MP2C

Massively Parallel Multi-Particle Collision Dynamics

Scientific area

Soft Matter Physics and mesoscopic hydrodynamics.

Short description

MP2C implements a hybrid representation of solvated particles in a fluid. Solutes are simulated atomistically by classical molecular dynamics which is coupled to the solvent, described by the Multi-Particle-Collision-Dynamics method (MPC). The MPC part of the program can be applied as stand-alone implementation for particle-based hydrodynamics, which takes into account thermal fluctuations. Solutes can mutually interact via short range and long range electrostatics interactions, where the long range part is handled by the external scalable library ScaFaCoS, which was developed in a JSC coordinated German network project. Parallelisation of MP2C is currently based on MPI and implements a domain decomposition, which can dynamically be adjusted by an adaptive load-balancing scheme in case of inhomogeneous systems. Scalable I/O is achieved by the external library SIONlib, which is used for e.g. large data check-pointing and trajectory files.

Flow field in a gas diffusion layer (GDL) of a polymer electrolyte fuel cell, modelled by MPC fluid in a matrix of random cylinders (not shown for clarity).

MP2C
MP2C

Scalability

  • 458,752 cores on BlueGene/Q
  • 262,144 cores on BlueGene/P
  • 17,664 cores (JUROPA cluster)

Programming language and model

  • Fortran
  • MPI
  • SIONlib for file I/O

Tested on platforms

  • BlueGene/Q and /P
  • x86 (JUROPA cluster / JUDGE cluster)
  • CRAY XT4/XT5

Application developers and contact

Simulation Laboratory Molecular Systems
Jülich Supercomputing Centre
Forschungszentrum Jülich GmbH
52425 Jülich
Germany

g.sutmann@fz-juelich.de

Last Modified: 10.05.2022