mirror of
https://github.com/ManimCommunity/manim.git
synced 2026-06-22 10:01:47 +00:00
859c1793f6* Guide on performance * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Update docs/source/contributing/performance.rst Co-authored-by: Benjamin Hackl <devel@benjamin-hackl.at> * Try to get code block to render * Update docs/source/contributing/performance.rst Co-authored-by: Benjamin Hackl <devel@benjamin-hackl.at> * try fix Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com> Co-authored-by: Benjamin Hackl <devel@benjamin-hackl.at>
81 lines
2.3 KiB
ReStructuredText
81 lines
2.3 KiB
ReStructuredText
=====================
|
|
Improving performance
|
|
=====================
|
|
|
|
One of Manim's main flaws as an animation library is its slow performance.
|
|
As of time of writing (January 2022), the library is still very unoptimized.
|
|
As such, we highly encourage contributors to help out in optimizing the code.
|
|
|
|
Profiling
|
|
=========
|
|
|
|
Before the library can be optimized, we first need to identify the bottlenecks
|
|
in performance via profiling. There are numerous Python profilers available for
|
|
this purpose; some examples include cProfile and Scalene.
|
|
|
|
Running an animation as a script
|
|
--------------------------------
|
|
|
|
Most instructions for profilers assume you can run the python file directly as a
|
|
script from the command line. While Manim animations are usually run from the
|
|
command-line, we can run them as scripts by adding something like the following
|
|
to the bottom of the file:
|
|
|
|
.. code-block:: python
|
|
|
|
with tempconfig({"quality": "medium_quality", "disable_caching": True}):
|
|
scene = SceneName()
|
|
scene.render()
|
|
|
|
Where ``SceneName`` is the name of the scene you want to run. You can then run the
|
|
file directly, and can thus follow the instructions for most profilers.
|
|
|
|
An example: profiling with cProfile and SnakeViz
|
|
-------------------------------------------------
|
|
|
|
Install SnakeViz:
|
|
|
|
.. code-block:: bash
|
|
|
|
pip install snakeviz
|
|
|
|
cProfile is included with in Python's standard library and does not need to be installed.
|
|
|
|
Suppose we want to profile ``SquareToCircle``. Then we add and save the following code
|
|
to ``square_to_circle.py``:
|
|
|
|
.. code-block:: python
|
|
|
|
from manim import *
|
|
|
|
|
|
class SquareToCircle(Scene):
|
|
def construct(self):
|
|
s = Square()
|
|
c = Circle()
|
|
self.add(s)
|
|
self.play(Transform(s, c))
|
|
|
|
|
|
with tempconfig({"quality": "medium_quality", "disable_caching": True}):
|
|
scene = SquareToCircle()
|
|
scene.render()
|
|
|
|
Now run the following in the terminal:
|
|
|
|
.. code-block:: bash
|
|
|
|
python -m cProfile -o square_to_circle.txt square_to_circle.py
|
|
|
|
This will create a file called ``square_to_circle.txt``.
|
|
|
|
Now, we can run SnakeViz on the profile file:
|
|
|
|
.. code-block:: bash
|
|
|
|
snakeviz square_to_circle.txt
|
|
|
|
A browser window or tab will open with a visualization of the profile, which should
|
|
look something like this:
|
|
|
|
.. image:: /_static/snakeviz.png
|