[gmx-users] REMD ... again a problem with analysis (on the fly)

[Berk Hess]

>From: <mollica.luca at hsr.it>
>Reply-To: Discussion list for GROMACS users <gmx-users at gromacs.org>
>To: gmx-users at gromacs.org
>Subject: [gmx-users] REMD ... again a problem with analysis (on the fly)
>Date: Sun, 23 Oct 2005 19:45:23 +0200
>
>Hi,
>
>unfortunately, waiting for the end of REMD simulation, I can
>only  ‘grep’ my log files for each replica run in order to check
>what’s going on … say, md0.log, md1.log …. up to md25.log .
>
>What I can see in each of my log file is (every 2 ps or 1000 steps,
>having set this value as exchange rate) something like this (this is
>extracted form replica 3):
>
>
>Replica exchange at step 48000 time 96
>Repl 3 <-> 4  dE =  4.036e+01  dpV =  1.591e-03  d =  4.037e+01
>Repl ex  0    1    2    3    4    5    6    7    8    9   10   11
>12   13   14   15   16   17   18   19   20   21   22   23   24   25
>Repl
>pr        .00       .00       .00       .00       .00       .00       .
>00       .00       .00       .00       .00       .00
>
>
>Replica exchange at step 49000 time 98
>Repl 2 <-> 3  dE =  5.380e+01  dpV =  9.683e-04  d =  5.380e+01
>Repl ex  0    1    2    3    4    5    6    7    8    9   10   11
>12   13   14   15   16   17   18   19   20   21   22   23   24   25
>Repl
>pr   .00       .00       .00       .00       .00       .00       .00
>     .00       .00       .00       .00       .00       .00
>
>
>Actually, I do not get the point of what is going to happen: does the
>statement
>
>Repl 3 <-> 4  dE =  4.036e+01  dpV =  1.591e-03  d =  4.037e+01
>
>mean that replica 3 and 4 are going to exchange ? This would mean in
>my case that a replica exchange event is happening every 2ps, but this
>seems to happen for every replica and it seems so strange also due to
>the fact that the probability for each exchange is set to .00 for
>every replica couple for each attempt ….
>Do I have to wait for the end of simulation and get the final
>statistics of the job ? Or am I missing completely the point ? If yes,
>are going to be provided any detailed explanations of the REMD
>features of GMX 3.3 in the near future ?

No, <-> means an exchange attempt.
But as the energy difference is 40 kJ/mol, the probability is extremely 
small
and it will not exchange.
At the beginning of the log file it says that x means exchange.
I assume you have never observed an x then, whith probably means
you temperature differences are too large.
I wrote a section on REMD in the manual, but it is not on the site yet
(it does not explain anything on the output format though).

Berk.