Add notes about Python profiling in 1.10.13

Author: rdb

optimization/performance-issues/pstats-python-time.png

@@ -20,6 +20,17 @@ not mean you need to abandon Python and rewrite all your code in C++ for that
 single bottleneck. Only that single function or class can be ported to C++ and
 called from your Python code.
 
+To determine whether it is the Python code that is slowing down an application,
+you should :ref:`use PStats <measuring-performance-with-pstats>`. As of Panda3D
+version 1.10.13, you can set the ``pstats-python-profiler`` Config.prc variable
+to get a detailed view of how much time each Python module, class and function
+is taking up. To access this view, in the Frame strip chart, double-click the
+App collector on the left side, then double-click the Python collector. Then,
+you can drill down further into the packages, modules, classes and functions.
+
+.. image:: pstats-python-time.png
+   :width: 1005
+
 Before considering writing any C++ code, consider other options:
 
 -  See if you can optimize your code. The Python website has a

optimization/performance-issues/python-calculation.rst

@@ -75,6 +75,9 @@ the profiling machine:
 
 .. only:: python
 
+   Profiling Python Code
+   ~~~~~~~~~~~~~~~~~~~~~
+
    If you are developing Python code, you may be interested in reporting the
    relative time spent within each Python task (by subdividing the total time
    spent in Python, as reported under "Show Code"). To do this, add the
@@ -94,8 +97,12 @@ the profiling machine:
 
       pstats-python-profiler 1
 
-Caveats
-~~~~~~~
+   To access this view, in the Frame strip chart, double-click the App collector
+   on the left side, then double-click the Python collector. Then, you can drill
+   down further into the packages, modules, classes and functions.
+
+Profiling GPU Time
+~~~~~~~~~~~~~~~~~~
 
 OpenGL is asynchronous, which means that function calls aren't guaranteed to
 execute right away. This can make performance analysis of OpenGL operations