You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This is related to why #684 needed fixing, and hopefully some of the alchemical pains here will go away with #672
Bug report
We noticed that the FIREMinimizationIntegrator applied to an explicitly solvated system will yield NaNs when using a CPU platform, but will run fine on CUDA, OpenCL, and Reference.
Minimum reproducible code:
from openmmtools.integrators import FIREMinimizationIntegrator
from openmmtools import testsystems
import openmm
t = testsystems.AlanineDipeptideExplicit()
system, positions = t.system, t.positions
integrator = FIREMinimizationIntegrator()
platform = openmm.Platform.getPlatformByName('CPU')
context = openmm.Context(system, integrator, platform)
context.setPositions(positions)
integrator.step(10000)
state = context.getState(getEnergy=True)
state.getPotentialEnergy()
Error output:
OpenMMException Traceback (most recent call last)
Cell In[1], line 11
9 context = openmm.Context(system, integrator, platform)
10 context.setPositions(positions)
---> 11 integrator.step(10000)
12 state = context.getState(getEnergy=True)
13 state.getPotentialEnergy()
File ~/software/mambaforge/install/envs/openfe-latest/lib/python3.10/site-packages/openmm/openmm.py:13583, in CustomIntegrator.step(self, steps)
13573 def step(self, steps):
13574 r""" 13575 step(self, steps) 13576 Advance a simulation through time by taking a series of time steps. (...) 13581 the number of time steps to take 13582 """> 13583 return _openmm.CustomIntegrator_step(self, steps)
OpenMMException: Particle coordinate is NaN. For more information, see https://github.com/openmm/openmm/wiki/Frequently-Asked-Questions#nan
Oh, I've been trying to get an example on how to reproduce this error because other users have also encountered it. Can you please share the scripts and data to reproduce this? It should help us test and debug behavior when we work on #672 . Thanks for reporting!
Context
This is related to why #684 needed fixing, and hopefully some of the alchemical pains here will go away with #672
Bug report
We noticed that the FIREMinimizationIntegrator applied to an explicitly solvated system will yield NaNs when using a
CPU
platform, but will run fine onCUDA
,OpenCL
, andReference
.Minimum reproducible code:
Error output:
mamba list output:
The text was updated successfully, but these errors were encountered: