GitLab

* Do not add constraints involving extra particles

* Added test case

* Max Drude distance defaults to 0.2 nm

* Fixed incorrect number in docs

* Fixed a test case

* Access atom element field

* Update GB-Neck2 force regression test

It allows to use a faster float-to-int64 in the HIP platform.

* Use importlib_metadata instead of pkg_resources

* Use built in importlib when available

* Revert change that didn't work

* Minor optimizations to computing single pairs

* Adjusted MAX_BITS_FOR_PAIRS on Ampere

* Ensure bond lengths are set before creating forces that depend on them

* Minor formatting

* Only call set switching function/distance on nonbonded force(s)

* Also catch CustomNonbondedForce

* Fix logic around setEwaldErrorTolerance

* DrudeLangevinIntegrator has getSystemTemperature()

* DrudeNoseHooverIntegrator has getSystemTemperature()

* StateDataReporter reports system temperature for Drude systems

* Fixed incorrect return type

* Add library path with os.add_dll_directory()

* Update path for Visual Studio

* Improved handling of CRYST1 records #3446

* Apply suggestions from code review
Co-authored-by: Jason Swails <jason@entos.ai>
Co-authored-by: Jason Swails <jason@entos.ai>

* Added documentation to clarify behavior of nonbonded exceptions

* Added documentation to clarify behavior of nonbonded exceptions

* Reference implementation of computed values for CustomNonbondedForce

* CPU implementation of computed values for CustomNonbondedForce

* Common implementation of computed values for CustomNonbondedForce

* Serialization of computed values

* ForceField supports computed values

Without this restriction, the compilation fails if the project folder has parent folders called "tests", "internal" or other words in these generic patterns.

* Fixed potential invalid memory access

* Fixed exception

* Support prmtop files created by chamber

* Support CHARMM CMAP flags

* Changed handling of SCEE and SCNB scale factors

* fix barostat frequency check

* fix other barostats

* fix error messages

* Add Fortran type translation for `long long`

* Replaced tabs with spaces
Co-authored-by: Peter Eastman <peter.eastman@gmail.com>

* Optimized CudaSort for non-uniformly distributed data

* Optimized OpenCLSort for non-uniformly distributed data

* Further tuned distributing elements between buckets

* Copied optimizations over to OpenCL

* Try older miniforge version

Use miniforge version that was known to be working

* ci bump

* switch back to using latest miniforge3 + cleanup miniconda folder at the end of testing for the self-hosted runner'

* ci bump

* ci bump

* Minor clarification to manual

* Another minor clarification