"Energy is NaN" error with AMBER prmtop and restart file

Carl Benzer · Post by **Carl Benzer** » Fri Oct 31, 2014 5:27 am

Hi,

i need some help figuring out what's wrong here. I manually compiled 6.1 with CUDA 6.5. All tests passed and all example *.py scripts run fine. When I try to use a specific AMBER prmtop and restart file created with AmberTools 14 on the CUDA and OpenCL platform, I get the following error:

Code: Select all

#"Step","Potential Energy (kJ/mole)","Temperature (K)"
Traceback (most recent call last):
  File "simulateAmber.py", line 17, in <module>
    simulation.step(5)
  File "/usr/lib64/python2.6/site-packages/simtk/openmm/app/simulation.py", line 127, in step
    reporter.report(self, state)
  File "/usr/lib64/python2.6/site-packages/simtk/openmm/app/statedatareporter.py", line 158, in report
    self._checkForErrors(simulation, state)
  File "/usr/lib64/python2.6/site-packages/simtk/openmm/app/statedatareporter.py", line 301, in _checkForErrors
    raise ValueError('Energy is NaN')
ValueError: Energy is NaN

The above simulation runs just fine on the Reference and CPU platform. The prmtop and inpcrd file from the example directory do work fine on all platforms.

Any hints or ideas?