erwin.coumans
aa6f352ddc
use #include "btScalar.h" instead of "LinearMath/btScalar.h"
...
fixes Issue 572
2011-12-12 16:56:39 +00:00
erwin.coumans
e1c9b67ef7
reduce dynamic memory allocations in btDbvt::rayTestInternal,
...
fixes Issue 574, thanks to Anton Breusov
2011-12-12 16:52:52 +00:00
erwin.coumans
606d65d95b
Reduce dynamic memory allocation, fixes Issue 573
...
Thanks to Anton Breusov for the patch
2011-12-12 16:47:41 +00:00
erwin.coumans
665829bd81
only use the getInterpolationWorldTransform for objects that are not isStaticOrKinematic
...
This should solver the issue discussed here: http://bulletphysics.org/Bullet/phpBB3/viewtopic.php?t=7681
remove some warning
2011-12-12 13:33:34 +00:00
erwin.coumans
76785d6e01
Only set the m_interpolationWorldTransform to m_worldTransform for kinematic objects, and use the new transform directly for static/dynamic objects.
...
update some cmake project files
2011-12-12 11:53:20 +00:00
erwin.coumans
a06ee28aa6
fix OpenCL kernel
2011-11-30 18:44:00 +00:00
erwin.coumans
866dc2c6c7
operator argument had to be const
2011-11-22 00:21:48 +00:00
erwin.coumans
0e6926c096
avoid compile errors in projects (btHashMap copied gives errors, see http://gamekit.org/forum/viewtopic.php?f=9&t=29&start=93
2011-11-22 00:03:46 +00:00
erwin.coumans
4e0aa42123
Disallow the operator= for btAlignedObjectArray, together with some fixes in the code (to avoid accidental deep copy)
...
Enable #define BT_ALLOW_ARRAY_COPY_OPERATOR in Bullet/LinearMath/btAlignedObjectArray.h to re-enable it
Fixes Issue 564, thanks to Tissen Peter for the report/fix
2011-11-15 20:09:05 +00:00
erwin.coumans
d3c6f9dd67
BulletWorldImporter didn't properly import m_3indices16 arrays, fixes Issue 562 thanks to Marius Elvert
...
Support for m_3indices8 import
Fix un-initialized memPtr->m_3indices8 at export time, add a workaround at loading time
2011-11-15 18:40:45 +00:00
erwin.coumans
f64fc0a26b
avoid array out-of-bounds access
2011-11-15 18:16:42 +00:00
erwin.coumans
21ecc26fde
fix compile issue under linux
2011-11-14 23:29:07 +00:00
erwin.coumans
0ccc82a2ab
disable the OpenCL asserts by default (BT_SUPPRESS_OPENCL_ASSERTS), and re-enable the shader initialzation in bool btOpenCLSoftBodySolver::checkInitialized()
2011-11-14 23:12:33 +00:00
erwin.coumans
0d1fcf7c48
Move btOclUtils.cpp/h and btOclCommon.cpp/h to btOpenCLUtils.cpp/h
...
Rename member numerator to m_numerator in btConvexHullComputer.cpp to avoid confusion (Thanks to Ian for the feedback)
Add btSoftBody::addAeroForceToNode and btSoftBody::addAeroForceToFace, thanks to Dongsoo Han/Saggita
2011-11-14 22:22:49 +00:00
erwin.coumans
a374f8c0d4
remove btOclGetMaxFlopsDev, it was buggy. Just use device 0 in the demos. The OpenCL initialization needs more work/some GUI.
2011-11-11 20:11:03 +00:00
erwin.coumans
66c349caa6
Remove btSoftBodySolver_CPU.*
...
Move btSoftBodySolverData.h to src/BulletMultiThreaded/GpuSoftBodySolvers/Shared/btSoftBodySolverData.h
Attempt to re-enable MiniCL version of OpenCLClothDemo (cloth-capsule collision still broken)
Add optional OpenCL acceleration to SerializeDemo (just for cloth)
2011-11-11 19:00:26 +00:00
erwin.coumans
14352169ab
Fix some array out-of-bounds access errors
...
Add an assert using MAX_DEGENERATE_STATS
Fixes Issue 561, thanks to jarno for the report
2011-10-24 18:24:31 +00:00
erwin.coumans
a415ef6067
apply patch to fix incorrect co-planar polygon merging
...
Thanks to fastflo for the report/patch. This fixes Issue 557
2011-10-18 15:41:14 +00:00
erwin.coumans
04285116ad
Use a different check for duplicate vertices,
...
Thanks to fastflo, see http://code.google.com/p/bullet/issues/detail?id=555
and http://bulletphysics.org/Bullet/phpBB3/viewtopic.php?f=9&t=7511&p=25762#p25762
Added missing Intel OpenCL build file, and fix some compile issue due to missing external
2011-10-12 22:40:43 +00:00
erwin.coumans
8b849e0d07
Fixed split impulse SIMD issue in the constraint solver (doesn't affect default behaviour, as split impulse is switched off by default)
...
Thanks to Laurent Coulon for the report/fix. http://bulletphysics.org/Bullet/phpBB3/viewtopic.php?f=9&t=7515&p=25766#p25766
2011-10-12 20:22:55 +00:00
erwin.coumans
961fb12985
also set the properties for the mirror version of the plane-convex algorithm (convex-plane)
...
This fixes Issue 548, thanks to ejtttje
2011-09-19 22:10:32 +00:00
erwin.coumans
4f25d731f9
fix autotools build
...
fix picking for tetrahedral soft body (demo 29/30)
add text for failing OpenCL kernel build, thanks to Greg
2011-09-16 23:58:38 +00:00
erwin.coumans
d3dc7afefd
fix in non-SIMD OpenCL cloth demo, kernel names changed.
...
Thanks Greg, see http://bulletphysics.org/Bullet/phpBB3/viewtopic.php?f=9&t=7442&p=25422#p25422
2011-09-16 21:58:01 +00:00
erwin.coumans
bb29d15058
create premake4 solutions for the various OpenCL configurations (AMD, NVIDIA, Intel OpenCL SDKs for Windows)
...
primarily to distribute project files that work.
CMake is still better supported in general (not all demos/libs have been ported)
Revert a recent change about warnings: %zu doesn't work in printf, %d does
2011-09-16 01:28:34 +00:00
erwin.coumans
ce62d7615e
Update version to 2.79 for a new bugfix release
2011-09-15 23:36:59 +00:00
erwin.coumans
654fe7872a
remove btContinuousDynamicsWorld, it is obsolete and was never completed.
...
If you need tunneling prevention, one option is to use the CCD motion clamping in btDiscreteDynamicsWorld. See also Bullet/Demos/CcdPhysicsDemo.
You can use an embedded sphere that prevents tunneling, using the following settings:
body->setCcdMotionThreshold(0.5);
body->setCcdSweptSphereRadius(0.9f);
Add a camera zoom setting for the demos, thanks to ejtttje, fixes Issue 543
2011-09-15 18:47:13 +00:00
erwin.coumans
69a932f13e
Choose a default 'safe' collision margin for very small convex collision shapes, in particular btBoxShape and btCylinderShape
...
and add some documentation in btConvexInternalShape.h
Thanks to Simon Lundmark for the suggestion
Fixes http://code.google.com/p/bullet/issues/detail?id=349
2011-09-15 18:17:20 +00:00
erwin.coumans
a13d22dc84
Fix a lot of warnings, thanks to ejtttje, Fixes issue 537
...
Fix soft body debug rendering
2011-09-15 01:17:34 +00:00
erwin.coumans
b68c64f2db
Disable DX11 cloth bounds update, it is too slow by default.
...
Use g_dx11Solver->setEnableUpdateBounds(true); to re-enable it (it enables GPU collision of cloth versus capsule)
2011-09-14 20:55:00 +00:00
erwin.coumans
385933b5ed
Expose the multi-point perturbation algorithm for btConvexPlaneCollisionAlgorithm
...
Usage:
m_collisionConfiguration = new btDefaultCollisionConfiguration();
m_collisionConfiguration->setPlaneConvexMultipointIterations();
It is only enabled for polyhedral convex shapes, because implicit/smooth surfaces such as spheres, cylinders, capsules and cones keep on rolling forever, due to the extra off-center contact points
2011-09-14 20:38:33 +00:00
erwin.coumans
fe6991822d
Use the original code for convex-plane collision in (btConvexPlaneCollisionAlgorithm),
...
rather than the perturbed rotated version (it is slower and harder to debug)
2011-09-14 01:28:40 +00:00
erwin.coumans
b690019e39
Removed global operator- in btQuaternion, and rewrote the 'getAxis' method
...
