-
Notifications
You must be signed in to change notification settings - Fork 7
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Fix restart #108
Comments
Hi @ElliottKasoar, sorry to piggy back on this issue, but I was just thinking about restart and heating, when designing a workflow using the heating. It seems that restart loads up the data file as you say but not explicitly the My assumption now was if I pass in @alinelena and I were just discussing also the possibility of having a "final" restart file that has a known filename to aid automation of workflows. Currently If I change the final temperature any follow on workflow will have to be changed to account for the new file name that depends on the step (I think). And it may not be the final step either if the user does not set the interval correctly. e.g. I might even suggest we could have an option to explicitly override the restart file name, so we always know it would be say I'd try to do it now, but currently catching up on how the attachments work Not sure what the best course of action is. |
Actually revisiting this, it's while it may not be entirely intuitive, I think it currently works as I originally intended for normal MD. In the original proof of concept, there's a separate auto restart option, which I didn't feel essential for the initial implementation. It's probably worth adding that option back in, but I'm not sure it's necessarily a problem that by default it relies on the user specifying the correct files.
Yes, I think that's right.
I'll have to think about this more/discuss it when I have more time, but I think it's useful to keep the step in the file names, as otherwise it's not necessarily obvious where you're picking up from. Assuming we keep multiple restart files, I'm not sure there's an obvious way of naming the others anyway? |
the idea is not to change the restart files but add another file that is the final frame, of the simulation or of the T for a ramp. |
As Alin said yes, keeping the file names so we know what the restart is for is definitely important, I think you are right on that Elliott. Alin and I iterated a little on this, culminating in this main...harveydevereux:janus-core:final_restart This would leave a work directory with these files, where
when called as so Whether that is the best way forward, I'm not sure. This might complicate auto restart though, a lot, I expect. |
i think we start to few issues in here.
|
So we now have the final file saved, and have updated the file names to account for heating and/or MD, but I think we're still missing:
Probably best to split the first two points into one issue, and the last into another, unless anyone disagrees with what needs doing? |
I do not think we want 1... adds a lot of extra complications for little gain.
|
Ok, created a specific issue for auto restart, otherwise sounds like we're ok. |
Restart correctly identifies the step from the data file, but I'm not sure it actually use the restart file to set the structure.
We can probably update the step-reading to use the data loader too.
Restart also needs some reworking for heating from #107
The text was updated successfully, but these errors were encountered: