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

Fix documentation typos and CMake framework setup for macOS.

Browse Source 
The previous commit missed out some recent changes in the internal repo. This commit brings the public GitHub repo in sync.

PiperOrigin-RevId: 450377010
Change-Id: I59a68ae4a39c5cd5ed5f7df77ba870273243d623
This commit is contained in:
Saran Tunyasuvunakool
2022-05-23 14:03:09 +01:00
parent 64bc6d27b2
commit 7eb4ab3a8f
8 changed files with 103 additions and 384 deletions
Download Patch File Download Diff File Expand all files Collapse all files
doc/overview.rst
+1 -1
View File
@@ -602,7 +602,7 @@ Divergence of a simulation happens when elements of the state tend quickly to in
manifested as an :ref:`mjWARN_BADQACC<mjtwarning>` warning. Divergence is endemic to all physics simulation and is not
necessarily indicative of a bad model or bug in the simulator, but is rather a hint that the timestep is too large for
the given choice of integrator. In physics simulation there is always a tension between speed (large time steps) and
stabillity (small timesteps). A model which is well-tuned for speed has the largest possible timestep that does not
stability (small timesteps). A model which is well-tuned for speed has the largest possible timestep that does not
diverge, which usually means that it *can* be made to diverge under extreme conditions. In that sense *rare* cases of
divergence can actually be indicative of a well-tuned model. In all cases it should be possible to prevent divergence by
reducing the timestep and/or switching to a more stable :ref:`integrator <geIntegration>`. If that fails, the culprit is