Add a unit test for testQuaternionGetAxisAngle
Thanks to ejtttje, fixes Issue 538
2011-09-13 23:37:05 +00:00
erwin.coumans
1e360bdb6b
add Intel build support for the OpenCL cloth demo, and use the bullet_logo.png
...
Thanks to ggjunker for the patch, see Issue 533
2011-09-13 01:52:42 +00:00
erwin.coumans
82234e6f33
add preliminary premake4 support (next to cmake) mainly to auto-generate Visual Studio projectfiles that can be redistributed (with no dependency on cmake, premake)
2011-09-02 22:24:08 +00:00
erwin.coumans
ed02816e68
add kst argument for VSolveLinks.cl kernel (VSolveLinksKernel method)
...
Note that this method is not used by default.
Thanks to Peter for the report and fix, this fixes Issue 511
2011-09-02 04:30:01 +00:00
erwin.coumans
06112592fd
Add another wind model, that doesn't clamp the maximum force.
...
Thanks to saggitasaggita for the patch, see http://code.google.com/p/bullet/issues/detail?id=532
2011-09-01 01:53:45 +00:00
erwin.coumans
642a6821a5
remove speculative contacts, they break restitution and cause other artifacts
...
use btConvexPolyhedron for debug rendering, if available
fixes in btConvexConvexAlgorithm for polyhedral contact clipping, if GJK separating normal is zero
don't shift vertices in btPolyhedralConvexShape
2011-08-30 03:28:32 +00:00
erwin.coumans
7743f5bce5
Update environment variable to CUDA_PATH (was NVSDKCOMPUTE_ROOT) and fix a few bugs that prevented the NVIDIA OpenCLClothDemo to run properly.
2011-08-17 22:08:10 +00:00
erwin.coumans
0336ed8637
prepare support for picking softbody tetrahedra
2011-08-06 21:50:09 +00:00
erwin.coumans
e83b5dac75
fixes in the DX11 cloth (for devices with physical wavefronts smaller than 32)
...
another fix for out-of-bounds check (there was no assert for this previously)
2011-08-03 20:21:23 +00:00
erwin.coumans
16cc5665de
fix in the breaking constraints feature
2011-07-01 21:51:38 +00:00
erwin.coumans
251bb4e727
remove the 'exit' calls from the OpenCL soft body solver. Use the solver->checkInitialized() method, after solver->optimize() to see if the kernels build OK.
2011-06-30 00:23:42 +00:00
erwin.coumans
fcb1448f21
fix member variable name
2011-05-22 11:01:46 +00:00
erwin.coumans
b80e5fd167
Remove co-planar faces from convex hull, using 2d Graham scan
...
Improve SAT performance, by skipping back-facing features
Add assert in array class (probably fires in places)
2011-05-20 12:29:24 +00:00
erwin.coumans
20e95be9cd
Fix regression issue: transform is not identity in case of softbody colliding against btCompoundShape.
...
Fix issue 512
2011-05-11 20:51:57 +00:00
erwin.coumans
cce582d905
Add support for AMDAPPSDKROOT environment variable
...
Initialize vSolveLinksKernel OpenCL cloth kernel (although kernel is unused)
Fixes Issue 508
2011-04-23 16:01:44 +00:00
erwin.coumans
6ccd5a644a
implement stub function for clGetProgramInfo to avoid linker errors
...
Fixes Issue 510
2011-04-23 15:47:49 +00:00
erwin.coumans
7dd0cbb726
fix OSX compilation
2011-04-15 19:17:34 +00:00
erwin.coumans
1b305562be
Added Pierre Terdiman's 'internal object' optimization to improve performance for separating axis tests.
...
Make the winding consistent in btConvexHullComputer (and related fixes in btPolyhedralConvexShape), thanks to Ole!
Some fixes in the btPolyhedralContactClipping implementation (never report a penetration deeper than GJK/EPA found, to avoid issues due to its approximate contact normal directions)
Properly visualize btPolyhedralConvexHullShape that have a btConvexPolyhedron (by calling initializePolyhedralFeatures() method)
2011-04-15 18:37:28 +00:00
erwin.coumans
7d37b3c472
add support for PosixThreadSupport for MiniCL (partial CPU OpenCL replacement)
2011-04-11 16:28:21 +00:00