Version 2.75a

Incorporates a few patches that were supposed to be included in release 2.75.

What's Changed

• Fix Issue #434: If defined, use Brownian Motion MP moves with NeMT by @LSchwiebert in #435
• Fix Issue #442 Tail Corrections Description by @LSchwiebert in #444
• change version and release date in README.md by @GregorySchwing in #446

Full Changelog: v2.75...v2.75a

Version 2.75

This version includes:

• Added support for Brownian-motion multiparticle move.
• Added fully NAMD file support, including *.dcd, *.vel, *.xsc and *.coor files.
• Additional improvements to the GPU performance.
• Improved error checking of the configuration files.
• Many bug fixes.

Note: We have switched from the GPL License to the MIT License. If you wish to continue to use GOMC under the GPL license, please do not upgrade.

What's Changed

• Multi-particle brownian motion by @msoroush in #160
• Added header to CUDAMemoryManager.cu by @YounesN in #247
• Update development with master changes by @YounesN in #248
• fixes issue #250 by @YounesN in #251
• DCD coordinate output by @msoroush in #252
• Fix to #256 by @YounesN in #257
• Read in restart binary coordinate and extended system file by @msoroush in #259
• Reading Multi parameter files by @msoroush in #260
• redirect error when intel not present #262 by @YounesN in #264
• Fix to ERROR: Error closing DCD file` at the end of simulation by @msoroush in #268
• Added atom equality operator to allow for user defined chemical poten… by @GregorySchwing in #270
• Simplified the logic between two box psf parser. by @GregorySchwing in #271
• Added googletest by @YounesN in #272
• Fixed psf restart headers by @GregorySchwing in #277
• Prints actual residue for proteins in PDB Restart file. Should be re… by @GregorySchwing in #278
• added the extra terms namd requires at the end of a psf file (imprope… by @GregorySchwing in #279
• Changinf xsc file name by @msoroush in #280
• Changed how we retrieve a moleculeKind. Instead of indexing into the… by @GregorySchwing in #286
• Reject swap/intra-swap move for bad configuraiton by @msoroush in #288
• Need to allocate one more character for null character by @YounesN in #291
• Recip patch by @LSchwiebert in #297
• Fixed checkpoint system by @YounesN in #298
• Probabilities and Coefficients are calculated using only swappable molecules for molecule transfer and MEMC. by @GregorySchwing in #301
• PSF Parser can handle MosDef format Porous Materials by @GregorySchwing in #304
• Revert "Merge pull request #304 from GregorySchwing/rebasedpsfparsert… by @GregorySchwing in #305
• Fixes compilation of GPU on windows by @YounesN in #303
• split psf and pdb parsers. We only use psf for determining number of… by @GregorySchwing in #309
• Faster-Math should have current Development Branch Code by @LSchwiebert in #310
• Fix to #issue312 by @YounesN in #317
• Circuit by @GregorySchwing in #320
• Fix to the BoxDimensions initialization from XSC files by @msoroush in #322
• Profiling GOMC with NVTX by @msoroush in #315
• Targeted swap by @msoroush in #293
• MultiParticle Brownian like motion by @msoroush in #327
• Revert "MultiParticle Brownian like motion" by @GregorySchwing in #332
• Revert "Revert "MultiParticle Brownian like motion"" by @GregorySchwing in #333
• Fix issues with MultiParticle Brownian. by @msoroush in #334
• Fix to the issue #338 and issue #335 by @msoroush in #339
• Reading, updating, and outputting binary Velocity file by @msoroush in #341
• Update readme by @msoroush in #342
• Manual Update to Correct Typo by @LSchwiebert in #347
• Consistent Molecule Trajectories across Checkpoints and Checkpoint is performed using Boost Serialization Library by @GregorySchwing in #351
• quick fix to use right index for kinds in InitPartVec by @GregorySchwing in #352
• Added methods for parsing all of the PSF (although NNB, NGRP, NCRTRM are currently deactivated) by @GregorySchwing in #353
• Added Config parameter "ExpertMode bool" (default false) by @GregorySchwing in #357
• Native header only serialization library used if boost not found by @GregorySchwing in #358
• Only support Cereal serialization to avoid additional build complexity of linking a compiled library by @GregorySchwing in #364
• Implemented box specific enthalpy and compressibility calculations and printing to block averages, as well as compressibility to log files. by @dyukovsm in #362
• Non-Equilibrium Molecule Transfer Monte Carlo by @msoroush in #369
• Z and H in log file by @dyukovsm in #367
• Add support for IntraTargetedSwap move by @msoroush in #371
• Continuous trajectory GOMC in py-MCMD tested and supported, Added "InitStep #" argument to conf file by @GregorySchwing in #374
• Define CODEOWNERS file to define who can approve updates by @GOMC-WSU in #376
• Revise qqFact to be a more accurate constant by @LSchwiebert in #381
• Fix #383: Resolve Warnings and Errors for Windows builds by @LSchwiebert in #385
• Fix #386: Allow GOMC to run on older GPUs by @LSchwiebert in #387
• Fix #380: More accurate qqFact with qqFact as a constant on the GPU by @LSchwiebert in #389
• Fix hvap and prevent undefined behavior from using unitialized variables by @GregorySchwing in #397
• Resolve Issue #375 - Restore original PDBAtom, PSF Data, and MolLookup data structures on checkpoints. by @GregorySchwing in #382
• Sep res name from map key and check MEMC Freq. Fixes issues #390 and #391 by @GregorySchwing in #395
• Improve CUDA Memory Manager behavior for some edge cases -- Fixes issue #401 by @LSchwiebert in #402
• Development by @GregorySchwing in #406
• Fix #407 Mie potential with non-integer repulsion exponents by @LSchwiebert in #408
• #392 Generate per-step output for energy and move information in debug mode by @LSchwiebert in #410
• Fix #414 GOMC prints console headers twice when printing every step by @LSchwiebert in #415
• Fix #379 Cleaner error messages in MolSetup.cpp by @LSchwiebert in #417
• Restart pdb uses the atoms' input occupancy instead of overwriting occ with 0. by @GregorySchwing in #413
• Made OCC PR a template by @GregorySchwing in #418
• Fix #405: Issue warning for free energy calculations if RCutLow != 0.0 by @LSchwiebert in #420
• Add support for Impulse Pressure Correction (IPC) by @msoroush in #422
• Fix issue #377 Make Multiparticle reversible by @LSchwiebert in #412
• Fix #421 GOMC terminates on input Checkpoint False by @LSchwiebert in #423
• Fix some minor issue in Multiparticle and brownian like Multiparticle by @msoroush in #424
• Fix #425 use the correct recip energy for IntraMEMC move acceptance by @LSchwiebert in #426
• Update license by @GregorySchwing in #430
• GOMC Release 2.75 by @GOMC-WSU in #431
• Add #include <stdint.h> to Endian.h by @GregorySchwing in #432

