PDHCG is an open-source first-order solver based on primal-dual hybrid gradient methods for large-scale convex quadratic and quadratic conic programming. It supports CPU and CUDA backends, including distributed multi-GPU execution.
Supported models include linear programs, sparse and structured low-rank quadratic objectives, and SOC, rotated SOC, exponential, power, and PSD cones.
PDHCG also integrates with CVXPY and qpsolvers.
The full Hessian
| Topic | Reference |
|---|---|
| Requirements and installation | Installation |
| Backends, threads, and multi-GPU execution | Devices |
| Python getting started | Quick start |
| Python API | Model, Parameters |
| Native API | C API |
| Command-line options | CLI |
| Modeling examples | Examples |
| Upgrading | Migration guide |
If you use PDHCG in research, see the citation and BibTeX.
This solver is built upon the infrastructure of cuPDLPx (originally developed by Haihao Lu). We gratefully acknowledge this project for providing the high-performance CUDA-C framework for Linear Programming (LP) that serves as the foundation for this QP solver.
Copyright 2024-2026 Hongpei Li, Haihao Lu.
Licensed under the Apache License, Version 2.0. See the LICENSE file for details.