From: Raman Preet Singh (ramanpreetsingh_at_hotmail.com)
Date: Wed Aug 12 2020 - 12:40:45 CDT
Thank you very much for the detailed, step-by-step info. Really helpful.
From: Alex Balaeff<mailto:abalaeff_at_polarisqb.com>
Sent: 12 July 2020 21:20
To: namd-l_at_ks.uiuc.edu<mailto:namd-l_at_ks.uiuc.edu>; Raman Preet Singh<mailto:ramanpreetsingh_at_hotmail.com>
Subject: Re: namd-l: Restarting aborted NAMD run
Typically, you restart your simulation from the latest restart point saved before the crash. See the parameters restartname/restartfreq in the manual:
And you set the firsttimestep to that of that latest restart point. Then you will have to stitch together the DCD files.
For example, let's say that your original simulation:
- was aborted at step 12975000
- saved the latest DCD frame at the step 12974000 (dcdfreq=2000)
- and saved the latest restart point at the step 12950000 (restartfreq=50000)
Then you'll restart the simulation with firsttimestep=12950000 (remember to supply the correct file names - those of your saved restart point - to bincoordinates/binvelocities/extendedsystem). And you will end up with 2 dcds. One from steps 0-12974000, the other from steps 12950000-20000000 (presuming you run for 20 mln steps total). As you see, these DCDs have a 13-frame overlap: timesteps 12950000, 12952000, ..., 12974000. One removes those 13 frames from the first DCD and stitches the two DCDs together. You can use catdcd for that:
On Sun, Jul 12, 2020 at 4:57 AM Raman Preet Singh <ramanpreetsingh_at_hotmail.com<mailto:ramanpreetsingh_at_hotmail.com>> wrote:
I am a new NAMD user and have a very trivial question.
I ran a simulation which got aborted at step 12975000 (the last step in .xst file) due to power outage. Looking through the NAMD mailing list and NAMD tutorial, I believe that I need to restart my simulation with firsttimestep set to 12975000. Then I need to concatenate the two dcd files. Is this the correct way?
-- ----- Dr. Alexander Balaeff Polaris Quantum Biotech www.PolarisQB.com<http://www.PolarisQB.com> (919)-270-5772
This archive was generated by hypermail 2.1.6 : Thu Dec 31 2020 - 23:17:14 CST