New Contributors

• @dyukovsm made their first contribution in #362

Full Changelog: v2.70...v2.75

GPU performance improvement

• This release focuses on GPU performance improvements
• Using Random123 to be able to generate the same random number on GPU and port Translation and Rotation of molecules to the GPU
• Upgraded to C++14
• Lots of bug fixes
• Upgraded our CMake to 3.8 and use built-in CUDA support

Version 2.60

This version introduces:

• Changed the way we calculate pair interactions. The new approach allows us to optimize GPU by reducing the amount of cudaMemcpy needed.
• Added C++11 support to our code base and removed any using namespace std; statements.
• GPU memory management class to watch the allocation and deallocation of GPU memories.
• Bug fixes related to lambda functionalities on GPU

Version 2.50

• Added support for force biased multiparticle move
• Added support for free energy calculations using TI and FEP methods
• Support for multiple simulation with different temperatures using MPI
• Added support for Exponential-6 forcefield
• Support for restarting simulation using checkpoint
• Added support for new GPU architectures 7.0 and 7.5
• Added an error message when GPU version was used but there was no GPU present (#126)
• Read PDB file more efficiently when recalculating trajectory (#131)
• Fixed detecting simulation box shape (orthogonal and non-orthogonal) (#134)
• Fixed bugs in PDB reader and support for HETATM field in PDB file
• Fixed bugs in PSF reader
• Fixed Case sensitive keywords (#100)
• Fixed the timing report on Windows (#89)
• Added minimum volume in case the difference in box sizes were large (#94)
• Fixed the issue where cutoff value is larger than half of minimum box length (#98)
• Added an error message when charges detected and ewald was set to false (#99)
• Fixed compiling issue on Visual Studio (#110)
• Fixed the issue where GOMC did not read CHARMM parameter files missing Urey-Bradley terms (#147)
• Reduced GOMC reposity size

v2.40

Main features included in this version:

• Support for cyclic molecules
• Inter Molecular Exchange Monte Carlo move in GCMC and GEMC simulation
• Intra Molecular Exchange Monte Carlo move
• Crankshaft move
• Console log format changed to scientific

Hotfixes

• Compiling problem fixed on CYGWIN
• Fix to the error checking of volume exchange if simulation volume became negative
• Fix the equation to impose fix angle if the angle is less than 90. Generate error if constrained angle is not possible.
• Fixed a bug where move timings were 0 in windows.
• Fixed the issue where input files imported from NVT to NPT had zero volume, default value of 500 will be assigned in those cases.

Regrowth move

• Added Regrowth move
• Cleaned up the code and removed any memory leaks and bugs.
• Redesigned Ewald, EwaldCached, and NoEwald to be more efficient and allocate less memory.
• Added an error when multiple atoms have zero coordinates.
• Added an error when reservoir is empty and should print an error message and exit.
• Fixed some typo in output.
• Fix a bug where seed number output and input was not the same size, so we changed int to uint while reading to avoid overflow. They both use uint now.
• Increase memory allocation for NPT simulation in Ewald.
• Fixed a bug in PickWeighted() function where it was return a value larger than the array.
• Printing individual timing information for each move.
• To avoid getting large energy we are recalculating the total system at equilbrium step.
• Print CBMC information including first atom trial, secondary atom trials, angle trials, and dihedral trials.
• Fix a bug where in windows the clock() function was returning a wrong value
• Fix compiler warning for macOS in ConfigSetup file

Non-orthogonal support

• Non orthogonal implementation
• Removed CUB library from our source code and we will automatically download CUB library before compiling (newest version all the time!).
• Removed compute_20 and compute_21 since they are depricated and CUDA 9 will generate fatal error.
• Fixed a problem where the input reader will skip random_seed because there was a "info" print in an "else if" and would block anything after that.
• Fixed a bug in MoveSettings where two for loops where not initialized.
• Fixed a bug where in simulation of polar molecules, if any atom of the molecule has no VDW parameter (e.g. water), there is a chance that during insertion, atom of opposite charge will overlap and generate a very large energy value. Current mechanism of avoiding overlap is not working properly
• Fixed a bug where STDOUT was printing out garbage because input reader was passing empty vector.
• Fixed a problem where we were receiving seg fault at the end of the simulation (removed [] from destructor in Static.cpp).

HotFix

Avoid box shrinking bellow 2Rcut for NPT simulation + fixed some compiler warnings in newer versions