[gmx-users] Error in generated Replica Exchange statistics in log file

[Abhi Acharya]

Hello GROMACS users,

I have a problem with the Replica Exchange statistics printed in log file.
I ran a 25 ns Replica exchange on a cluster through a queuing system.
However, due to insufficient wall time the job stopped abruptly around 24
ns and no stats were printed in the log file. I then extended the run by
10000 steps using -nsteps  and provided additional -cpi flag for
continuation. The simulation ran without errors but the replica exchange
stats that were printed in the log files of each of the replica is as
follows:

Replica exchange statistics
Repl  10 attempts, 5 odd, 5 even
Repl  average probabilities:
Repl     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   26   27   28
  29   30   31
Repl      .23  .45  .57  .12  .28  .20  .82  .46  .44  .50  .07  .74  .53
 .46  .44  .51  .22  .56  .24  .56  .23  .82  .30  .28  .59  .54  .34  .53
 .75  .29  .62
Repl  number of exchanges:
Repl     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   26   27   28
  29   30   31
Repl        1    2    3    0    0    2    4    2    2    2    0    4    3
 2    2    4    1    4    0    3    2    4    1    0    2    4    2    1
 4    1    2
Repl  average number of exchanges:
Repl     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   26   27   28
  29   30   31
Repl      .20  .40  .60  .00  .00  .40  .80  .40  .40  .40  .00  .80  .60
 .40  .40  .80  .20  .80  .00  .60  .40  .80  .20  .00  .40  .80  .40  .20
 .80  .20  .40



I rate of exchange attempt used was 1000, which means that the above stats
somehow only take into account the last 10000 steps of the simulations. Is
this some bug or I did something wrong ? More importantly, is there a way
to regenerate the replica exchange stats from the log file ?


Regards,
Abhishek Acharya
Shasara Research Foundation