**Proposed layout for 'multidir' scaling plots**
The left axis shows the individual (per system, blue circles) and cumulative performance (green crosses), while the left axis shows the performance relative to the 1GPU & 1 system case. Notes:
- Gromacs requires ranks to be a multiple of the number of simulations. Here, they were set equal.
- In all cases, the maximum allowed threads gave the best performance. Still, the optimal combination should be tested.
- All possible interactions of the Martini test system (no PME!) were offloaded to gpu using
srun gmx_mpi mdrun -nb gpu -bonded gpu -v -maxh 0.25 -resethway -noconfout -deffnm topol

Originally posted by @balazsfabian in #179 (comment)
The left axis shows the individual (per system, blue circles) and cumulative performance (green crosses), while the left axis shows the performance relative to the 1GPU & 1 system case. Notes:
srun gmx_mpi mdrun -nb gpu -bonded gpu -v -maxh 0.25 -resethway -noconfout -deffnm topolOriginally posted by @balazsfabian in #179 (comment)