GROMACS development

09/23/2018

06:34 AM GROMACS Task #1868: implement mdrun -rerun better, simplifying do_md

> Maybe just don't report the virial unless explicitly requested (with an error if not possible). Preserving old beha...

08/30/2018

04:04 AM GROMACS Task #2616: Model for MD state

> Where do we already already need copies of state objects (such as t_state in minimize)? And are there exceptions ei...

12/30/2017

10:29 PM GROMACS Bug #2264: Changes in bonded interactions are not considered in expanded ensemble simulations

So, I believe there may be an error here (there are lots of different ways of doing the settings that have not all be...

10/23/2017

01:45 PM GROMACS Bug #2264: Changes in bonded interactions are not considered in expanded ensemble simulations

Berk Hess wrote:
> I voted -1 on the fix on gerrit because the group scheme non-bonded potentials are treated with i...

09/07/2017

06:35 PM GROMACS Bug #2244: Test rerun with FEP

Berk, could you post a sample file that cause the problem? I'd love to get a regression test in the free energy sect...

05/23/2017

09:24 PM GROMACS Feature #2186 (New): Potential change for logical improvements: move control of constraints purely to the .mdp

Here's a suggestion that could control flow control within the program logic, though it might not be worth doing beca...

03/11/2017

08:32 AM GROMACS Feature #2137 (New): Preliminary refactoring of constraints and update machinery.

As per Mark Abraham, preliminary refactoring of the constraints or update machinery, with a view to evolving a more c...

08:30 AM GROMACS Bug #2020: Possible issue with md-vv integrator

Did this get resolved?
Regardless, bumping so I remember to look at this when we are testing the physical consi...

02/25/2017

11:21 PM GROMACS Revision 198bc2a9: Fix for an incorrect check that nstlog !=0

The original version was accidentally reversed, causing it to
fail when nstlog =/= 0. Wasn't noticed because of lack ...

02/19/2017

05:04 AM GROMACS Feature #1137: Proposal for integrator framework (do_md) in future GROMACS

> Saving output should probably be done by handing off a buffer according to the user's schedule to a thread e.g. doi...