Commits · 5fb78b59dda3c887f2bc845b8d9f32972b477b1a · tsoc / openmm

23 Oct, 2012 1 commit
- Updated to newer version of OpenCL C++ bindings · 5fb78b59
  Peter Eastman authored Oct 23, 2012
  
  5fb78b59
20 Oct, 2012 1 commit
- Limit thread block size to avoid exceeding available shared memory · 6d82a920
  Peter Eastman authored Oct 20, 2012
  
  6d82a920
19 Oct, 2012 1 commit
- The test for identical exceptions between nonbonded forces is order insensitive · ff7f1035
  Peter Eastman authored Oct 19, 2012
  
  ff7f1035
18 Oct, 2012 4 commits
- Assorted cleanup and bug fixes · 1de2e2a0
  Peter Eastman authored Oct 18, 2012
  
  1de2e2a0
- CustomNonbondedForce.cpp, CudaNonbondedUtilities.cpp and CudaKernels.cpp are... · 93497df6
  Lee-Ping Wang authored Oct 18, 2012
```
CustomNonbondedForce.cpp, CudaNonbondedUtilities.cpp and CudaKernels.cpp are now fully reverted to 3478.
```
  93497df6
- Rolled back "list sorting" change but kept commented debug printout · a0a0cac3
  Lee-Ping Wang authored Oct 18, 2012
  
  a0a0cac3
- LPW attempted bugfixes: (1) Changed unit multipliers in... · 68593ba6
  Lee-Ping Wang authored Oct 18, 2012
```
LPW attempted bugfixes: (1) Changed unit multipliers in AmoebaVdwForce::getParticleParameters (2) Sorted vectors in exclusionLists so it agrees with atomExclusions
```
  68593ba6
17 Oct, 2012 2 commits
- Finished implementing double precision for OpenCL · 59f8cf87
  Peter Eastman authored Oct 17, 2012
  
  59f8cf87
- Continuing to implement double precision in OpenCL · c8dac206
  Peter Eastman authored Oct 17, 2012
  
  c8dac206
16 Oct, 2012 2 commits
- Bug fix · 34938e2c
  Peter Eastman authored Oct 16, 2012
  
  34938e2c
- Beginnings of mixed/double precision support in OpenCL · 8d6a2a01
  Peter Eastman authored Oct 16, 2012
  
  8d6a2a01
12 Oct, 2012 2 commits
- Fixed compilation errors and warnings on Windows · 9fd73edf
  Peter Eastman authored Oct 12, 2012
  
  9fd73edf
- Flush the queue after each time step on Windows to reduce UI lag. · 65c389ff
  Peter Eastman authored Oct 12, 2012
  
  65c389ff
10 Oct, 2012 1 commit

OpenCLArray is no longer templatized and doesn't provide a host buffer. This... · 1107aa83

Peter Eastman authored Oct 10, 2012

OpenCLArray is no longer templatized and doesn't provide a host buffer.  This is in preparation for adding mixed/double precision support.

1107aa83

05 Oct, 2012 2 commits
- Fixed performance bug in mixed precision · 5980100d
  Peter Eastman authored Oct 05, 2012
  
  5980100d
- Eliminated mixed precision accumulation for nonbonded forces, which had negligible benefit · b0161939
  Peter Eastman authored Oct 05, 2012
  
  b0161939
04 Oct, 2012 1 commit
- Finished implementing mixed precision model · 8cda1b79
  Peter Eastman authored Oct 04, 2012
  
  8cda1b79
02 Oct, 2012 1 commit
- Began implementing new mixed precision model that does integration in double... · 9ad85ebd
  Peter Eastman authored Oct 02, 2012
```
Began implementing new mixed precision model that does integration in double precision and force evaluation in single precision
```
  9ad85ebd
28 Sep, 2012 2 commits
- Renamed cuda2 platform to cuda · 58b094ce
  Peter Eastman authored Sep 28, 2012
  
