[gmx-developers] Re: NMR refinement

[santhosh]

Hello,

I have done some testing with -multi 2 option. Basically there are 2
simultaneous simulations which are running and I want them to sync with each
other with regards to timestep. (Both the simulations running at the same
time step). 

How do i achieve it without using -replex option? Currently the value is
'mysteriously' ten. 
I inserted a custom code, as described above in a link, which takes the
replicaId and the current timestep(in ps) and when ever the replicaId was 0,
I made it wait 0.5 secs. Here is the output of the run :

replicaID 0   timeStep 0.0 #waits 0.5 sec
replicaID 1   timeStep 0.0
replicaID 1   timeStep 0.001
replicaID 0   timeStep 0.001 #waits 0.5 sec
replicaID 1   timeStep 0.002
replicaID 1   timeStep 0.003
replicaID 1   timeStep 0.004
replicaID 1   timeStep 0.005
replicaID 1   timeStep 0.006
replicaID 1   timeStep 0.007
replicaID 1   timeStep 0.008
replicaID 1   timeStep 0.009
replicaID 1   timeStep 0.01
replicaID 0   timeStep 0.002
replicaID 0   timeStep 0.003
replicaID 0   timeStep 0.004
replicaID 0   timeStep 0.005
replicaID 0   timeStep 0.006
replicaID 0   timeStep 0.007
replicaID 0   timeStep 0.008
replicaID 0   timeStep 0.009
replicaID 0   timeStep 0.01
replicaID 1   timeStep 0.011
replicaID 0   timeStep 0.011
replicaID 1   timeStep 0.012
replicaID 1   timeStep 0.013
replicaID 1   timeStep 0.014
replicaID 1   timeStep 0.015
replicaID 1   timeStep 0.016
replicaID 1   timeStep 0.017
replicaID 1   timeStep 0.018
replicaID 1   timeStep 0.019
replicaID 1   timeStep 0.02
replicaID 0   timeStep 0.012
replicaID 0   timeStep 0.013
replicaID 0   timeStep 0.014
replicaID 0   timeStep 0.015
replicaID 0   timeStep 0.016
replicaID 0   timeStep 0.017
replicaID 0   timeStep 0.018
replicaID 0   timeStep 0.019
replicaID 0   timeStep 0.02

I am unable to find the piece of code which makes it wait until it makes it
sync. Any hint is what controls this parameter? 




--
View this message in context: http://gromacs.5086.n6.nabble.com/NMR-refinement-tp5004517p5004776.html
Sent from the GROMACS Developers Forum mailing list archive at Nabble.com.