Commit Graph

143 Commits

Author SHA1 Message Date
erwin.coumans
561066af75 + make compound versus soft body work (soft body uses interpolated transform)
+ fixed issue with persistent manifold, warmstarting values were not initialized properly
+ don't clear manifold in sphere-sphere collision (need warmstarting)
+ added support for 'split impulse', decouple positional error correction from velocity correction
This avoids adding momentum due to penetration correction, it can be tuned using following variables:
solverInfo.m_splitImpulse = true/false (disable/enable)
solverInfo.m_splitImpulsePenetrationThreshold (below this value, baumgarte/mixed velocity/penetration is used (cheaper, looks more plausible)
solverInfo.m_linearSlop (less jitter, when small amound of penetration is allowed)
2008-05-23 09:05:37 +00:00
erwin.coumans
bd97c5e569 Fixed warnings in Bullet/src core library
Thanks Martijn Reuvers from Two Tribes B.V. (www.twotribes.com) for the patch

To make this work more visible, suppress warnings in external libraries in Extras (COLLADA_DOM, libxml and glui contain many warnings)
Added PreprocessorDefinitions: _CRT_SECURE_NO_DEPRECATE;_CRT_NONSTDC_NO_DEPRECATE to vcproj files
2008-05-10 18:00:21 +00:00
ejcoumans
a4e63aed00 - DemoApplication, debug text rendering was broken, glLoadIdentity() missing
- copy user data over to child shape
- added applied impulse to btManifoldPoint
- add ContactProcessedCallback (needs test/demo)
- didn't copy over m_additionalDampingFactor into btRigidBody.
2008-02-07 08:00:16 +00:00
ejcoumans
c0055bdb4a removed some profiling in the innerloop, it hurts performance 2008-01-18 06:48:55 +00:00
ejcoumans
17a214a2b3 - Added btRigidBodyConstructionInfo, to make it easier to set individual setting (and leave other untouched) during rigid body construction.
This was harder using default arguments. Thanks Vangelis Kokkevis for pointing this out.
- Fixed memoryleak in the ConstraintDemo and Raytracer demo.
- fixed issue with clearing forces/gravity at the end of the stepSimulation, instead of during internalSingleStepSimulation.
Thanks chunky for pointing this out: http://www.bulletphysics.com/Bullet/phpBB3/viewtopic.php?f=9&t=1780
- Disabled additional damping in rigid body by default, but enable it in most demos. Set btRigidBodyConstructionInfo m_additionalDamping to true to enable this.
- Removed obsolete QUICKPROF BEGIN/END_PROFILE, and enabled BT_PROFILE. Profiling is enabled by default (see Bullet/Demos/OpenGL/DemoApplication.cpp how to use this).
User can switch off profiling by enabling define BT_NO_PROFILE in Bullet/src/btQuickprof.h.
2007-12-17 04:26:36 +00:00
ejcoumans
e252f5cadd Bugfix: allow btCollisionObjects (non-btRigidBody) to interact properly with btRigidBody for cache-friendly btSequentialImpulseConstraintSolver.
Thanks Andy O'Neil for pointing this out.
2007-11-28 03:17:09 +00:00
ejcoumans
cab75b53ec added hierarchical profiling (needs more cleanup)
avoid dynamic allocations in btRaycastVehicle
2007-11-21 03:00:40 +00:00
ejcoumans
a101719687 - removed warnings in GLUI, thanks Jorrit Tyberghein for the patch!
- removed memory leaks from BasicDemo and Bullet library (other demos needs to be cleaned up!)
- added memory leak debugging functionality in btAlignedAlloc.h: #define BT_DEBUG_MEMORY_ALLOCATIONS
2007-10-31 08:00:03 +00:00
ejcoumans
ec76f2e0a3 - keep track of all memory allocations (gNumAllignedAllocs/gNumAllignedFree)
All memory allocations in Bullet go through btAlignedAlloc/btAlignedFree
Fix in hinge constraint constructors, thanks Marcus Hennix!
2007-10-22 22:23:10 +00:00
ejcoumans
e7caaa28d3 re-organized memory (stack and pool) allocators. this lets the user pass in their own memory allocators. 2007-10-20 02:23:39 +00:00
ejcoumans
eff4fe8ec8 more work on hashed pairmanager. growing doesn't work yet, so need to allocate enough room for the overlapping pairs in advance.
boxbox reports contact point in B, rather then average point
box, cylinder use halfextents corrected for scaling and margin. made the margin in this halfextents explicit in the 'getHalfExtentsWithMargin' and 'getHalfExtentsWithoutMargin'
integrated changed for ODE quickstep solver
replaced inline with SIMD_FORCE_INLINE
some minor optimizations in the btSequentialImpulseConstraintSolver

added cone drawing (for X,Y,Z cones)
2007-10-12 02:52:28 +00:00
ejcoumans
0405ce278a use size_t instead of int, for allocator
added hashtable based PairManager, thanks Pierre Terdiman and Erin Catto
improved friction in 'cachefriendly' solver
moved 'refreshcontactpoints' into collision detection, instead of solver
avoid linear search for contact manifolds, by storing an index
ignore margin for sphere shape (its entire radius is already margin)
avoid alignment checks in BVH serialization, they don't compile on 64-bit architectures
made 'bomb' box more heavy
2007-10-11 03:17:54 +00:00
ejcoumans
811c105c24 align btQuadWord on 16byte boundary
slightly improved friction model
2007-09-25 06:58:53 +00:00
ejcoumans
f50f3ad722 fix for restitution
added preliminary continuous dynamics world (under construction, not ready yet)
fix crash on when restarting simplex demo using spacebar, Thanks Jorrit Tyberghein for reporting!
2007-09-20 05:24:03 +00:00
ejcoumans
15df0a7d7a revert to previous version of constraint solver, until the memory allocation issues are sorted properly. 2007-09-10 19:19:52 +00:00
ejcoumans
ab33083a9c make visual studio 6 compile/run (doesn't support 'long long int', so replace by char* pointers... 2007-09-10 01:53:07 +00:00
ejcoumans
b054f375bc - Added serialization to btBvhTriangleMeshShape/btOptimizedBvh. See ConcaveDemo for example usage.
- added bt32BitAxisSweep3, which co-exists without recompilation, using template class. This broadphase is recommended for large worlds with many objects (> 16384), until btMultiSwap is finished.
- Fixed some recent issues in Bullet 2.57 related to compound (thanks Proctoid) and memory allocations
2007-09-10 01:14:42 +00:00
ejcoumans
87df3d0f32 Based on feedback from another professional game company, there are several improvements, including some API change...
Some dynamic memory allocations have been replace by pool allocation or stack allocations.
quantized aabb versus quantized aabb overlap check is made branch-free (helps a lot on consoles PS3/XBox 360)
Collision algorithms are now created through a new btDefaultCollisionConfiguration, to decouple dependency (this is the API change):
Example:
	btDefaultCollisionConfiguration* collisionConfiguration = new btDefaultCollisionConfiguration();
	m_dispatcher = new	btCollisionDispatcher(collisionConfiguration);
2007-09-08 05:40:01 +00:00
ejcoumans
0df4cedc3e added 'getConstraintSolver' and ability to 'reset' (set randseeds back to zero, and clear internal caches) 2007-07-27 23:54:22 +00:00
ejcoumans
5f8873c339 support diamondsubdivision (thanks Jay for contribution)
support user-defined broadphase collision filtering (thanks Marten)
make sure btSimpeDynamicsWorld doesn't use cache friendly optimization in the solver (added an assert, and updated BasicDemo)
2007-06-29 22:13:15 +00:00
ejcoumans
82c8687881 - initialize m_unusedW to avoid some warnings.
- added angularFactor to 'cachefriendly' constraint solver
2007-04-23 10:52:38 +00:00
ejcoumans
bc3f9535ad cleaned up, removed warning under MSVC2005 (Level 4)
Mostly related to alignment and unused variables
2007-04-13 01:37:21 +00:00
ejcoumans
c1a54d9edc Attempts to improve performance. Not much gain yet, but good to experiment what has effect and what hasn't.
Added 'DO_BENCHMARK_PYRAMID' to CcdPhysicsDemo.
2007-03-20 20:12:23 +00:00
ejcoumans
f199cd1347 make visual studio 6 compile again 2007-03-20 05:06:01 +00:00
sjbaker
6b102c65b3 Cleanup some trivial compilation warnings. 2007-03-19 04:27:59 +00:00
ejcoumans
b8ce13adb1 bugfix, also convert rigidbody into solverbodies for constraints (without contact manifolds), and do additional check for validity 2007-03-17 08:29:05 +00:00
ejcoumans
2b87104184 more solver optimizations, can be disabled using solver->setSolverMode(SOLVER_RANDMIZE_ORDER) 2007-03-17 07:59:27 +00:00
ejcoumans
151cd4b9da prepare and added constraint solver optimizations, not activated yet. 2007-03-17 00:09:12 +00:00
ejcoumans
adfdea5ca6 replaced #include "new.h" to #include <new> 2007-03-06 23:40:06 +00:00
ejcoumans
054d672592 - removed STL from the Bullet library: replace std::vector by btAlignedObjectArray. Also removed the std::set for overlapping pair set, and turned it into an overlapping pair array. The SAP only adds objects, never removed. Removal is postponed for during traversal of overlapping pairs (duplicates and non-overlapping pairs are removed during that traversal).
- added heap sort and binary search/linear search to btAlignedObjectArray
- fixed wrong cast, thanks Hamstray, http://www.continuousphysics.com/Bullet/phpBB2/viewtopic.php?t=1015
2007-03-06 09:59:17 +00:00
ejcoumans
b73900bc60 Improved performance of convex collision shapes, cache local AABB instead of recomputation. This fixes issue with very slow performance in larger .bsp levels.
Moved some asserts into 'btFullAssert', which is disabled by default (see btScalar.h to enable them). This is to speed-up debugging.
2007-02-26 04:59:05 +00:00
ejcoumans
bf68a742cc by default, disable warmstarting, enable random ordering,
user can override this using solver->setSolverMode(...)
2007-02-13 16:14:55 +00:00
ejcoumans
df9230327c Contribution to add optional double precision floating point support. Define BT_USE_DOUBLE_PRECISION for all involved libraries/apps. 2006-12-16 05:51:30 +00:00
ejcoumans
6dff5a218e process contact and non-contact constraints inside the same iteration loop
added first draft for hingeConstraint motor
2006-12-12 03:15:11 +00:00
ejcoumans
7dc6aa215a some more fixes to get msvc6 happy, and constraint solver: make non-randomizing solver default 2006-11-23 02:57:38 +00:00
ejcoumans
b07bb88a2d Added new EPA penetration depth solver
improved gjk/minkowski sampling pd method
added original solver variant, use btSequentialImpulseConstraintSolver2
Added Pierre Terdimans PD testbed
2006-11-16 06:14:12 +00:00
ejcoumans
23c64fb0db more solver experiments, randomize the order of contact points, not just manifolds
use #defines for constants, rather then const btScalar
2006-11-04 05:22:36 +00:00
ejcoumans
07873854a4 more small refactoring: some methods were still upper-case. 2006-11-03 04:32:48 +00:00
ejcoumans
4050da0e2f One of the last parts of the refactoring (hopefully), made most members of btCollisionObject protected.
Also did some work on improving the constraint solver.
2006-11-02 03:42:53 +00:00
ejcoumans
3fe3b11924 use Dispatcher in ConcaveConvexCollisionAlgorithm (so it uses the registered collision algorithm, not hardcoded convexconcave)
improved performance of constraint solver by precalculating the cross product/impulse arm
added collision comparison code: ODE box-box, also sphere-triangle
added safety check into GJK, and an assert for AABB's that are very large
write partid/triangle index outside of GJK
2006-10-28 02:06:19 +00:00
ejcoumans
2b1657b1dd Refactoring: another huge number of changes, renamed methods to start with lower-case. 2006-09-28 01:11:16 +00:00
ejcoumans
eb23bb5c0c merged most of the changes from the branch into trunk, except for COLLADA, libxml and glut glitches.
Still need to verify to make sure no unwanted renaming is introduced.
2006-09-27 20:43:51 +00:00
ejcoumans
0e04cfc806 First stage in refactoring Bullet: moved Bullet Collision and Dynamics and LinearMath into src folder, and all files in Collision Detection and Dynamics have bt prefix.
Made all buildsystems to work again (jam, msvc, cmake)
2006-09-25 08:58:57 +00:00