- mjTIMER_ACCELERATION is not very useful, it's usually very tiny (0.5% for humanoid, 1% for 22humanoids). It basically times a single call to `mj_solveM`.
- Added `mjTIMER_ADVANCE` to time `mj_Euler` and `mj_implicit` (which can be significant).
- Moved `mj_transmission` to `mjTIMER_KINEMATICS` where it belongs (and used to be).
PiperOrigin-RevId: 565936880
Change-Id: I26b8a067133594e5500ca1c809beb3518b4a8d9c
See https://gitlab.kitware.com/cmake/cmake/-/issues/21146
BEGIN_PUBLIC
Prevent patches to be applied to an already patched dependency.
END_PUBLIC
PiperOrigin-RevId: 565624168
Change-Id: I29b36d2cd02f58d5d9e84f9cae5e5bbf8be1b404
The following dependencies are disabled with a cmake patch file.
- Enoki
- OpenMP
- Assimp
- FCPW
PiperOrigin-RevId: 565352386
Change-Id: Icff9de68408e0afa2032bb83b24491fc4e41d181
- See public description below.
- Stopped incrementing the iteration count in saveStats().
- `sizeof(mjSolverStat) == 40`, so this ends up costing 160KB, up from 40KB.
BEGIN_PUBLIC
Changed the size of `mjData.solver`, the structure used to collect solver diagnostic information. The array is now of length `mjNISLAND * mjNSOLVER`, where each row of length `mjNSOLVER` contains separate solver statistics for each constraint island. Until solver islanding is implemented, only row 0 is used.
- The new constant `mjNISLAND` is set to 20.
- `mjNSOLVER` is reduced from 1000 to 200.
- Added `mjData.solver_nisland`, the number of islands for which the solver ran.
- `mjData.solver_niter` (renamed from mjData.solver_iter) and `mjData.solver_nnz` are now integer vectors of length `mjNISLAND`.
END_PUBLIC
PiperOrigin-RevId: 565030093
Change-Id: I773e918805c6ced79f0dab5f19ea23956760c8c5
Add missing `ctrl` attribute to keyframe indexer.
END_PUBLIC
Context: Making an LQR notebook for the Skydio drone and wanted to read the ctrl attribute in the keyframe stored in the XML.
PiperOrigin-RevId: 564806350
Change-Id: I5af3189cc8be87915f9d320d95f0fab551ef87fa
The SDF is defined in the `geom` frame. Using a mesh changes this frame so the inverse transformation is needed. So far the full `geom_pos` and `geom_quat` were used, which possibly contained user transformation. This CL uses the new structures which only contain transformations due to mesh processing.
PiperOrigin-RevId: 564721461
Change-Id: I6fb28001b38eaa50663427bce452cb1f05a56ac0
Zero all of act_dot when there are custom actuator dynamics, but no callback.
END_PUBLIC
PiperOrigin-RevId: 564712105
Change-Id: I107481727a42eb2577d62e29ac4e1496c56bb634
This makes mj_forward calls idempotent if warmstart is disabled and quats in qpos are normalized.
PiperOrigin-RevId: 564442666
Change-Id: I9c43add82bcfb4101c19fbfbfc8c30091b8906d5
- Make mjTask non-opaque and remove C++ Task class.
- Make C++ thread pool a subclass of a skeletal mjThreadPool C struct.
- Make the mjTask status enum more consistent with the rest of MuJoCo.
- Change mju_threadPoolEnqueue to take just the mjTask. Users must now prepare the mjTask by assigning the function pointer and argument into the struct.
- Rename files in thread/ to be more consistent with the rest of MuJoCo.
- Run threading tests in CMake.
- Allow use of C++20 designated initializers.
Otherwise the functionality remains identical.
PiperOrigin-RevId: 564374675
Change-Id: I37c9894566bc39faf217e5aa97a4e2713a70e467
std::function can have an allocation cost when it's created, and also adds overhead to function calls. Since this API must be used from C, using a raw function pointer is good enough.
PiperOrigin-RevId: 562760324
Change-Id: I2b714a13e07a2b71566d013f21f76c506e817a52
Also add asan instrumentation to detect stack frame leakages (i.e. `mj_markStack` without a corresponding `mj_freeStack` in the same caller function).
PiperOrigin-RevId: 562625645
Change-Id: I4e3ff66ca0b9d08ed0a95cef45393db8e3053e22
This results in approximately 4-5% speedup in collision detection for particle.xml.
PiperOrigin-RevId: 562526634
Change-Id: Ia6487b1908618064d858bac484fe6861ed37a402
Alessio Quaglino