Commits · 63bb38a98c9360dbb3f7f4872a09f18490fbc8fd · tsoc / openmm

02 Jun, 2015 1 commit
- Cleaned up CUDA FFT code · 82069311
  Peter Eastman authored Jun 02, 2015
  
  82069311
30 May, 2015 1 commit
- Ported OpenCL FFT to CUDA · d3e91b15
  peastman authored May 29, 2015
  
  d3e91b15
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
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
10 Mar, 2015 1 commit
- Fixed a performance regression in multi-GPU on CUDA · 9d3a655b
  peastman authored Mar 10, 2015
  
  9d3a655b
06 Mar, 2015 2 commits
- 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
- Load balancing between GPUs forces the neighbor list to be rebuilt · 62f7c44c
  peastman authored Mar 06, 2015
  
  62f7c44c
13 Feb, 2015 1 commit
- Created plugin for CUDA runtime compiler · 7e29671a
  Peter Eastman authored Feb 13, 2015
  
  7e29671a
07 Jan, 2015 1 commit
- Fixed some methods that should have been marked const · 17c06daa
  Peter Eastman authored Jan 07, 2015
  
  17c06daa
05 Jan, 2015 1 commit
- Beginning of CUDA support for triclinic boxes · 61458a5f
  Peter Eastman authored Jan 05, 2015
  
  61458a5f
05 Nov, 2014 1 commit
- Further improvements to multi-GPU performance · 67a3b4c0
  peastman authored Nov 05, 2014
  
  67a3b4c0
21 Oct, 2014 2 commits
- Workaround for driver bugs that affect GTX 980 · 8e2fc4ea
  Peter Eastman authored Oct 21, 2014
  
  8e2fc4ea
- Optimizations to multi-GPU calculations · 390e0a6b
  root authored Oct 21, 2014
  
  390e0a6b
13 Oct, 2014 1 commit
- GBSAOBCForce allows the surface area energy to be changed · 34a604b8
  peastman authored Oct 13, 2014
  
  34a604b8
08 Oct, 2014 1 commit
- CUDA platform does PME on a separate stream · 832e7f04
  peastman authored Oct 08, 2014
  
  832e7f04
29 Aug, 2014 1 commit
- Minor optimization to CUDA version of CustomManyParticleForce · a306aa58
  peastman authored Aug 29, 2014
  
  a306aa58
22 Aug, 2014 1 commit
- Optimizations to CUDA version of CustomManyParticleForce · 150d943a
  peastman authored Aug 22, 2014
  
  150d943a
21 Aug, 2014 1 commit
- CUDA version of CustomManyParticleForce uses neighbor list · bddaf4e7
  peastman authored Aug 21, 2014
  
  bddaf4e7
20 Aug, 2014 1 commit
- Wrote code to build neighbor list for CUDA version of CustomManyParticleForce · e3b631f6
  peastman authored Aug 20, 2014
  
  e3b631f6
19 Aug, 2014 2 commits
- Implemented exclusions for CUDA version of CustomManyParticleForce · 786e1ee7
  peastman authored Aug 19, 2014
  
  786e1ee7
- Implemented type filters for CUDA version of CustomManyParticleForce · 65b56ee1
  peastman authored Aug 19, 2014
  
  65b56ee1
18 Aug, 2014 1 commit
- Beginning of CUDA implementation of CustomManyParticleForce · 890963b9
  peastman authored Aug 18, 2014
  
  890963b9
24 Apr, 2014 1 commit
- Created CUDA and OpenCL implementations of LocalCoordinatesSite · c88213f8
  peastman authored Apr 24, 2014
  
  c88213f8
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
- Same bug fix as the previous commit, but for CUDA · 1e857c31
  peastman authored Jan 10, 2014
  
  1e857c31
13 Dec, 2013 1 commit
- Added option for setting CUDA host compiler · b989fd9b
  peastman authored Dec 13, 2013
  
  b989fd9b
08 Nov, 2013 1 commit
- Added some missing virtual destructors · 9d207e8d
  peastman authored Nov 08, 2013
  
  9d207e8d
07 Aug, 2013 1 commit
- Created CUDA implementation of restricting CustomNonbondedForce to selected interaction groups · 09611266
  peastman authored Aug 07, 2013
  
  09611266
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

30 Jul, 2013 1 commit
- Can set OPENMM_CACHE_DIR to specify where PTX files should be cached · bf34e947
  peastman authored Jul 30, 2013
  
  bf34e947
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 2 commits
- OpenCL platform can use the CPU based PME · f2548616
  peastman authored Jun 27, 2013
  
  f2548616
- Created platform property for specifying whether to use the CPU based PME · d942ac2f
  peastman authored Jun 27, 2013
  
  d942ac2f
24 Jun, 2013 1 commit
- Moved CPU based PME to a separate plugin · 5e0b0e3a
  peastman authored Jun 24, 2013
  
  5e0b0e3a
08 Apr, 2013 1 commit
- Use peer to peer copying between GPUs when available · 41bcc20f
  Peter Eastman authored Apr 08, 2013
  
  41bcc20f