[gmx-users] multiple time step algorithm

[Ran Friedman]

David van der Spoel wrote:
> Ran Friedman wrote:
>>   Berk Hess wrote:
>>> It would also make the code more complicated than it already is.
>>>
>>> Berk.
>> It may be useful to incorporate these things as part of the source
>> but to compile them only if requested by the user, as done e.g. in
>> some features of charmm. This will make the main code only minimally
>> more complex I guess.
> This is basically how things work with some libraries, like FFT etc
> and also with the QMMM interface.
>
>>
>> There were recently several contributions that didn't make it to the
>> main code. This is very understandable, but I guess that as the GMX
>> code evolves, those contributions become outdated and force their
>> users to use older versions (at least for a while) or simply phase out.
>
> Just wondering, but what are you thinking of here? Stuff that goes to
> the contribution page on the website?
Could be, but I was thinking about things like GMX/CPMD,
GMX/metadynamics and interfaces with programs that solve the PB equation.

Ran.