[gmx-developers] FW: repeatability of runs?

[Shirts, Michael R. (mrs5pt)]

>Note that we don't guarantee that -reprod will give reproducable runs.
>But without a GPU (and excluding the possibility of non-deterministic
>MPI reductions), I don't see anything that could interfere with
>reproducibility. 

Yes, that's what I would expect.  Note that I don't think that we need
MANY ways that the trajectories should be reproducible -- but it would be
useful if there was one way that it was.

>Note that restarting from checkpoint files should
>result in reproducible results. This can help with debugging, since you
>don't need to run all those ns, only from the last checkpoint and you
>can then reduce the checkpoint interval until it crashes in a few steps.

I agree, they should, but they haven't been, at least with this example,
on my machine (I'll add some more details in the issue).  I've started off
with just the tpr, running on one thread, with the assumption that that's
is simpler than .cpt.

Best,
~~~~~~~~~~~~
Michael Shirts
Associate Professor
Department of Chemical Engineering
University of Virginia
michael.shirts at virginia.edu
(434) 243-1821