HOOMD-blue is a general-purpose particle simulation toolkit. It scales from a single CPU core to thousands of GPUs.
You define particle initial conditions and interactions in a high-level python script. Then tell HOOMD-blue how you want to execute the job and it takes care of the rest. Python job scripts give you unlimited flexibility to create custom initialization routines, control simulation parameters, and perform in situ analysis.
Download and get started using HOOMD-blue today. Please cite HOOMD-blue if you use it published work.
import hoomd, hoomd.md
hoomd.context.initialize()
unitcell=hoomd.lattice.sc(a=2.0, type_name='A')
hoomd.init.create_lattice(unitcell=unitcell, n=10)
nl = hoomd.md.nlist.cell()
lj = hoomd.md.pair.lj(r_cut=3.0, nlist=nl)
lj.pair_coeff.set('A', 'A', epsilon=1.0, sigma=1.0)
all = hoomd.group.all();
hoomd.md.integrate.mode_standard(dt=0.005)
hoomd.md.integrate.langevin(group=all, kT=1.2, seed=4)
hoomd.run(10e3)$ hoomd run.py --mode=cpu
$ hoomd run.py --mode=gpu
$ mpirun -n 256 hoomd run.py --mode=cpu
$ mpirun -n 64 hoomd run.py --mode=gpu
Fast GPU performance
On a single NVIDIA GPU, HOOMD-blue performs an order of magnitude faster than a multi-core CPU in typical benchmarks. Obtain faster time to solution, execute larger simulations than were previously possible, or increase job throughput through a cluster given a fixed monetary budget.
Scalable
HOOMD-blue scales up to thousands of GPUs on Titan and Blue Waters, two of the largest GPU accelerated supercomputers in the world. It runs just as efficiently from one to dozens of GPUs on smaller clusters like Comet (XSEDE), Maverick (XSEDE), your local GPU cluster, and the workstation at your desk.
Flexible
Want to run a Molecular Dynamics simulation using a custom force field? Or maybe you are simulating granular dynamics, dissipative particle dynamics, hard particle Monte Carlo, or active particles? HOOMD-blue is the right tool for you, with a variety of potentials and integration modes.
Hard particle Monte Carlo capabilities
Execution
- Canonical hard particle MC on a single CPU core
- Parallel update scheme on a single GPU
- Parallel updates on many CPU cores / GPUs using MPI
- Frenkel-Ladd free energy determination
- Wall constraints (sphere, cylinder, plane)
Shape classes
- Spheres / disks
- Union of spheres
- Convex polygons
- Convex spheropolygons
- Simple polygons
- Ellipsoids / ellipses
- Convex polyhedra
- Union of convex polyhedra
- Convex spheropolyhedra
- Faceted spheres
- General polyhedra
Ensembles
- NVT
- NPT
- Implicit depletants
- Grand canonical ensemble
- Gibbs ensemble
Analysis
- Pressure sampling in NVT
- Free volume
Molecular dynamics capabilities
Integrators
Apply many independent integrators to separate particle groups. HOOMD-blue has integrators built in for many different thermodynamic ensembles and energy minimization. Many of them support integration of orientational degrees of freedom.
- NVE, NVT, NPH, NPT
- Langevin dynamics
- Brownian dynamics
- Berendsen thermostat
- Dissipative Particle Dynamics (DPD)
- FIRE energy minimization
Pair potentials
- Buckingham
- CGCMM
- DLVO
- DPD
- Lennard-Jones
- Gaussian
- Mie
- Moliere
- Morse
- Reaction field
- Yukawa
- ZBL
- User-defined (table)
Bond potentials
- Harmonic
- FENE
- User-defined (table)
- Special pair lj
- Special pair coulomb 1-4
Angle potentials
- Harmonic
- CGCMM
- cosine squared
- User-defined (table)
Rigid bodies
Combine particles together to form rigid bodies. Potentials still evaluate particle to particle, but the integrator keeps particles in rigid bodies together. Most of the standard integrators work in conjunction with the rigid body constraint.
Advanced neighbor list algorithms
Efficiently handle large cutoff radius disparities using a BVH tree, and/or stenciled cell list to compute the neighbor list. These algorithms make possible large colloid and explicit solvent simulations with.
Other forces
- Active force
Anisotropic potentials
- Gay-berne
- Dipole
- DEM (faceted shapes)
Wall potentials
- Sums of planar, cylindrical, and spherical half-spaces
- lj, gauss, slj, yukawa, morse, force_shifted_lj, and mie interactions
Long-ranged potentials
- PPPM electrostatics
- PPPM with Debye screening
Constraints
- Sphere
- Ellipsoid
- Distance
- Rigid body
- One dimensional motion
Dihedral potentials
- Harmonic
- OPLS
- User-defined (table)
Improper potentials
- Harmonic
External potentials
- Periodic
- Constant force
- Electric field
Many-body potentials
- EAM
- Tersoff
- square density
General features
Simulation box
Simulations in HOOMD-blue can be 2D or 3D. The box is a parallelepiped with arbitrary lengths and angles allowing for generic triclinic symmetry with periodic boundary conditions in all directions.
Initializers
Extend HOOMD-blue
Need a capability not listed here? Fork the project and modify the code. HOOMD-blue is written in well documented C++ and its design is modular. Extensions that just add a new pair or bond potential don't even require CUDA code.
If your work is useful to a wide audience of users, submit it to the HOOMD-blue developers as a pull request for inclusion into the main code.
Update properties
- Resize box over time
- Temperature rescale
- Zero system momentum
- Arbitrary manipulation of system and particle properties from python
- Add/remove particles while the simulation progresses
- Access particle properties via numpy arrays
- Introduce flow with the Mueller/Plathe method
Analysis
- Realtime visualization in VMD using IMD.
- Mean-squared displacement
- Log thermodynamic properties to text or HDF5
- Query logged values from python
- Arbitrary analysis in python
- Execute inside Jupyter notebooks
News
- 2017/09/08 - HOOMD-blue v2.2.0 released.
- 2017/08/22 - HOOMD-blue v2.1.9 released.
- 2017/07/19 - HOOMD-blue v2.1.8 released.
- 2017/05/11 - HOOMD-blue v2.1.7 released.