Commits · 7893f59a2763be55e8099a3aab30c2b693e48517 · tsoc / openmm

02 Sep, 2015 1 commit
- Created OpenCL version of CustomCentroidBondForce · 7893f59a
  peastman authored Sep 02, 2015
  
  7893f59a
10 Aug, 2015 1 commit
- Ensure neighbor list is large enough on first step · cb599d67
  Peter Eastman authored Aug 10, 2015
  
  cb599d67
29 Jul, 2015 1 commit
- Very minor optimizations · ef627240
  peastman authored Jul 29, 2015
  
  ef627240
17 Jul, 2015 1 commit
- OpenCL CustomIntegrator supports if and while blocks · 9ddded35
  peastman authored Jul 17, 2015
  
  9ddded35
09 Jul, 2015 1 commit
- OpenCL allows different force groups to have different cutoffs · f065df01
  Peter Eastman authored Jul 09, 2015
  
  f065df01
15 May, 2015 1 commit
- Switched over to new Windows pthreads library · 6cf75568
  peastman authored May 15, 2015
  
  6cf75568
22 Apr, 2015 1 commit
- Fixed bug in CustomIntegrator · 4cf66cab
  Peter Eastman authored Apr 22, 2015
  
  4cf66cab
15 Apr, 2015 1 commit
- OpenCL FFT stores only non-redundant elements · ad62d81e
  Peter Eastman authored Apr 15, 2015
  
  ad62d81e
08 Apr, 2015 1 commit
- Implemented an optimized real-to-complex FFT · 9c7898e0
  peastman authored Apr 08, 2015
  
  9c7898e0
03 Apr, 2015 1 commit
- OpenCL FFT supports real-to-complex transforms · 913c6dc3
  Peter Eastman authored Apr 03, 2015
  
  913c6dc3
17 Mar, 2015 1 commit
- CUDA and OpenCL implementations of CMAPTorsionForce::updateParametersInContext() · 6dc2f9a8
  peastman authored Mar 16, 2015
  
  6dc2f9a8
16 Mar, 2015 1 commit
- Added updateParametersInContext() to CMAPTorsionForce, and implemented in reference platform · 474bbf7a
  peastman authored Mar 16, 2015
  
  474bbf7a
06 Mar, 2015 1 commit
- For multi-GPU, exceeding the size of the neighbor list causes the forces to be marked as invalid · bd666c27
  peastman authored Mar 06, 2015
  
  bd666c27
07 Jan, 2015 1 commit
- Began implementing OpenCL support for triclinic boxes · 195bcecf
  Peter Eastman authored Jan 07, 2015
  
  195bcecf
13 Oct, 2014 1 commit
- GBSAOBCForce allows the surface area energy to be changed · 34a604b8
  peastman authored Oct 13, 2014
  
  34a604b8
09 Oct, 2014 1 commit
- Further optimizations to NonbondedForce · 3d1b2186
  peastman authored Oct 09, 2014
  
  3d1b2186
08 Oct, 2014 1 commit
- OpenCL platform does PME on a separate stream · 94aa8c3f
  peastman authored Oct 08, 2014
  
  94aa8c3f
03 Sep, 2014 1 commit
- Created OpenCL version of CustomManyParticleForce (not yet fully debugged) · 6f7dee30
  peastman authored Sep 03, 2014
  
  6f7dee30
02 Jun, 2014 1 commit
- Fixed bug when multiple virtual sites depend on the same particles (see bug 2019) · 496f469a
  peastman authored Jun 02, 2014
  
  496f469a
24 Apr, 2014 1 commit
- Created CUDA and OpenCL implementations of LocalCoordinatesSite · c88213f8
  peastman authored Apr 24, 2014
  
  c88213f8
09 Apr, 2014 1 commit
- Disable OpenCL platform on known-buggy macs · 9edaf653
  Robert McGibbon authored Apr 09, 2014
  
  9edaf653
27 Jan, 2014 1 commit

Tabulated function parameters are hardcoded in the kernel instead of being... · 0a1a011d

peastman authored Jan 27, 2014

Tabulated function parameters are hardcoded in the kernel instead of being stored in an array.  This makes the code simpler and may help performance slightly.

0a1a011d

25 Jan, 2014 1 commit
- Created Discrete2DFunction and Discrete3DFunction · a773952e
  peastman authored Jan 24, 2014
  
  a773952e
24 Jan, 2014 2 commits
- Created Discrete1DFunction · 56e36449
  peastman authored Jan 24, 2014
  
  56e36449
- Converted CUDA and OpenCL platforms to process tabulated functions with the TabulatedFunction API · 81f26683
  peastman authored Jan 23, 2014
  
  81f26683
11 Jan, 2014 1 commit
- Fixed a bug in OpenCL implementation of CustomGBForce · 44e65cb1
  peastman authored Jan 10, 2014
  
  44e65cb1
26 Nov, 2013 1 commit
- Bug fix to selecting workgroup size for findInteractingBlocks() · a3fd6c23
  peastman authored Nov 26, 2013
  
  a3fd6c23
08 Nov, 2013 1 commit
- Added some missing virtual destructors · 9d207e8d
  peastman authored Nov 08, 2013
  
  9d207e8d
23 Aug, 2013 1 commit
- bugfixes to previous commit · 27b802cf
  Robert McGibbon authored Aug 22, 2013
  
  27b802cf
08 Aug, 2013 1 commit
- Began creating OpenCL implementation of CustomNonbondedForce interaction groups · 54c0ca3f
  peastman authored Aug 08, 2013
  
  54c0ca3f
01 Aug, 2013 1 commit

Converted the routine for identifying molecules from a recursion to a loop. ... · b2d5520d

peastman authored Aug 01, 2013

Converted the routine for identifying molecules from a recursion to a loop.  This avoids stack overflows with very long linear molecules.

b2d5520d

08 Jul, 2013 1 commit

Platform specific header files get installed. This allows plugins to be built... · 300758a5

peastman authored Jul 08, 2013

Platform specific header files get installed.  This allows plugins to be built with just an OpenMM installation, not a full source tree.

300758a5

27 Jun, 2013 1 commit
- OpenCL platform can use the CPU based PME · f2548616
  peastman authored Jun 27, 2013
  
  f2548616
04 Apr, 2013 1 commit
- Ensure that CUDA is chosen over OpenCL when both are available · 4e50d721
  Peter Eastman authored Apr 04, 2013
  
  4e50d721
28 Mar, 2013 1 commit
- Added platform properties to report the CUDA/OpenCL device name and OpenCL platform name · 0f83704f
  Peter Eastman authored Mar 28, 2013
  
  0f83704f
22 Mar, 2013 1 commit
- Replaced lots of tabs with spaces · b723e4af
  Peter Eastman authored Mar 22, 2013
  
  b723e4af
12 Dec, 2012 1 commit
- Fixed compilation errors on Windows · b256555e
  Peter Eastman authored Dec 12, 2012
  
  b256555e
16 Oct, 2012 1 commit
- Beginnings of mixed/double precision support in OpenCL · 8d6a2a01
  Peter Eastman authored Oct 16, 2012
  
  8d6a2a01
11 May, 2012 1 commit
- Fixed bugs in checkpointing · fcdba25c
  Peter Eastman authored May 11, 2012
  
  fcdba25c
15 Nov, 2011 1 commit
- By default, use OpenCL instead of CUDA · b9c09b79
  Peter Eastman authored Nov 15, 2011
  
  b9c09b79