1efc388 | Hammad Mazhar | 25 August 2013, 22:54:21 UTC | Jacobian and rhs code will be deprecated in favor of a constraint based approach. Rhs and Jacobians along with shur matrix vector multiplies are done on a per constraint type basis, this helps to keep the code clean and modular. Fluid-Fluid and Rigid-Fluid code is not enabled yet Bilaterals need to be tested still. | 25 August 2013, 22:54:21 UTC |
ae8c79b | Hammad Mazhar | 25 August 2013, 20:03:05 UTC | Remvoed combined jacobian and body id data structures, separated into specific ones for contacts and bilaterals. This removed extra copy operations and streamlines code slightly. | 25 August 2013, 20:03:05 UTC |
6d9e8c9 | Hammad Mazhar | 25 August 2013, 19:47:23 UTC | Reorganized data structures for host and device data. Currently working on host side, cleaning up memory accesses and removing redundant data. For host simulations memory should be drastically reduced | 25 August 2013, 19:47:23 UTC |
f8b1d10 | Hammad Mazhar | 22 August 2013, 18:59:10 UTC | Stabilization for APGD can be enabled or disabled at runtime | 22 August 2013, 18:59:10 UTC |
7238db5 | Hammad Mazhar | 22 August 2013, 03:59:26 UTC | parallel region was not correctly used, removed | 22 August 2013, 03:59:26 UTC |
a823bc8 | Hammad Mazhar | 22 August 2013, 01:38:34 UTC | Added cone asset and geometry type Tweaked velocity stabilization Updated Opengl to display cone geometry Added a new function to add cone to a collision model, not implemented for Bullet collision model | 22 August 2013, 01:38:34 UTC |
00a4450 | Hammad Mazhar | 21 August 2013, 04:47:00 UTC | Modified triangle mesh for gpu velocity stabilization active by default opengl code handles triangle meshes better | 21 August 2013, 04:47:00 UTC |
b7bafe1 | Justin Madsen | 19 August 2013, 16:05:09 UTC | allows a ChFunction to be passed into the PID controller, which manages a smooth transition between the current and desired controller inputs over a specified time period. | 19 August 2013, 16:05:09 UTC |
b9ea24e | Hammad Mazhar | 09 August 2013, 21:44:39 UTC | Moved setBody to the CHcCollisionModel class A chBody can now set the body in the constructor when it is given a collision model as an argument There is no need for ChBodyGPU, class is removed Removed all dependency on ChBodyGPU class Added velocity based stabilization for bilateral constraints, need to clean this up still Added compliance as a per object parameter | 09 August 2013, 21:44:39 UTC |
cf8193e | Andrea Favali | 07 August 2013, 20:18:03 UTC | added GetStrain/GetStress for hexahedrons | 07 August 2013, 20:18:03 UTC |
9c3a69d | Hammad Mazhar | 03 August 2013, 23:32:02 UTC | removed old include for gpu jacobi solver | 03 August 2013, 23:32:02 UTC |
c02a063 | Hammad Mazhar | 03 August 2013, 22:48:55 UTC | Added new base GPU class that contains data that most GPU related code requires Added a new class for RHS computation Moved Jacobi solver to the main solver class Adding functionality for stabilization related solver step | 03 August 2013, 22:48:55 UTC |
d47f860 | Andrea Favali | 03 August 2013, 01:04:57 UTC | | 03 August 2013, 01:04:57 UTC |
