chenlin/Mujoco_WASM
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

Precompute rotated stiffness matrix before CG in implicit flex integrator.

Browse Source 
This gives a 3x speedup in implicitfast.

Also cleanup old code that was used in the dense factorization of the stiffness matrix before we switched to CG.

PiperOrigin-RevId: 915900315
Change-Id: Id6973c4bfd7d371a43ec6db982703969b23a3550
This commit is contained in:
Alessio Quaglino
2026-05-15 02:57:18 -07:00
committed by Copybara-Service
parent 85bf2b9660
commit 7ca1bc6a56
5 changed files with 104 additions and 107 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/engine/engine_derivative.h
+5 -1
View File
@@ -44,8 +44,12 @@ MJAPI void mjd_passive_vel(const mjModel* m, mjData* d);
MJAPI void mjd_rne_vel_dense(const mjModel* m, mjData* d);
// compute res += (s1 + s2*damping) * J'*K*J * vec, for all interpolated flexes
// K_rot_cache: if non-NULL, use pre-cached K_rot (same layout as m->flex_stiffness)
MJAPI void mjd_flexInterp_mul(const mjModel* m, mjData* d, mjtNum* res, const mjtNum* vec,
mjtNum s1, mjtNum s2);
mjtNum s1, mjtNum s2, const mjtNum* K_rot_cache);
// precompute unscaled K_rot for all elements into cache (same layout as m->flex_stiffness)
MJAPI void mjd_flexInterp_cacheKrot(const mjModel* m, mjData* d, mjtNum* K_rot_out);
// compute res += scale * K_bend * vec for standard (non-interp) flex bending
// scale = s1 + s2 * flex_damping[f] per flex