Commits · 3031c4da50187cbf97e4b62dd9add7dfe5fa468d · tsoc / openmm

09 Sep, 2014 4 commits
- Fixed bug on Intel GPUs · 3031c4da
  peastman authored Sep 09, 2014
  
  3031c4da
- Improved performance of building neighbor list on AMD GPUs · ae466134
  peastman authored Sep 09, 2014
  
  ae466134
- Bug fixes · f9fbe12a
  peastman authored Sep 09, 2014
  
  f9fbe12a
- Ported the new version of findBlocksWithInteractions() to OpenCL (not yet debugged) · 70bd61a6
  peastman authored Sep 09, 2014
  
  70bd61a6
04 Sep, 2014 2 commits
- Fixed errors compiling kernels in double precision mode · 6a38039f
  peastman authored Sep 04, 2014
  
  6a38039f
- Bug fixes to OpenCL version of CustomManyParticleForce · 9f77efc2
  Peter Eastman authored Sep 04, 2014
  
  9f77efc2
03 Sep, 2014 2 commits
- Bug fix · 9cd18aeb
  peastman authored Sep 03, 2014
  
  9cd18aeb
- Created OpenCL version of CustomManyParticleForce (not yet fully debugged) · 6f7dee30
  peastman authored Sep 03, 2014
  
  6f7dee30
18 Aug, 2014 1 commit
- avoid syntax error if tabulated function min is negative · e30242b6
  Mike Garrahan authored Aug 18, 2014
  
  e30242b6
12 Aug, 2014 2 commits
- Fixed uninitialized memory error when using multiple GPUs · 2279888f
  peastman authored Aug 12, 2014
  
  2279888f
- Avoid integer overflow when dealing with very large systems · edd91641
  peastman authored Aug 12, 2014
  
  edd91641
30 Jul, 2014 1 commit
- Fix crash with empty interaction groups · 92d5b696
  peastman authored Jul 30, 2014
  
  92d5b696
23 Jul, 2014 1 commit
- Fixed error in OpenCL MonteCarloBarostat in double precision · debe1662
  peastman authored Jul 23, 2014
  
  debe1662
15 Jul, 2014 1 commit
- Fixed bug when running on a CPU device · c9c5997b
  peastman authored Jul 15, 2014
  
  c9c5997b
27 Jun, 2014 1 commit
- Fixed a rare memory corruption error when roundoff causes tile indices to be wrong · ece75910
  peastman authored Jun 27, 2014
  
  ece75910
03 Jun, 2014 1 commit
- Fixed a crash with virtual sites in double precision mode · 3bf0edb4
  peastman authored Jun 03, 2014
  
  3bf0edb4
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
14 May, 2014 1 commit
- Preliminary support for ARM/Android. · 454caac9
  peastman authored May 14, 2014
  
  454caac9
06 May, 2014 1 commit
- Fixed bug using multiple devices with PME · f1991b3c
  peastman authored May 06, 2014
  
  f1991b3c
24 Apr, 2014 1 commit
- Created CUDA and OpenCL implementations of LocalCoordinatesSite · c88213f8
  peastman authored Apr 24, 2014
  
  c88213f8
10 Apr, 2014 1 commit
- Fixed bug in OpenCL implementation of GBSA on CPU devices · cd0ee42e
  peastman authored Apr 09, 2014
  
  cd0ee42e
09 Apr, 2014 1 commit
- Disable OpenCL platform on known-buggy macs · 9edaf653
  Robert McGibbon authored Apr 09, 2014
  
  9edaf653
06 Mar, 2014 1 commit
- Removed unnecessary synchronization that hurt performance on Nvidia · 73643858
  Peter Eastman authored Mar 05, 2014
  
  73643858
31 Jan, 2014 1 commit
- Renamed methods for dealing with TabulatedFunctions to avoid problems in... · 75a04e7f
  peastman authored Jan 31, 2014
```
Renamed methods for dealing with TabulatedFunctions to avoid problems in wrapper APIs.  Also fixed a few bugs.
```
  75a04e7f
30 Jan, 2014 1 commit
- Created OpenCL and CUDA implementations of Continuous3DFunction · 3279c73a
  peastman authored Jan 30, 2014
  
  3279c73a
29 Jan, 2014 2 commits
- Fixed an error on platforms where the maximum workgroup size can vary between kernels · f7f70136
  peastman authored Jan 29, 2014
  
  f7f70136
- Created OpenCL and CUDA implementations of Continuous2DFunction · 9434c6e8
  peastman authored Jan 29, 2014
  
  9434c6e8
27 Jan, 2014 2 commits
- CPU, CUDA, and OpenCL platforms can now be built statically · 8285d658
  Yutong Zhao authored Jan 27, 2014
  
  8285d658
- 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
23 Jan, 2014 1 commit
- Prevent trying to use the CPU OpenCL device on OS X · ee3fcb01
  peastman authored Jan 23, 2014
  
  ee3fcb01
11 Jan, 2014 1 commit
- Fixed a bug in OpenCL implementation of CustomGBForce · 44e65cb1
  peastman authored Jan 10, 2014
  
  44e65cb1
09 Jan, 2014 1 commit
- Fixed a bug that could potentially cause LJ energy to be nan · 5c983283
  peastman authored Jan 09, 2014
  
  5c983283
07 Jan, 2014 1 commit
- Very minor optimization · 06db7ac4
  peastman authored Jan 07, 2014
  
  06db7ac4
06 Jan, 2014 1 commit
- Improved logic for selecting a platform to better deal with exceptions in context initialization · d91e3f19
  peastman authored Jan 06, 2014
  
  d91e3f19
26 Nov, 2013 1 commit
- Bug fix to selecting workgroup size for findInteractingBlocks() · a3fd6c23
  peastman authored Nov 26, 2013
  
  a3fd6c23
23 Nov, 2013 1 commit
- Workarounds for devices whose maximum workgroup size varies between kernels · 35b9d787
  peastman authored Nov 22, 2013
  
  35b9d787
01 Oct, 2013 1 commit
- Ignore constraints that connect two massless particles (feature request 1915) · 803faa81
  peastman authored Oct 01, 2013
  
  803faa81