axes

axes is a utility-library for Physical Simulation in Computer Graphics.

System Requirement:

gcc>=12, or
clang>=16.0, or
MSVC: VS2022

axes use C++20 code standard.

What's in

A basic library for general purpose CG physical simulation, containing
1. Math library using Eigen;
2. Utilities library, such as abseil's Status, Logging, Flags;
3. A tiny optimization library
4. Basic geometry functionalities
Graphics for visulization. which supports
1. Trimesh/line/point visualization (with phong shading)
2. ImGUI integration.
3. and something more...
A trival node system, for fast prototype development, with moderate performance and high flexibility.

System Dependencies

These settings are tested on:

MacOS-latest (m1)
Windows 11 latest, with Visual Studio 2022 latest (x64)
Arch Linux latest (x64)

Common Dependencies

Refer to axdeps, and run the install script.

For Unix-like System

export SDK_PATH="path/to/axes"
export BUILD_TYPE=Release
path/to/axdeps/build.sh

For Windows

$env:SDK_PATH="path/to/axes"
$env:BUILD_TYPE="Release"
path/to/axdeps/build.ps1

(Optional) Python binding, please refer to requirements_cpu.txt:

pip install -r requirements_cpu.txt

for additional GPU dependencies, please refer to requirements_gpu.txt:

pip install -r requirements_gpu.txt

The prefered Deep Learning Framework is PyTorch, version=2.3.0.

Platform Dependent Dependencies

For mac users, because we do not support Metal, nothing is required.

Windows:

Install Visual Studio 2022 with C++ Development Kit;
(Optional) CUDA Toolkit from NVIDIA.

Although the library is able to compile on Windows, the performance of both compilation and runtime is not guaranteed.

Arch Linux:

yay -S libcap gperf libsystemd libmount liblzma xz dbus
# If you prefer a clang compiler
yay -S clang15
# (Optional) CUDA Toolkit, However CUDA is always the latest from arch linux install.
yay -S cuda

For Ubuntu:

# CPU Version.
sudo apt install base-devel

CUDA is also available for boost up the performance. Refer to the CUDA official installation guide for more information.

My CUDA version: 12.4.131

Performance Illustration