Introduce trilinear flex parametrization.

These flexes use only 24 DOFs (3 per vertex of the bounding box), while colliding with the full high resolution mesh.

On an 8x8x8 cube, the performance using DOFs at all vertices is

```
 Simulation time      : 18.74 s
 Steps per second     : 533
 Realtime factor      : 0.53 x
 Time per step        : 1874.4 µs

 Contacts per step    : 114.88
 Constraints per step : 3322.51
 Degrees of freedom   : 1536
```

With the new implementation, it is the following:

```
 Simulation time      : 1.82 s
 Steps per second     : 5507
 Realtime factor      : 5.51 x
 Time per step        : 181.6 µs

 Contacts per step    : 38.84
 Constraints per step : 155.36
 Degrees of freedom   : 24
```

PiperOrigin-RevId: 721008829
Change-Id: I833df027527db578d86667cc4b24295bcf6f7d22

doc/changelog.rst

@@ -5,8 +5,27 @@ Changelog
 Upcoming version (not yet released)
 -----------------------------------
 
+
+Feature promotion
+^^^^^^^^^^^^^^^^^
+.. youtube:: qJFbx-FR7Bc
+   :align: right
+   :width: 240px
+
+- Introduced a new kind of **fast deformable body**, activated by setting :ref:`flexcomp/dof<body-flexcomp-dof>` to
+  "trilinear". This type of :ref:`deformable<CDeformable>` flex object has the same collision geometry as a regular
+  flex, but has far fewer degrees of freedom. Instead of 3 dofs per vertex, only the corners of the bounding box are
+  free to move, with the positions of the interior vertices computed with trilinear interpolation of the 8 corners, for
+  a total of 24 dofs for the entire flex object (or less, if some of the corners are pinned). This limits the types of
+  deformation achievable by the flex, but allows for much faster simulation. For example, see the video on the right
+  comparing `full <https://github.com/google-deepmind/mujoco/blob/main/model/flex/gripper.xml>`__ and `trilinear
+  <https://github.com/google-deepmind/mujoco/blob/main/model/flex/gripper_trilinear.xml>`__ flexes for modeling
+  deformable gripper pads.
+
 General
 ^^^^^^^
+- Separate collision and deformation meshes for :ref:`flex<deformable-flex>`. This enables a fixed cost for the soft
+  body computations, while preserving the fidelity of high-resolution collisions.
 - Added :ref:`mjs_setDeepCopy` API function. When the deep copy flag is 0, attaching a model will not copy it to the
   parent, so the original references to the child can be used to modify the parent after attachment. The default
   behavior is to perform such a shallow copy. The old behavior of creating a deep copy of the child model while