14:36
> To: rmcprofile-users at shadow.nd.rl.ac.uk <rmcprofile-users at shadow.nd.rl.ac.uk>
> Cc: Zhang, Yuanpeng <zhangy3 at ornl.gov>
> Subject: Re: [RMCProfile-users] [EXTERNAL] SWAP_MULTI
>
> Hi Isaac,
>
> It seems that the program is still reading the 'SWAP ::' section in the .dat file for some reason.
>
> The 'SWAP_MULTI ::' for sure will work in 6.7.9 if you set it up correctly. Could you please help sharing the .dat file with me (zhangy3 at ornl.gov) so I can check?
>
> Cheers,
> Yuanpeng
>
> --
>
>
> Dr. Yuanpeng Zhang
> Neutron Scattering Scientist
> Spallation Neutron Source
>
> Oak Ridge National Laboratory
> 1 bethel valley road, Oak Ridge, TN 37830, US
> Mobile: +1(865)-335-4818
> Email: zhangy3 at ornl.gov | Web: https://www.iris2020.net <https://www.iris2020.net/>
>
> On 9/15/22, 08:20, "rmcprofile-users on behalf of Isaac Abrahams via rmcprofile-users" <rmcprofile-users-bounces at shadow.nd.rl.ac.uk on behalf of rmcprofile-users at shadow.nd.rl.ac.uk> wrote:
>
>      Dear all
>
>      has anyone had managed to get SWAP_MULTI :: to work in version 6.7.9?
>      SWAP :: works fine but if I try to the following
>
>      SWAP_MULTI ::
>       > SWAP_ATOMS :: 1 2 0.5
>       > SWAP_ATOMS :: 3 4 0.5
>
>      I get
>
>        No atom swap moves will be generated,
>        atoms will only be translated.
>
>      Any thoughts?
>
>      Isaac
>
>      _______________________________________________
>      rmcprofile-users mailing list
>      rmcprofile-users at rmcprofile.org
>      hxxp://lists.rmcprofile.org/mailman/listinfo/rmcprofile-users
>
>
> _______________________________________________
> rmcprofile-users mailing list
> rmcprofile-users at rmcprofile.org
> http://lists.rmcprofile.org/mailman/listinfo/rmcprofile-users
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <https://lists.rmcprofile.org/pipermail/rmcprofile-users/attachments/20220915/b9a813c4/attachment.htm>
> _______________________________________________
> rmcprofile-users mailing list
> rmcprofile-users at rmcprofile.org
> http://lists.rmcprofile.org/mailman/listinfo/rmcprofile-users