Currently the order of particles in memory will not be consistent if you run the same simulation twice on the GPU. This usually doesn't matter, but for cases where it does, this adds a slower, stable option.

The proposed changes:

• fix a bug or incorrect behavior in AMReX
• add new capabilities to AMReX
• changes answers in the test suite to more than roundoff level
• are likely to significantly affect the results of downstream AMReX users
• include documentation in the code and/or rst files, if appropriate