3ba0694 | Hammad Mazhar | 30 July 2013, 21:01:59 UTC | Updated ChBody with an extra constructor that takes a collision model as an argument Added a unique identifier to Chbody for indexing purposes ChBodyGPU is now a wrapper around ChBody, there is no custom code in ChBodyGPU, it calls the new ChBody constructor with a GPU collision model | 30 July 2013, 21:01:59 UTC |
c873911 | Hammad Mazhar | 30 July 2013, 19:14:49 UTC | Updated data manager copy functions to be more generic cleaning up the ChSystemGPU code | 30 July 2013, 19:14:49 UTC |
dc0ae5d | Andrea Favali | 30 July 2013, 03:13:59 UTC | | 30 July 2013, 03:13:59 UTC |
cae7acd | Andrea Favali | 30 July 2013, 03:03:11 UTC | Added the classes GaussPoint and GaussIntegrationRule to to manage the volume integration process | 30 July 2013, 03:03:11 UTC |
798e648 | Andrea Favali | 30 July 2013, 03:00:32 UTC | | 30 July 2013, 03:00:32 UTC |
81a2ed3 | Justin Madsen | 29 July 2013, 15:41:33 UTC | create bodies where the COG is offset from the mesh origin | 29 July 2013, 15:41:33 UTC |
02e30dd | Justin Madsen | 25 July 2013, 17:02:08 UTC | Justin - adding functions to easily insert a body with a mesh whose origin is offset from the COG of the body | 25 July 2013, 17:02:08 UTC |
e737ba7 | Hammad Mazhar | 24 July 2013, 19:00:36 UTC | Some extra qualifications removed from header files | 24 July 2013, 19:00:36 UTC |
9699fe3 | Andrew Seidl | 24 July 2013, 18:46:42 UTC | Fixing backslashes in #includes that Visual Studio inserted | 24 July 2013, 18:46:42 UTC |
430db26 | Andrea Favali | 22 July 2013, 22:20:01 UTC | | 22 July 2013, 22:20:01 UTC |
c86bc49 | Andrea Favali | 22 July 2013, 22:18:31 UTC | Added QUARDATIC HEXAHEDRON Updated demo_FEMbasic.cpp | 22 July 2013, 22:18:31 UTC |
080b296 | Hammad Mazhar | 19 July 2013, 11:35:18 UTC | cleaning up GPU code removed optimized APGD code temporarily (might be placed in it's own solver class) removed dependency on gpu data structures, all classes should now get the full data and determine which data to pick. | 19 July 2013, 11:35:18 UTC |
1326bc3 | Andrea Favali | 18 July 2013, 21:05:17 UTC | Added linear hexaedral element class Updated fem/demo_FEMbasic.cpp | 18 July 2013, 21:05:17 UTC |
17b2bdc | Hammad Mazhar | 18 July 2013, 14:47:24 UTC | refactor: change CHGPUDataManager class to ChGPUDataManager | 18 July 2013, 14:47:24 UTC |
d6229f3 | Hammad Mazhar | 18 July 2013, 09:57:59 UTC | Fixing compilation issues, slashes were in the wrong direction | 18 July 2013, 09:57:59 UTC |
16977bd | Andrea Favali | 16 July 2013, 16:28:13 UTC | | 16 July 2013, 16:28:13 UTC |
cac828e | Andrea Favali | 16 July 2013, 16:24:57 UTC | Added quadratic tetrahedron | 16 July 2013, 16:24:57 UTC |
f3f6b71 | Alessandro Tasora | 10 July 2013, 15:16:07 UTC | Setup changed to SetupInitial in ChElementBase etc. Also, SetupInitial() added to ChMesh too, for collective initialization. | 10 July 2013, 15:16:07 UTC |
65cee33 | Alessandro Tasora | 10 July 2013, 14:49:56 UTC | Bug in ChLcpConstraintTwoGeneric - fixed. The FEM test with constrained tetahedron is working. | 10 July 2013, 14:49:56 UTC |
7cb3ec4 | Alessandro Tasora | 10 July 2013, 09:24:37 UTC | Bug in ChNodeBody, the 'attached position' of a point was always 0,0,0 after Initialize. Now fixed to get the actual pos of the node. | 10 July 2013, 09:24:37 UTC |
