Skip to content
RobothecaBeta
Log in

Cupoch

  • Software
  • Perception
  • Library

A library for fast 3D point cloud and robotics computation on NVIDIA GPUs using CUDA, based on Open3D, for tasks like SLAM, collision checking and path planning.

Source on GitHub: github.com/neka-nat/cupoch (opens in a new tab). Facts read at commit 3415b5f (opens in a new tab).

Licenses

Code
MIT (opens in a new tab)

Read by Robotheca from LICENSE (opens in a new tab) (MIT) at commit 3415b5f (opens in a new tab), on 22 Oct 2026.

Commercial use
Yes

What it does

  • Provides GPU-accelerated point cloud processing, including nearest-neighbor search, registration (ICP, Generalized ICP, Fast Global Registration), features and keypoints.
  • Offers visual odometry, Kinect Fusion, stereo matching, occupancy grids and distance transforms for robot perception.
  • Includes collision checking, path finding on graphs and path planning for collision avoidance.
  • Reads and writes several file types (pcd, ply, stl, obj, urdf) and ROS messages.
  • Supports memory pools, managed allocators, an interactive OpenGL/imgui GUI, and DLPack interoperability with PyTorch and CuPy.
  • Can be installed with pip or built from source with CMake, and also builds for AMD GPUs with ROCm when configured with USE_HIP.

Written by Robotheca from the project's README (opens in a new tab) at commit 3415b5f.

Getting started

Difficulty
Advanced
Hardware it needs, as the project states it
NVIDIA GPU with CUDA 12.9 on 64-bit Ubuntu Linux 24.04 (benchmarks run on an Intel Core i7-7700HQ CPU and GTX1070 GPU); AMD GPUs via ROCm are also supported; Jetson Nano is supported with JetPack.

Maintenance

Status
Maintained: a commit in the last 12 months
Newest commit
25 Jun 2026 on main
Latest release
v0.2.11.0 (8 Jun 2024)

Checked on GitHub on 22 Oct 2026.

Links

Robots that can run it

No robot on Robotheca is linked to it yet. Links come from a robot's robot.yaml, from its dependency files read as text, or from Robotheca's editors.

Robotheca lists software as a reference: nothing is hosted or run here, and every fact links to where it was read.