Well, you're in luck. OpenMM 6.0 RC has a static builds of OpenMM, OpenMMCUDA, OpenMMOpenCL, OpenMMCPU platforms, in addition to the PME Plugin. We hoped this would cover most use cases. Set the cmake OPENMM_BUILD_STATIC_LIB option to true, which will the generate *_static libraries. Note that to us...

Hi Stephan,

OpenMM does not parallelize across multiple nodes in a cluster environment. It's designed for single node use with powerful GPUs and CPUs.

Hi,

You can use force groups to separate each force into a different group. When you call getState later on, you can query by the forceGroup. Take a look at the API in the docs, and if you have trouble understanding it let us know

Unfortunately this information isn't exposed directly. In fact, we don't even build a canonical neighborlist on the GPU. We find neighbors for a block of atoms instead of a single atom. The relevant code are located here: https://github.com/SimTk/openmm/blob/master/platforms/cuda/src/kernels/findInt...

Hi Peter and Mark, When I moved the svn repo into github, to avoid transferring literally everything in the root of /svn/openmm, I instead copied over /svn/openmm/trunk/ - which contained all of the pertinent information in that folder. The tags folder IIRC was in /svn/openmm/tags which wasn't copie...

Hi Carl, Context class's createCheckpoint creates a checkpoint that is only loadable on the exact computer used to create the checkpoint. It basically writes out to a binary stream (or string in case of python) containing various internal states of the system. ie. it tries to maintain the same GPU s...