[gmx-users] Incompatible subsytems for REMD when running on multiple nodes

[Daniel Kozuch]

Hello,

I am running GROMACS 2019.4 (with GPUs) using the following command on two
nodes (each with 28 processors, and 4 GPUs).

srun -n 56 gmx mdrun -s sim -cpi sim -append no -deffnm sim -plumed
plumed.dat -multidir $mydirs -replex 500 -ntomp 1

It starts fine, but when I restart I get the incompatible systems error
(see below). I looked at the following discussion:
https://mailman-1.sys.kth.se/pipermail/gromacs.org_gmx-users/2013-November/085739.html,
but using -append no does not produce numbered cpt files and I would like
to avoid wasting every run this happens for. Is there a more current
solution?


Initializing Replica Exchange
Repl  There are 56 replicas:
Multi-checking the number of atoms ... OK
Multi-checking the integrator ... OK
Multi-checking init_step+nsteps ...
init_step+nsteps is not equal for all subsystems
  subsystem 0: 33291099
  subsystem 1: 33291099
  subsystem 2: 33291099
  subsystem 3: 33291099
  subsystem 4: 33291099
  subsystem 5: 33291099
  subsystem 6: 33291099
  subsystem 7: 33291099
  subsystem 8: 33291099
  subsystem 9: 33291000
  subsystem 10: 33291000
  subsystem 11: 33291000
  subsystem 12: 33291000
  subsystem 13: 33291099
  subsystem 14: 33291099
  subsystem 15: 33291000
  subsystem 16: 33291000
  subsystem 17: 33291000
  subsystem 18: 33291000
  subsystem 19: 33291000
  subsystem 20: 33291000
  subsystem 21: 33291099
  subsystem 22: 33291099
  subsystem 23: 33291099
  subsystem 24: 33291099
  subsystem 25: 33291099
  subsystem 26: 33291099
  subsystem 27: 33291000
  subsystem 28: 33291000
  subsystem 29: 33291099
  subsystem 30: 33291039
  subsystem 31: 33291039
  subsystem 32: 33291039
  subsystem 33: 33291039
  subsystem 34: 33291039
  subsystem 35: 33291039
  subsystem 36: 33291039
  subsystem 37: 33291039
  subsystem 38: 33291039
  subsystem 39: 33291039
  subsystem 40: 33291039
  subsystem 41: 33291039
  subsystem 42: 33291039
  subsystem 43: 33291039
  subsystem 44: 33291039
  subsystem 45: 33291000
  subsystem 46: 33291000
  subsystem 47: 33291039
  subsystem 48: 33291039
  subsystem 49: 33291039
  subsystem 50: 33291039
  subsystem 51: 33291039
  subsystem 52: 33291049
  subsystem 53: 33291049
  subsystem 54: 33291049
  subsystem 55: 33291049

-------------------------------------------------------
Program:     gmx mdrun, version 2019.4
Source file: src/gromacs/gmxlib/network.cpp (line 745)
MPI rank:    0 (out of 56)

Fatal error:
The 56 subsystems are not compatible