245dd98 | Radu Serban | 09 July 2013, 16:53:07 UTC | Modified the vector3dfCH() constructor and the ChIrrTools::alignIrrlichtNodeToChronoCsys() function to take const references as parameters (as they are not modified). This allows passing rvalues (e.g. temporaries) which would be prohibited if these functions accepted non-const references (error with gcc; Visual C++ binds rvalues to non-const references as a "feature"). | 09 July 2013, 16:53:07 UTC |
48ed43c | Alessandro Tasora | 09 July 2013, 16:49:34 UTC | | 09 July 2013, 16:49:34 UTC |
8bad975 | Hammad Mazhar | 09 July 2013, 08:20:27 UTC | Modified cmake to reflect changed license and readme file names GPU unit now compiles with gcc 4.7.3 | 09 July 2013, 08:20:27 UTC |
1566796 | Hammad Mazhar | 09 July 2013, 08:03:37 UTC | Updated license to BSD3 renamed license file and readme files Fixed compilation errors with fem code modified opengl code | 09 July 2013, 08:03:37 UTC |
50a2140 | Andrea Favali | 05 July 2013, 13:42:49 UTC | | 05 July 2013, 13:42:49 UTC |
928c86b | Andrea Favali | 04 July 2013, 20:58:49 UTC | | 04 July 2013, 20:58:49 UTC |
6d19a4f | Andrea Favali | 04 July 2013, 20:56:51 UTC | | 04 July 2013, 20:56:51 UTC |
972a4df | Andrea Favali | 04 July 2013, 20:56:22 UTC | | 04 July 2013, 20:56:22 UTC |
d21f7bb | Andrea Favali | 04 July 2013, 20:46:51 UTC | added CHelement3D.h, CHelementTetra_4.h/.cpp updated ChContinuumElastic class | 04 July 2013, 20:46:51 UTC |
395ab2e | Andrea Favali | 04 July 2013, 20:45:39 UTC | | 04 July 2013, 20:45:39 UTC |
50d70f2 | Andrea Favali | 04 July 2013, 20:39:44 UTC | | 04 July 2013, 20:39:44 UTC |
9c82b85 | Alessandro Tasora | 03 July 2013, 21:24:39 UTC | | 03 July 2013, 21:24:39 UTC |
49b8a27 | Alessandro Tasora | 02 July 2013, 21:24:59 UTC | | 02 July 2013, 21:24:59 UTC |
13fafbd | Alessandro Tasora | 02 July 2013, 21:24:19 UTC | Obsolete code removed from the FEM directory. The CHfem.h file has been splitted in multiple files. | 02 July 2013, 21:24:19 UTC |
6e1dc9e | Andrea Favali | 02 July 2013, 12:27:54 UTC | Updated Det() Added MatrInverse(matrA) (max 4x4 matrices) | 02 July 2013, 12:27:54 UTC |
be34dfe | Alessandro Tasora | 01 July 2013, 14:03:00 UTC | | 01 July 2013, 14:03:00 UTC |
f434efa | Justin Madsen | 30 June 2013, 21:15:43 UTC | soil bin with: controls, wheel and soil info panels | 30 June 2013, 21:15:43 UTC |
e25d1b7 | Justin Madsen | 26 June 2013, 14:41:00 UTC | fixed some bin controller issues | 26 June 2013, 14:41:00 UTC |
01d7911 | Justin Madsen | 26 June 2013, 14:00:33 UTC | soilbin now has working controls, output display for wheel and soil infos | 26 June 2013, 14:00:33 UTC |
a7a1a42 | Justin Madsen | 25 June 2013, 15:53:01 UTC | working version of soilbin, with controls | 25 June 2013, 15:53:01 UTC |
6c54b6c | Hammad Mazhar | 20 June 2013, 07:33:43 UTC | cleaned up APGD code, improved performance | 20 June 2013, 07:33:43 UTC |
06758ea | Justin Madsen | 19 June 2013, 04:13:59 UTC | fixed interface app bug | 19 June 2013, 04:13:59 UTC |
3ed3fcb | Hammad Mazhar | 18 June 2013, 18:47:34 UTC | Removed minimum residual computation from main loop | 18 June 2013, 18:47:34 UTC |
be4236f | Hammad Mazhar | 18 June 2013, 18:19:30 UTC | fixed compilation issue | 18 June 2013, 18:19:30 UTC |
3d5204d | Hammad Mazhar | 18 June 2013, 18:16:22 UTC | Moved reduction operations to thrust calls | 18 June 2013, 18:16:22 UTC |
