GitLab

* CPU platform checkpoints random number generator

* Fix Windows compilation error

* Another Windows compilation error

* Use Intel OpenCL for CI

* Set environment variables

* Try to get CI to run

* Debugging

* Debugging

* Fixes for Intel OpenCL

* updateParametersInContext() can modify parameter offsets

* Reordering respects parameter offsets

* Implemented for CUDA and HIP

* Cleanup of AMOEBA code

* Fixed another compilation error

* More Windows compilation errors

* Fixed incorrect syntax

* Update the AMOEBA OpenMM API for vdW, GK and WCA

* Changes needed for the Corrigan et al Generalized Kirkwood model and minor changes to the vdW force to support CpHMD

* Add casts to real for uses of POW in GK; Pass force by reference within the WCA kernel

* Update swigInputConfig for Amoeba vdW and GK forces

* Update TestAPIUnits.testAmoebaVdwForce

* Set the units for getSolventDielectric and getSoluteDielectric to None

* Update default dispersion offset parameter for the AmoebaWcaDispersionForce

* Remove overloaded getParticleParameters and setParticleParameters from AmoebaGeneralizedKirkwoodForce

* Update the AmoebaWcaDispersionForce TestAPIUnits tests to reflect using the correct units for the C++ parameter default values; Update the alanine-dipeptide-amoeba-forces to reflect the updated GK model

* Move neck descreening constants into AmoebaGeneralizedKirkwoodForceImpl; set the default GK dielecticOffset to 0.09; set the default WCA shctd parameter to 0.82

* Fix Python test cases for WCA and GK

* Load AMOEBA/GK parameters into an array of float4

* Cleaned up the AmoebaGeneralizedKirkwoodForce based on feedback from Peter; the one case where backwards compatibility remains a challenge is application of the dielectric offset parameter - in the prior code this was only applied to the nonpolar cavity term, but not to calculation of Born radii; in this revision the dielectric offset is applied to BOTH the nonpolar cavity term and to calculation of Born radii. At this point I can't think of elegant way to maintain backwards compatibility that isn't confusing, nor does it make sense (at least to me) to only apply the concept of the dieletric offset to one aspect (i.e. only to nonpolar cavity or only to Born radii calculation) but not to both.

* Remove 'using std::vector' from AmoebaGeneralizedKirkwoodForceImpl.h; divide by 10 instead of multiplying by 0.1f in amoebaGk.cc

* Added a parameter called descreenOffset, which is applied during calculation of effective Born radii for GK. The parameter dielectricOffset is only used for the nonpolar cavity term consistent with its prior use. All tests in TestAmoebaGeneralizedKirkwoodForce.h are now backwards compatible with their behavior prior to this PR.

* Change two constants in amoebaGk.cc to single precision; Improved the documentation for getNeckConstants in AmoebaGeneralizedKirkwoodForceImpl.h

* Fix comment for setTanhRescaling in AmoebaGeneralizedKirkwoodForce.h, Fix comment for setTanhParameters in AmoebaReferenceGeneralizedKirkwoodForce.h; set the type of parameter GeneralizedKirkwoodTanhRescaling to bool in AmoebaGeneralizedKirkwoodForceProxy.cpp; In ReferenceCalcAmoebaGeneralizedKirkwoodForceKernel return references of per particle parameters instead of copies; update AmoebaReferenceKernels.h method signatures for per particle parameters to return const vector references

* Minor tweaks to the documentation for the tanh rescaling flag

* Improve the comments for the get and setTanhParameters in AmoebaGeneralizedKirkwoodForce.h and AmoebaReferenceGeneralizedKirkwoodForce.h

* Reduced memory use while identifying molecule groups

* Further reduce memory use

* Benchmark enhancements

- Add `--disable-pme-stream` to benchmark.py to easily compare performance without having to build with different default values for `DisablePmeSteam`
- Replace utcnow to remove DeprecationWarning
- Check for nvidia-smi before attempting to launch it
- Mention HIP where appropriate

* Adjust option

* Updated Python versions for CI

* Install setuptools

* updated units in energy minimisation user guide

* Made units more readable

* added type checking for Simulation.step()

* changed how to check if step is an integer number

* allow for dicts to be returned from Reporter.describeNextReport
remove deprecated getState parameters ( #4437 )

* convert old format into new format

* update docstring

* nested set comprehension to set.union

* Allow 'periodic':None
update describeNextReport in all occurrences in the code

* debug

* update documentation

* add a reporter for energyParameterDerivative

* Revert "add a reporter for energyParameterDerivative"

This reverts commit 1d44dc3f60153defb6252ab56a3b85350fa24826.

