Dear Daniel and Shi,
Thank you for your responses.
Sorry I forgot to send starting command:
srun -n 11 ${AMBERHOME}/bin/pmemd.cuda.MPI -rem 3 -remlog remt1.log -ng
11 -groupfile inputs/ti.groupfile
Sorry, groupfile was also confusing. I used posted groupfile in first
calculation, in my attempt to restart calculation I changed -O to -A
and adjusted -c with *ti.rst7, but calculation resulted in error
"unknown flag: -A". I guess it is not implemented for H-REMD. I was just
used to do it in gromacs and found it quite convenient. I will write to
new files as you recommend.
HMR comment is interesting, it was not used in FE-Workflow. Shi, have
you tested HMR in TI? And by deleting noshakemask do you mean, I could
turn of shake completely after HMR?. I noticed that shake was turned off
in tutorial mdins (ntc=1), but in current FE-Workflow they turned it on
(ntc=2).
Juraj
On 11. 10. 23 1:17, Shi Zhang wrote:
> Hello Juraj,
>
> Do you mean that you want to append the restart simulation outputs to
> the ones from your previous run? My suggestion is to restart another
> whole new H-REMD simulation using the final *ti.rst7 from the previous
> run as the new starting structure for the next run. For your following
> analysis stage, if you want to analyze the free energy using the
> edgembar module in AmberTools23, you can combine all the extracted
> energy info from each run and perform the analysis. Besides that, the
> replica exchange log file (remlog) cannot be restarted at the moment,
> which still requires another set of output files.
>
> I notice that you are using dt=1fs, if you intend to run the H-REMD
> simulation longer, you may want to increase it to 4fs (with
> hydrogen-mass repartitioning) and some suggested changes to consider:
>
> 1. use parmed to turn on HMR for the unisc.parm7
> 2. change dt to 0.004
> 3. you can delete the "noshakemask" flags
>
> Hope this helps!
>
> Shi
> ------------------------------------------------------------------------
> *From:* Daniel Roe via AMBER <amber.ambermd.org>
> *Sent:* Tuesday, October 10, 2023 8:59 AM
> *To:* Juraj Dobias <juraj.dobias.uochb.cas.cz>; AMBER Mailing List
> <amber.ambermd.org>
> *Subject:* Re: [AMBER] restart H-REMD simulation
> Hi,
>
> On Tue, Oct 10, 2023 at 2:54 AM Juraj Dobias via AMBER
> <amber.ambermd.org> wrote:
> > According to manual T-REMD can be restarted using -A flag instead of -O,
> > but amber does not recognize -A flag. I tried changing starting rst7
> > file in groupfile, but all output files were overwritten. I need to
> > append new data to existing mdout file for subsequent analysis.
>
> In the groupfile you posted, you have -O flags set, not -A. I think
> you also might need to make sure you don't have -O specified on the
> command line itself.
>
> However, my advice is to just avoid -A altogether. Just restart from
> your old run and write to a new set of files. Personally, when I'm
> running REMD or any other ensemble type jobs, I put separate runs in
> their own directories, e.g.
>
> run.000
> run.001
> run.002
> etc
>
> This provides a nice logical separation and ensures no one file gets
> too big and unwieldy. Hope this helps,
>
> -Dan
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> https://nam02.safelinks.protection.outlook.com/?url=http%3A%2F%2Flists.ambermd.org%2Fmailman%2Flistinfo%2Famber&data=05%7C01%7Csz550%40chem.rutgers.edu%7C019f8bdbe0cb41a0b40308dbc990debe%7Cb92d2b234d35447093ff69aca6632ffe%7C1%7C0%7C638325396291738310%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=8VII7HQYNBf6FWqI5ZHnZquZcgwZgSL2Mq1kpZ1ge0U%3D&reserved=0
> <http://lists.ambermd.org/mailman/listinfo/amber>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Oct 11 2023 - 01:00:02 PDT