  58b094ce
- Renamed old CUDA platform to cuda-old · 5a25b4ec
  Peter Eastman authored Sep 28, 2012
  
  5a25b4ec
27 Sep, 2012 1 commit
- Fixed test cases that used a Context without setting positions · 01269257
  Peter Eastman authored Sep 27, 2012
  
  01269257
25 Sep, 2012 1 commit
- Fixed bug when using multiple Contexts at once · 3d98a96f
  Peter Eastman authored Sep 25, 2012
  
  3d98a96f
21 Sep, 2012 1 commit
- Can use OPENMM_CUDA_COMPILER environment variable to set the path to the CUDA compiler · 20e01186
  Peter Eastman authored Sep 21, 2012
  
  20e01186
11 Sep, 2012 1 commit
- Minor cleanup to PME · 235a88e5
  Peter Eastman authored Sep 11, 2012
  
  235a88e5
10 Sep, 2012 1 commit
- Committing FFTWs to work with complex conjugate arrays. Decoupled energy and... · b27a0bd6
  Yutong Zhao authored Sep 10, 2012
```
Committing FFTWs to work with complex conjugate arrays. Decoupled energy and force calculation into two separate kernels.
```
  b27a0bd6
07 Sep, 2012 1 commit
- fixed sigmaEpsilon padding errors. Gets rid of computeNonbonded cuda-memcheck errors · 9fffe6da
  Yutong Zhao authored Sep 07, 2012
  
  9fffe6da
29 Aug, 2012 1 commit

Continuing to convert AmoebaMultipoleForce: implemented routines to get system... · 5c071a07

Peter Eastman authored Aug 29, 2012

Continuing to convert AmoebaMultipoleForce: implemented routines to get system multipole moments and potential at grid points.  Some bug fixes.

5c071a07

28 Aug, 2012 1 commit
- Fixed errors on Intel OpenCL · b8e15090
  Peter Eastman authored Aug 28, 2012
  
  b8e15090
21 Aug, 2012 1 commit
- Fixed compilation error on compute level 2.1 GPUs · 3af4da86
  Peter Eastman authored Aug 21, 2012
  
  3af4da86
20 Jul, 2012 1 commit
- Continuing converting AMOEBA to the new CUDA platform: starting on AmoebaMultipoleForce · d0426ba9
  Peter Eastman authored Jul 20, 2012
  
  d0426ba9
12 Jul, 2012 1 commit
- Fixed a memory leak · a90c90e2
  Peter Eastman authored Jul 12, 2012
  
  a90c90e2
10 Jul, 2012 2 commits
- Implemented mixed precision mode · 18fb6efc
  Peter Eastman authored Jul 10, 2012
  
  18fb6efc
- Set default cache configuration for the context · 76716f1c
  Peter Eastman authored Jul 10, 2012
  
  76716f1c
06 Jul, 2012 1 commit
- Optimizations to new CUDA platform · 2bb875b0
  Peter Eastman authored Jul 06, 2012
  
  2bb875b0
03 Jul, 2012 2 commits
- New CUDA platform works on Windows · ca7fd533
  Peter Eastman authored Jul 03, 2012
  
  ca7fd533
- Fixed bugs when using double precision · cf112a25
  Peter Eastman authored Jul 03, 2012
  
  cf112a25
02 Jul, 2012 1 commit
- Changed number of nonbonded thread blocks to improve performance · dac874af
  Peter Eastman authored Jul 02, 2012
  
  dac874af
28 Jun, 2012 3 commits
- Use the __popc intrinsic for counting bits · c77af6e9
  Peter Eastman authored Jun 28, 2012
  
  c77af6e9
- Optimization: use mapped memory to communicate when CCMA is converged · f0c2e89c
  Peter Eastman authored Jun 28, 2012
  
  f0c2e89c
- Merged lots of kernels into fewer files to reduce compilation time · 1afb079f
  Peter Eastman authored Jun 28, 2012
  
  1afb079f