* Edit documentation

* Use miniforge instead of mambaforge

* Fixed version number

* Try miniforge-variant

* Added documentation on HIP

* Instructions for building HIP on Windows

* PME_ORDER threads process one atom;
* PME_ORDER threads access consecutive addresses;
* No need to permute z indices with zindexTable;
* finishSpreadCharge is needed only with fixed point charge spreading;

* XTCReporter overwrite existing xtc file

* add position size check for xtcfile

* Unified lots of parallel computation code between platforms

* Unified test code between platforms

* Eliminated duplicated timing code

* Optimize CustomNonbondedForce.updateParametersInContext()

* Optimized uploading changed values to GPU

* Optimized updateParametersInContext() for lots of bonded forces

* Optimized updateParametersInContext() for CustomExternalForce

* Optimized updateParametersInContext() for NonbondedForce

* Code changes for HIP platform

* CI build for HIP

* Removed some unneeded dependencies

* Fix test for which libraries are present

Prevent deadlock using CustomCPPForceImpl with multiple GPUs

HIP platform

CustomCPPForceImpl for writing forces in C++

    https://github.com/openmm/openmm/commit/9a0db72
    https://github.com/openmm/openmm/pull/4231

Virtual sites can depend on other virtual sites

    https://github.com/openmm/openmm/commit/71f4b3f

Use LF-Middle for LangevinIntegrator and VariableLangevinIntegrator

    https://github.com/openmm/openmm/commit/86988b9

Merged more code into common platform

    https://github.com/openmm/openmm/commit/5739788

    * Common implementation of BondedUtilities
    * Common implementation of UpdateStateDataKernel

Fixed periodic box changing from rectangular to triclinic

    https://github.com/openmm/openmm/commit/75d4f29

Skip neighbor list for very small systems

    https://github.com/openmm/openmm/pull/4070

Store bounding box sizes in half precision

    https://github.com/openmm/openmm/commit/2ae50f9

Use large blocks to optimize building the neighbor list

    https://github.com/openmm/openmm/commit/3955033

Improved sorting of blocks when building neighbor list

    https://github.com/openmm/openmm/commit/796ffaa

Fixed bug in large blocks optimization with triclinic boxes

    https://github.com/openmm/openmm/commit/4c10732

Optimize sorting of non-uniformly distributed data

    https://github.com/openmm/openmm/commit/71d9bb1

Co-authored-by: bdenhollander <44237618+bdenhollander@users.noreply.github.com>

Co-authored-by: Emilio Gallicchio <emilio.gallicchio@gmail.com>

Use a small kernel for copying interactionCounts to host memory

    hipMemcpy's CopyDeviceToHost operation has higher latency.

Do not set stream and event blocking/spin related flags

    Let the runtime choose the best option because overriding does not
    improve performance in most cases.

Remove NULL streams and use nonblocking streams explicitly

Make HipContext::pushAsCurrent/popAsCurrent thread-safe as they can be
called simultaneously from different threads via ContextSelector.

Allow peer access to be enabled more than once (if there are multiple
simulations one after another, like in benchmark.py).

Create peerCopyStream on a corresponding device

Use two-speed load balancing for multi GPU runs

    First 100 steps do coarse balancing, next 100 - fine tuning.
    Also ignore the slowest device (usually 0) if its fraction has
    reached 0, (i.e. no work can be transfered to other devices) and
    balance other devices.

Do not download inteactionCounts in parallel nonbonded tasks

    This is not required because updateNeighborListSize has been called
    and valid flag changed.

Initialize tilesAfterReorder properly

    It may contain a garbage value, and if it is large then
    updateNeighborListSize does not force reorder atoms after 25 steps
    in extremal cases.

Use cuCtxPushCurrent() and cuCtxPopCurrent() for selecting CUDA context

    https://github.com/openmm/openmm/pull/3258

Fixed uninitialized memory access

    https://github.com/openmm/openmm/issues/3392
    https://github.com/openmm/openmm/pull/3399

Fixed potential invalid memory access

    See https://github.com/openmm/openmm/pull/3428

Improved temperature reporting for Drude particles

    https://github.com/openmm/openmm/pull/3486
    https://github.com/openmm/openmm/commit/a5e42f5

Fixed race condition with multiple GPUs

    https://github.com/openmm/openmm/commit/6fb1c8a41edff980862750bc086f6a204eb50941

Use blocking sync when creating events

    https://github.com/openmm/openmm/commit/fe21d5ee4f14673a4ea38b7244991772a64ceec2

Very minor optimizations

    https://github.com/openmm/openmm/commit/109f6b2535da4e0c0dd88007d6ca06b4add2ce81

Use PocketFFT

    https://github.com/openmm/openmm/commit/1dac981a63300a2a53a7925f570995914f7163ed

Improved logic for deciding when to reorder atoms

    https://github.com/openmm/openmm/commit/48664a1f1a4490a4dabc277757545ac070e7b898

Ensure valid atom order after loading a checkpoint

    https://github.com/openmm/openmm/commit/a056d5a3754e193105409afa12c9f0c9a2d972a2

Improve performance running on multiple GPUs

    https://github.com/openmm/openmm/commit/0c82c2647de98da5c6dab7bf7a7b8b19705aadc0

Fixed errors when running on multiple GPUs

    https://github.com/openmm/openmm/commit/ed9df876d43c037c08d4762721e73e5caae086d9

Optimized reducing energy

    https://github.com/openmm/openmm/commit/2975f44

* VkFFT-based 3D FFT;
* Caching of compiled VkFFT kernels;
* Extend FFT tests with more sizes.