Optimize mj_flex performance.

Two performance optimizations for interpolated flex objects:

1. Hoist loop-invariant stride calculations in `mju_cellLookup` out of nested loops.
   This reduces multiplications from 16 to 6 (linear) and 54 to 12 (quadratic)
   per vertex.

2. Extract and consolidate optimized 3D interpolation logic into a new
   reusable utility `mju_evalBasisArray` in `engine_util_misc.c`. This function
   uses nested loops and precomputed 1D shape functions to avoid expensive
   dynamic `phi` calls and branching, and leverages stack-buffered outputs to
   eliminate compiler pointer-aliasing barriers. We propagate this optimization
   to both kinematics (`mju_interpolate3D`) and constraint setup
   (`engine_core_constraint.c`).

Together these changes yield a ~50% overall speedup in `mj_fwdKinematics` for interpolated flexes with ~10k vertices in the collision meshes and ~10 nodes in the deformation grid.

PiperOrigin-RevId: 922198093
Change-Id: I9c804e65cd532e2f9ac6e9a71d01c30cd64186fa

src/engine/engine_util_misc.h

@@ -89,6 +89,9 @@ MJAPI void mju_defGradient(mjtNum res[9], const mjtNum p[3], const mjtNum* dof,
 // evaluate the basis function at x for the i-th node
 MJAPI mjtNum mju_evalBasis(const mjtNum x[3], int i, int order);
 
+// evaluate the basis functions at x for all nodes in the cell
+MJAPI void mju_evalBasisArray(mjtNum* basis, const mjtNum x[3], int order);
+
 // map global parametric coord to cell-local coord and build node indices
 MJAPI int mju_cellLookup(const mjtNum coord[3], const int cellnum[3], int order, mjtNum local[3],
                          int* nodeindices);