Commits · 6f3d9fd54234d27c3377ae4e43deffc84ea55e48 · tsoc / openmm

04 Dec, 2013 2 commits
- Enforce memory alignment to improve performance of vector operations. Also... · b1be68d8
  peastman authored Dec 04, 2013
```
Enforce memory alignment to improve performance of vector operations.  Also fixed bugs in an earlier optimization.
```
  b1be68d8
- Optimizations to CPU nonbonded forces: better load balancing between threads,... · be04ec57
  peastman authored Dec 04, 2013
```
Optimizations to CPU nonbonded forces: better load balancing between threads, use linear splines instead of cubic
```
  be04ec57
03 Dec, 2013 1 commit
- Optimization to building neighbor list on CPU platform · 56cf0fde
  peastman authored Dec 03, 2013
  
  56cf0fde
25 Nov, 2013 1 commit
- Fixed bugs in computing PME energy with CPU platform · 0ee8c785
  peastman authored Nov 25, 2013
  
  0ee8c785
13 Nov, 2013 1 commit
- Fixed errors on OS X 10.9: link against libc++ instead of libstdc++ · c1f64b3f
  peastman authored Nov 13, 2013
  
  c1f64b3f
07 Nov, 2013 2 commits
- Tabulated log(x) to improve GB performance · f39a5545
  peastman authored Nov 07, 2013
  
  f39a5545
- Fixed compilation errors on gcc · 7714f7ff
  peastman authored Nov 07, 2013
  
  7714f7ff
06 Nov, 2013 1 commit
- Optimizations to CPU version of GBSA · 563914a7
  peastman authored Nov 06, 2013
  
  563914a7
05 Nov, 2013 1 commit
- Minor optimizations to CPU version of GBSAOBCForce · 775866b6
  peastman authored Nov 04, 2013
  
  775866b6
04 Nov, 2013 2 commits
- Created optimized CPU version of GBSAOBCForce · c97ca97b
  peastman authored Nov 04, 2013
  
  c97ca97b
- Bug fixes · 7b608514
  peastman authored Nov 04, 2013
  
  7b608514
01 Nov, 2013 1 commit
- Centralized the forces, positions, and thread pool so they can be shared between kernels · be61ee5b
  peastman authored Nov 01, 2013
  
  be61ee5b
30 Oct, 2013 3 commits
- Also use the previous optimization with reaction field · 79f064af
  peastman authored Oct 30, 2013
  
  79f064af
- Minor optimizations to CPU platform: avoid applying periodic boundary conditions unnecessarily · 62c64e3c
  peastman authored Oct 30, 2013
  
  62c64e3c
- Further optimizations to CPU nonbonded force · a69227ca
  peastman authored Oct 30, 2013
  
  a69227ca
29 Oct, 2013 1 commit
- Created ThreadPool to simplify parallel code · f3fe550a
  peastman authored Oct 29, 2013
  
  f3fe550a
24 Oct, 2013 3 commits
- Minor optimizations to CPU platform · a5bcc4be
  peastman authored Oct 24, 2013
  
  a5bcc4be
- Minor change to the neighbor list padding distance · 062a2db0
  peastman authored Oct 23, 2013
  
  062a2db0
- Do a more complex test to decide when to rebuild the neighbor list · 8d9950da
  peastman authored Oct 23, 2013
  
  8d9950da
23 Oct, 2013 3 commits
- Minor optimizations to neighbor list construction · 6ad86974
  peastman authored Oct 23, 2013
  
  6ad86974
- Parallelized the processing of exclusions for Ewald · d6b2eb75
  peastman authored Oct 23, 2013
  
  d6b2eb75
- Optimized algorithm for building neighbor list. · f301042f
  peastman authored Oct 23, 2013
  
  f301042f
22 Oct, 2013 1 commit
- Restructured nonbonded calculation to allow more efficient vectorization · 56902b45
  peastman authored Oct 22, 2013
  
  56902b45
21 Oct, 2013 1 commit
- Created vector classes to simplify SSE code · 6eacad45
  peastman authored Oct 21, 2013
  
  6eacad45
17 Oct, 2013 1 commit
- Fixed segfaults on some computers · 8f168506
  peastman authored Oct 17, 2013
  
  8f168506
15 Oct, 2013 4 commits
- Prevent segfaults when a simulation blows up · d389e504
  peastman authored Oct 15, 2013
  
  d389e504
- Bug fix · 2943d628
  peastman authored Oct 15, 2013
  
  2943d628
- Bug fix, plus a minor optimization to reference platform · dd175b23
  peastman authored Oct 15, 2013
  
  dd175b23
- Assorted optimizations · 79011828
  peastman authored Oct 14, 2013
  
  79011828
14 Oct, 2013 2 commits
- Tabulate the Ewald scale factor · f1533707
  peastman authored Oct 14, 2013
  
  f1533707
- Fixed a race condition · 9491eb72
  peastman authored Oct 14, 2013
  
  9491eb72
12 Oct, 2013 1 commit
- Do not rebuild the neighbor list every time step · 18f1cd3e
  peastman authored Oct 11, 2013
  
  18f1cd3e
11 Oct, 2013 4 commits
- Direct space nonbonded is multithreaded · b278bb6d
  peastman authored Oct 11, 2013
  
  b278bb6d
- Use optimized PME implementation when available · 3c435fb8
  peastman authored Oct 11, 2013
  
  3c435fb8
- PME works correctly in CPU platform · e22b8f53
  peastman authored Oct 11, 2013
  
  e22b8f53
- Continuing to optimize CPU nonbonded routines · 6b10b909
  peastman authored Oct 11, 2013
  
  6b10b909
10 Oct, 2013 4 commits
- Began creating optimized implementation of direct space nonbonded calculation · 9bc194b9
  peastman authored Oct 10, 2013
  
  9bc194b9
- Separated hardware query functions into a separate file. CPU platform checks... · 50d4f9d0
  peastman authored Oct 10, 2013
```
Separated hardware query functions into a separate file.  CPU platform checks for the correct hardware.
```
  50d4f9d0
- Neighborlist construction is multithreaded. · f4a1847a
  peastman authored Oct 10, 2013
  
  f4a1847a
- More optimizations to neighborlist construction, including SSE intrinsics · 3b3fd578
  peastman authored Oct 09, 2013
  
  3b3fd578