OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
ZK & Privacy

Heongpu

Heongpu, published as HEonGPU, is an open-source C++ and CUDA library for accelerating fully homomorphic encryption computations on GPUs. Its multi-stream architecture supports concurrent processing and is designed to reduce CPU–GPU transfer overhead. The library lists support for BFV, CKKS and TFHE schemes.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Heongpu — OpenViST project profile in ZK & Privacy.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Consider Heongpu when an application needs fully homomorphic encryption computations and can use a CUDA GPU; its concurrent, multi-stream design is aimed at reducing compute and transfer overhead for larger workloads.

Who it is for

  • Developers building applications that compute on encrypted data.
  • Teams implementing privacy-preserving analytics or other large FHE workloads on CUDA GPUs.
  • C++ developers who want a library interface that encapsulates CUDA kernels.

Documented uses

  • Run BFV, CKKS or TFHE computations with GPU parallelism.
  • Use listed bootstrapping and multiparty or collective bootstrapping capabilities in supported FHE workloads.
  • Build a GPU-backed component for encrypted computation rather than adopting a prebuilt end-user service.
Before you choose

Requirements & limits

The project is listed as open source under the Apache-2.0 license.

  • The documented build requires CUDA Toolkit 12.0 or later and CMake 3.30.4 or later.
  • Default CUDA architecture targets are 80, 86, 89 and 90; the project documents an override option.
  • The README lists BGV as “Very Soon,” while identifying BFV, CKKS and TFHE as supported schemes.
  • Published benchmark figures specify an NVIDIA RTX 4090 and a particular parameter set, so results are hardware- and setup-specific.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Build and install Heongpu from source

Configure, compile and install the library on a CUDA-equipped Linux system.

Before you start

  • CMake 3.30.4 or later
  • GCC, GMP, CUDA Toolkit 12.0 or later, OpenSSL 1.1.0 or later, and ZLIB
  • A CUDA architecture target; documented defaults are 80, 86, 89 and 90

Steps

  1. Configure the source with cmake -S . -B build -D CMAKE_BUILD_TYPE=Release.

  2. Compile it with cmake --build ./build/.

  3. Install it with sudo cmake --install build.

Expected outcome

The documented process installs the Heongpu library and its headers on the system.

Guide limitations

  • The instructions describe building and installing the library, not implementing or running a particular encrypted computation.
  • Override the CUDA architecture setting when the target GPU requires a value other than the documented defaults.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
14 October 2026
Original website title
enCRYPTON-labs/HEonGPU · Official homepage title · retained in its source language
How we review sourcesSuggest a correction