# Hamiltonian REX: different exchange probability with -nb gpu
**URL:** https://gromacs.bioexcel.eu/t/hamiltonian-rex-different-exchange-probability-with-nb-gpu/2688
**Category:** User discussions
**Tags:** mdrun
**Created:** [August 17, 2021, 6:09pm UTC](https://gromacs.bioexcel.eu/t/hamiltonian-rex-different-exchange-probability-with-nb-gpu/2688 "2021-08-17T18:09:03Z")
**Posts on this page:** 3
**Page:** 1
### Author:  [@mandarkulkarni](https://gromacs.bioexcel.eu/u/mandarkulkarni)
#### Post date: [August 17, 2021, 6:09pm UTC](https://gromacs.bioexcel.eu/t/hamiltonian-rex-different-exchange-probability-with-nb-gpu/2688/1 "2021-08-17T18:09:03Z")
GROMACS version: 2019.4
GROMACS modification: Yes (PLUMED 2.5.7 patched)
Dear all,
I am testing the hamiltonian replica exchange for a small RNA tetraloop system (NVT condition, hydrogen mass repartition 4 fs time step). As a part of testing, I am using the same scaled topology for four replicas to check exchange probability. The test run was performed for 400 ps. However, when I use the “-nb gpu” option with mdrun, I get a slightly low exchange rate. The simulations are repeated and the results are consistent.
Also, replex (250) is set as multiple of nstlist (25) and dynamic load balancing is also off (-dlb no).
Please, can you suggest whether using “-nb gpu” could affect energies and thus exchange probability in general?
mdrun command line:
mpirun -np 4 gmx\_mpi mdrun -v -deffnm md -plumed plumed.dat -replex 250 -hrex -multidir dir\_0 dir\_1 dir\_2 dir\_3 -dlb no -nb gpu -nsteps 100000 \>& log
**With “-nb gpu”:**
dir\_0/md.log:Repl average probabilities:
dir\_0/md.log-Repl 0 1 2 3
dir\_0/md.log-Repl .96 **.84**.97
dir\_0/md.log-Repl number of exchanges:
dir\_0/md.log-Repl 0 1 2 3
dir\_0/md.log-Repl 190 **172** 192
dir\_0/md.log-Repl average number of exchanges:
dir\_0/md.log-Repl 0 1 2 3
dir\_0/md.log-Repl .95 .86 .96
**without “-nb gpu”:**
…/No\_nb\_gpu/dir\_0/md.log:Repl average probabilities:
…/No\_nb\_gpu/dir\_0/md.log-Repl 0 1 2 3
…/No\_nb\_gpu/dir\_0/md.log-Repl .96 **.97**.97
…/No\_nb\_gpu/dir\_0/md.log-Repl number of exchanges:
…/No\_nb\_gpu/dir\_0/md.log-Repl 0 1 2 3
…/No\_nb\_gpu/dir\_0/md.log-Repl 198 **192** 197
…/No\_nb\_gpu/dir\_0/md.log-Repl average number of exchanges:
…/No\_nb\_gpu/dir\_0/md.log-Repl 0 1 2 3
…/No\_nb\_gpu/dir\_0/md.log-Repl .99 .96 .99
---
### Author:  [@alevilla](https://gromacs.bioexcel.eu/u/alevilla)
#### Post date: [August 18, 2021, 8:38am UTC](https://gromacs.bioexcel.eu/t/hamiltonian-rex-different-exchange-probability-with-nb-gpu/2688/2 "2021-08-18T08:38:50Z")
Hi,
Calculating the nonbonded interaction on gpu or not should not affect the results.
If you are running on the same machine and GROMACS is compiled accounting for gpu the non-bonded interactions are calculated on GPU with and without explicitly add the option -nb gpu).
My guess is that the difference you observed are due too not enough sampling.
Best regards
Alessandra
---
### Author:  [@mandarkulkarni](https://gromacs.bioexcel.eu/u/mandarkulkarni)
#### Post date: [August 19, 2021, 1:22pm UTC](https://gromacs.bioexcel.eu/t/hamiltonian-rex-different-exchange-probability-with-nb-gpu/2688/3 "2021-08-19T13:22:05Z")
Hi Alessandra,
Thanks for your reply. I agree that I should observe the same results.
I observed these results by repeating these tests, so I was a bit worried.
I will run longer simulations and will post the outcome here.
Thanks,
Mandar Kulkarni