Evan Pretti

Some notes after looking over the API documentation and theory guide: - For many classes, the theory guide just has an equation that could simply be moved over to the...

Without all of the input files, I can't test to verify what I believe may be the problem, but your PDB file appears to use the Amber convention for lipids,...

Right, if the Topology doesn't have bonds, OpenMM won't be able to match residue templates to what it considers "non-standard" residues. There is actually a mechanism that OpenMM uses to...

There is a check that was added a long time ago (in #1664) that is supposed to prevent overflows like this from happening. Without knowing what's in the script you're...

It looks like you're missing a call to `getState()` after the last `setPositions()` and before the last `getPotentialEnergy()`, so the energy returned is from the previous state evaluation when the...

You're still using the old `state`; you'll need `state = sim.context.getState(energy=True)`. Taking a closer look at the force field, doing that will still give you 0, though, because the `Discrete1DFunction`...

At one point we had updated the behavior of the other template generators to use user charges as is if they are present, and otherwise assign charges per the preferred...

Absolutely, we should have a test case that covers all of the cases and checks that the appropriate charges get assigned. I should be able to fix this; I just...

There are currently at least two remaining issues with checkpoints in OpenMM from a user experience standpoint: - `CheckpointReporter` overwrites checkpoints atomically, but `simulation.saveCheckpoint()` does not, so when using it...

I believe that the `charmm36.xml` file was generated using an old version of the [conversion tools](https://github.com/openmm/openmmforcefields/tree/main/charmm) in openmmforcefields. When I updated them to handle the new 2024 release of the...