Need help with taichi_elements?
Click the “chat” button below for chat support from the developer who created it, or find similar developers for support.

About the developer

263 Stars 47 Forks MIT License 249 Commits 17 Opened issues


High-performance multi-material continuum physics engine in Taichi

Services available


Need anything else?

Contributors list

Taichi Elements Build Status Code Coverage

Taichi elements is a high-performance multi-material continuum physics engine (work in progress). Features:

  • Cross-platform: Windows, Linux, and OS X
  • Supports multi-threaded CPUs and massively parallel GPUs
  • Supports multiple materials, including water, elastic objects, snow, and sand
  • Supports (virtually) infinitely large simulation domains
  • Supports sparse grids
  • Highly efficient and scalable, especially on GPUs

in Python

  • Install taichi with
    python3 -m pip install taichi
  • Execute
    to download model files used by the demos
  • Execute
    python3 demo/
    and you will see

  • Execute
    and you will see a 3D simulation visualized in 2D
  • Execute
    python3 demo/
    to get a visual understanding of Taichi sparse computation

To simulate and render an example 3D scene with Python

  • Make sure you have a modern NVIDIA GPU (e.g. GTX 1080)
  • Execute
    to download model files
  • Run
    python3 demo/
    (wait for at least 10 frames)
    • A binary particle folder with a timestamp in its time (e.g.,
      ) will be created under the current folder.
  • Example:
python3 engine/ \
     -i ./path/to/particles \
     -b 0 -e 400 -s 1 \
     -o ./path/output \
     --gpu-memory 20 \
     -M 460 \
     --shutter-time 0.0 \
     -r 128
  • Images are in the
    folder. For example, 100 million MPM particles simulated in 8 hours on a V100 GPU:

[Watch on YouTube]

  • Here is a 460-million-particle simulation on a GPU with
    GB memory. Each particle takes 40 bytes, thanks to quantization.

in Blender

Installing Taichi for Blender

(Not sure if it is the standard approach, but it works for now.) - Find the Python3 executable bundled in Blender. Open a console in Blender and type in

 import sys
The output looks like
, which means python3 is located at
- Install pip using that Python executable - Install Taichi:
./python3.7m -m pip install --upgrade taichi
(Note: as of Oct 8 2021, Taichi version is
. Please use the latest version.)

Installing taichi_elements (experimental) for Blender

  • Set the environment variable
    , e.g.
  • Go to
  • Execute
    to install the addon
    • If you are doing development and wish to keep refreshing the installed addon, add an argument
  • Restart Blender to reload the addon

We use cookies. If you continue to browse the site, you agree to the use of cookies. For more information on our use of cookies please see our Privacy Policy.