[gmx-developers] Re: [gmx-users] issue in replica exchange

[XAvier Periole]

Ok, the redmine is filled up and anybody who has time to help finding the issue is welcome :) I can't do much more!

As an alternative a colleague suggested that I could potentially get around the problem by using a compilation combining Open-MP or thread-MPI for each replica running on one node (shared memory=no need of decomposition) and MPI across nodes. 
Would it work? 

But I might have a problem, the CG uses shifted potentials and they seem not to be implemented with OpenMP and thread-MPI is not compatible with MPI … 

Any other solution one could imagine?

XAvier.

On May 9, 2013, at 1:01 PM, XAvier Periole <x.periole at rug.nl> wrote:

> 
> I finally could reproduce the problem in gmx461 and have fled up a red mine report. 
> 
> I hope we can fix this easily but I am not sure how things go go from now! Someone will get the bug assigned and fix it when ever possible or something else?
> 
> Thank you all for the help,
> XAvier.
> 
> On May 2, 2013, at 10:15 PM, XAvier Periole <x.periole at rug.nl> wrote:
> 
>> 
>> I'll look at the 4.6.1 version next week, I could install it but I got a conflict between the environmental variable defining openMP variable but I turned it off during compilation … 
>> 
>> You could try to run on particle decomposition to see if you get a problem … it should one quite quick.
>> 
>> On May 2, 2013, at 2:36 PM, Michael Shirts <mrshirts at gmail.com> wrote:
>> 
>>> Both.  So if 4.6.1 doesn't work, I want to know so we can patch it
>>> before 4.6.2 comes out.  If it does work, then there is probably stuff
>>> that can be backported.
>>> 
>>> On Thu, May 2, 2013 at 8:32 AM, XAvier Periole <x.periole at rug.nl> wrote:
>>>> 
>>>> You mean working with or working on the code?
>>>> 
>>>> I'll try gmx-4.6.1
>>>> 
>>>> On May 2, 2013, at 2:26 PM, Michael Shirts <mrshirts at gmail.com> wrote:
>>>> 
>>>>> Quick check here -- is 4.6 behaving correctly?  I actually spent some
>>>>> time working on REMD in 4.6, and it seems to be behaving  correctly in
>>>>> my hands with temperature and pressure control.
>>>>> 
>>>>> Thanks for any additional info on this!
>>>>> 
>>>>> On Thu, May 2, 2013 at 8:18 AM, Mark Abraham <mark.j.abraham at gmail.com> wrote:
>>>>>> On Thu, May 2, 2013 at 12:58 PM, XAvier Periole <x.periole at rug.nl> wrote:
>>>>>> 
>>>>>>> 
>>>>>>> I saw that redmine report, which could be related but it seems to happen
>>>>>>> only for runs done outside the domain and particle decompositions.
>>>>>>> 
>>>>>>> I'll fill up a red mine.
>>>>>>> 
>>>>>>> Anything I could do to help speeding the fix?
>>>>>>> 
>>>>>> 
>>>>>> What'd be really nice is some thought on how one can demonstrate that the
>>>>>> implementation of the exchange matches what would be expected from the
>>>>>> theory. For T-exchange under NVT, it is sufficient to rescale velocities
>>>>>> and quantities derived from them by the correct factor. That includes
>>>>>> various things like T-coupling history and integrator half-step quantities
>>>>>> (and does REMD with leap-frog make sense anyway?). For NPT, there's
>>>>>> probably also some P-coupling quantities to scale, and the box to exchange.
>>>>>> Anything I've missed? Hopefully virial contributions don't matter either
>>>>>> way?
>>>>>> 
>>>>>> Perhaps a decent first step is to hack the code to do a "self exchange," by
>>>>>> clearing the entire state and rebuilding with what would/should be received
>>>>>> from an exchange with a hypothethetical replica in an identical
>>>>>> pre-exchange state. Only if the code can do that (i.e. mdrun -reprod
>>>>>> produces a trajectory indistinguishable from a run that does not attempt
>>>>>> this self exchange) is it worth considering proper state exchanges, and the
>>>>>> process of making the code do the former should illustrate what is required
>>>>>> for the latter.
>>>>>> 
>>>>>> Mark
>>>>>> --
>>>>>> gmx-users mailing list    gmx-users at gromacs.org
>>>>>> http://lists.gromacs.org/mailman/listinfo/gmx-users
>>>>>> * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
>>>>>> * Please don't post (un)subscribe requests to the list. Use the
>>>>>> www interface or send it to gmx-users-request at gromacs.org.
>>>>>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>>>>> --
>>>>> gmx-users mailing list    gmx-users at gromacs.org
>>>>> http://lists.gromacs.org/mailman/listinfo/gmx-users
>>>>> * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
>>>>> * Please don't post (un)subscribe requests to the list. Use the
>>>>> www interface or send it to gmx-users-request at gromacs.org.
>>>>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>>>> 
>>>> --
>>>> gmx-users mailing list    gmx-users at gromacs.org
>>>> http://lists.gromacs.org/mailman/listinfo/gmx-users
>>>> * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
>>>> * Please don't post (un)subscribe requests to the list. Use the
>>>> www interface or send it to gmx-users-request at gromacs.org.
>>>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>>> -- 
>>> gmx-users mailing list    gmx-users at gromacs.org
>>> http://lists.gromacs.org/mailman/listinfo/gmx-users
>>> * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
>>> * Please don't post (un)subscribe requests to the list. Use the 
>>> www interface or send it to gmx-users-request at gromacs.org.
>>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>> 
>> --
>> gmx-users mailing list    gmx-users at gromacs.org
>> http://lists.gromacs.org/mailman/listinfo/gmx-users
>> * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
>> * Please don't post (un)subscribe requests to the list. Use the
>> www interface or send it to gmx-users-request at gromacs.org.
>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
> 
> --
> gmx-users mailing list    gmx-users at gromacs.org
> http://lists.gromacs.org/mailman/listinfo/gmx-users
> * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
> * Please don't post (un)subscribe requests to the list. Use the
> www interface or send it to gmx-users-request at gromacs.org.
> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists