That change removed some spurious contacts returned by mjc_BoxBox, but also removed some desirable contacts that occur during very deep penetration of two boxes (when one box is completely inside another box). This is now fixed.
PiperOrigin-RevId: 591036252
Change-Id: I84b51f2179fd6fe29618a4e908e86934c6fc6941
The box-box collider sometimes outputs bad contact points which are far outside one of the two boxes. This happens at configurations which are very rare numerically, but rare configurations (e.g. perfect box alignment) are often induced by the designer and the physics, making them actually not-so-rare.
Since the box collision functions are very difficult to understand, we circumvent the issue by extending the post processing code to detect and remove such contacts.
PiperOrigin-RevId: 589178197
Change-Id: I82b8a888ef04f9196111c960757e3314d0ab6d8e
- Fixed bug in timing of kinematics, was under-reported.
- Moved timing of inertia and collision into respective functions, in preparation for multi-threading.
- Detailed profiling in `testspeed` is always on, removed command-line flag.
- Improved `testspeed` self-documentation.
PiperOrigin-RevId: 571989548
Change-Id: If0e3cc17f6f3511b98fc36339f406a19968a9361
This results in approximately 4-5% speedup in collision detection for particle.xml.
PiperOrigin-RevId: 562526634
Change-Id: Ia6487b1908618064d858bac484fe6861ed37a402
At the leaves of the midphase trees in `mj_collideTree`, a bounding sphere test is performed. The margin was incorrectly set to zero instead of reading it from `geom_margin`.
PiperOrigin-RevId: 559782262
Change-Id: Icfceb8445bbf6977582e0061566d7bd4def2f0ae
Add -Wgnu-empty-initializer to Clang builds to catch this issue in the future. Initialising arrays with empty braces is not part of the C standard and fails on MSVC. This commit should catch the issue earlier by detecting it in Clang.
PiperOrigin-RevId: 517373961
Change-Id: I518006e84151e8d5b45a7fe00ba7675dfb031ca0
mj_contactFilter removed geom pairs where their bodies share the same weldid, but this check can be applied at broadphase instead of at geom-geom level.
PiperOrigin-RevId: 507435848
Change-Id: Ia37c741071dc5e1c26686dd319f42c9bd334d305
Sometimes broadphase returns the same body pair multiple times. In those cases, avoid creating every contact twice.
PiperOrigin-RevId: 499869235
Change-Id: I58606168b0723d40b4fdd43b1fafad79bf35d76d
- Add private function `mj_arenaAlloc`. This is used internally to allocate memory from the arena.
- Add private function `mj_nefc` to count constraints. This function returns a tight upper bound on `d->nefc`. The number of counted constraints can be slightly bigger than exact `d->nefc` in the case of constraints with empty Jacobian, as when placing a frictional tendon between two world sites.
- Add new `memory` attribute to the `size` XML element for specification of arena memory size. This attribute is mutually exclusive with `nstack` and `njmax` specifications, which are now deprecated (but left around for the time being for legacy compatibility).
- Move `d->stack` to the end of the new arena space. The stack now grows in reverse from the end.
PiperOrigin-RevId: 479341539
Change-Id: Ie019c202e0908577ffc6f833a37920858116f667
Yuval Tassa