d9d4148 | Justin Madsen | 18 June 2013, 17:40:39 UTC | demo_soilbin.cpp working (somewhat). fixed some typos in demo_forklift.cpp | 18 June 2013, 17:40:39 UTC |
3767866 | Justin Madsen | 17 June 2013, 20:54:39 UTC | removed direct.h | 17 June 2013, 20:54:39 UTC |
f54aa5b | Justin Madsen | 17 June 2013, 20:50:18 UTC | fixed demo_soilbin, put it in the correct directory. | 17 June 2013, 20:50:18 UTC |
9155194 | Hammad Mazhar | 17 June 2013, 20:49:04 UTC | Fixed cohesion code for the frictional case for unit GPU, added bilaterals to GPU code | 17 June 2013, 20:49:04 UTC |
ac8c41e | Justin Madsen | 17 June 2013, 00:31:02 UTC | beginning integration of soft soil terrain database functionality | 17 June 2013, 00:31:02 UTC |
53294a4 | Hammad Mazhar | 14 June 2013, 22:12:35 UTC | added cohesion to gpu code, only supported in APGD solver at the moment, need to clean this up still | 14 June 2013, 22:12:35 UTC |
bf8d686 | Hammad Mazhar | 14 June 2013, 04:46:39 UTC | Updated OpenGL code, can draw triangle meshes now Added triangle mesh asset GPU collision detection supports triangle meshes again removed reference in CHstream.h as it was causing an error in the new fem demo when a ChVector<> was streamed out to the log file. Working on getting bilateral working with all solvers on GPU | 14 June 2013, 04:46:39 UTC |
af670f5 | Alessandro Tasora | 12 June 2013, 14:37:00 UTC | Added support for statics (linear) in ChSystem, using DoStaticsLinear. Note that the LCP solver must be PMINRES for the moment, other cannot work. Also note that it is still experimental. Look at demo_FEMbasic.cpp Also, obsolete code from ChObject removed. | 12 June 2013, 14:37:00 UTC |
e3e05b1 | Alessandro Tasora | 10 June 2013, 21:12:51 UTC | | 10 June 2013, 21:12:51 UTC |
a04aba5 | Alessandro Tasora | 10 June 2013, 14:45:35 UTC | | 10 June 2013, 14:45:35 UTC |
f87ce7f | Alessandro Tasora | 07 June 2013, 21:43:58 UTC | Developement of FEM framework in progress. | 07 June 2013, 21:43:58 UTC |
d3b4e81 | Hammad Mazhar | 07 June 2013, 00:30:37 UTC | Fixing compilations issued on OSX due to pthread_spin functions not being defined. | 07 June 2013, 00:30:37 UTC |
c08e390 | Hammad Mazhar | 06 June 2013, 02:55:17 UTC | adding in some linear interpolation math for unit GPU | 06 June 2013, 02:55:17 UTC |
4a0c8dd | Andrew Seidl | 05 June 2013, 21:32:44 UTC | Adding MPI_Wtime support | 05 June 2013, 21:32:44 UTC |
4897c56 | Andrew Seidl | 05 June 2013, 20:17:40 UTC | Removing extraneous message | 05 June 2013, 20:17:40 UTC |
2aa6513 | Alessandro Tasora | 02 June 2013, 09:25:42 UTC | The PMINRES solver is now able to solve also problems with ChLcpKstiffness items. Also, the ChLcpSystemDescriptor uses OpenMP as much as possible. A small optimization on flags in ChLcpConstraints | 02 June 2013, 09:25:42 UTC |
63bf2a5 | Hammad Mazhar | 31 May 2013, 21:56:36 UTC | Fixed issues with drawing ellipsoids and cylinders | 31 May 2013, 21:56:36 UTC |
e0593ae | Hammad Mazhar | 31 May 2013, 21:55:56 UTC | streamlined APGD code, trying to get the optimized version working again. Tested if using single threaded Shur A is faster for small problems, it isn't | 31 May 2013, 21:55:56 UTC |
6872243 | Alessandro Tasora | 30 May 2013, 23:08:34 UTC | The ChLcpIterativePMINRES solver now also supports the presence of ChLcpKstiffness blocks in the system description. For the moment, it is the only solver that is capable of doing so. | 30 May 2013, 23:08:34 UTC |
1246bb5 | Alessandro Tasora | 30 May 2013, 22:53:30 UTC | | 30 May 2013, 22:53:30 UTC |
69e6db1 | Hammad Mazhar | 28 May 2013, 21:37:58 UTC | SMall bug with convergence/number of iterations performed in APGD solver | 28 May 2013, 21:37:58 UTC |
660c24d | Hammad Mazhar | 24 May 2013, 20:16:12 UTC | Trying to resolve freeze issue for APGD solver | 24 May 2013, 20:16:12 UTC |
edf287b | Alessandro Tasora | 17 May 2013, 17:46:36 UTC | Started the developement of FEM code. A basic framework is planned in the source/fem directory. Old stuff (CHfem.h and such) is moved into /fem, but will be modified/removed in future. Also, new functionality for LCP solver (classes ini source/lcp) for handling stiffness block matrices overlayed to the mass matrix, for finite element simulations. Under progress, to be tested. | 17 May 2013, 17:46:36 UTC |
b3ca99b | Hammad Mazhar | 16 May 2013, 20:39:55 UTC | Added timers to collision system GPU | 16 May 2013, 20:39:55 UTC |
6bb9d2c | Alessandro Tasora | 16 May 2013, 20:03:43 UTC | Tested with Python 3.3. | 16 May 2013, 20:03:43 UTC |
dab0eeb | Alessandro Tasora | 16 May 2013, 20:03:22 UTC | Tested with Python 3.3. | 16 May 2013, 20:03:22 UTC |
55c642b | Alessandro Tasora | 13 May 2013, 19:20:57 UTC | | 13 May 2013, 19:20:57 UTC |
f8c2b4f | Alessandro Tasora | 10 May 2013, 15:49:25 UTC | Started to introduce the support for FEM. The LCP solver data structures needed some modifications. The ChLcpKstiffness classes have been added too, but there is not yet a solver that supports them. To be continued... | 10 May 2013, 15:49:25 UTC |
46f8eb0 | Hammad Mazhar | 07 May 2013, 06:10:46 UTC | tweaked solver code, testing changes to use less CD memory | 07 May 2013, 06:10:46 UTC |
cf12a43 | Hammad Mazhar | 07 May 2013, 03:30:54 UTC | Added code to get residual from solvers | 07 May 2013, 03:30:54 UTC |
b9045e2 | Hammad Mazhar | 02 May 2013, 16:02:34 UTC | Fixed issue with ellipsoids and a bug in the narrowphase | 02 May 2013, 16:02:34 UTC |
e1adec3 | Hammad Mazhar | 30 April 2013, 23:23:11 UTC | Added ellipsoid asset and geometry classes Fixed issue with ellipsoid CD in narrowphase with a new support function | 30 April 2013, 23:23:11 UTC |
c0c4452 | Hammad Mazhar | 30 April 2013, 16:08:18 UTC | Mitsuba export was not looping over all Chassets | 30 April 2013, 16:08:18 UTC |
a29a0ba | Alessandro Tasora | 29 April 2013, 21:54:32 UTC | | 29 April 2013, 21:54:32 UTC |
a47dd14 | Alessandro Tasora | 29 April 2013, 21:54:20 UTC | | 29 April 2013, 21:54:20 UTC |
eca5362 | Alessandro Tasora | 29 April 2013, 21:54:07 UTC | | 29 April 2013, 21:54:07 UTC |
3b57d87 | Alessandro Tasora | 29 April 2013, 21:29:36 UTC | | 29 April 2013, 21:29:36 UTC |
c485658 | Alessandro Tasora | 29 April 2013, 21:29:29 UTC | | 29 April 2013, 21:29:29 UTC |
2728901 | Alessandro Tasora | 29 April 2013, 21:25:16 UTC | | 29 April 2013, 21:25:16 UTC |
a850f7c | Alessandro Tasora | 29 April 2013, 21:19:47 UTC | | 29 April 2013, 21:19:47 UTC |
5d9945c | Alessandro Tasora | 29 April 2013, 20:59:17 UTC | | 29 April 2013, 20:59:17 UTC |
441f4f6 | Hammad Mazhar | 29 April 2013, 01:28:30 UTC | APGD and GPU code now fully parallel, performance is much better and scales with added cores. | 29 April 2013, 01:28:30 UTC |