Skip to main content

GANSU2

PyPI PyPI Downloads

GANSU2 (GPU Accelerated Numerical Simulation Utility) is a GPU-native quantum chemistry package: Hartree-Fock, Kohn-Sham DFT, post-HF (MP2, CCSD, CCSD(T), DLPNO-CCSD, DMET, FCI), excited states (CIS, ADC(2), EOM-CCSD, STEOM-CCSD), analytic gradients / Hessians and geometry optimization, all in double precision on NVIDIA GPUs. It is distributed as a small pip wheel plus a licensed shared library that is fetched on first use.

Full documentation: https://gansu2.github.io/

Install

pip install gansu2

No CUDA toolkit, no compiler, no cmake on your side. The wheel is a thin Python wrapper; the CUDA 12 runtime libraries come in as pip dependencies (nvidia-cublas-cu12, nvidia-cusolver-cu12, ...). On first use GANSU2 downloads the GPU shared library libgansu2.so and the gansu2 command-line tool from the matching release, verifies them against their published SHA-256, and caches them under ~/.cache/gansu2/<version>/.

Requirements

  • Linux x86_64 (manylinux_2_28: Ubuntu 20.04+, Debian 10+, RHEL 8+ and equivalents)
  • Python 3.8+
  • NVIDIA driver compatible with the CUDA 12.x runtime (>= 525.60.13)
  • NVIDIA GPU with Compute Capability 8.0+ (Ampere / Ada / Hopper / Blackwell)

Quick test

import gansu2

gansu2.init()
m = gansu2.Molecule("H2O.xyz", basis="sto-3g")
print(f"RHF energy: {m.run(method='RHF').total_energy:.6f} Ha")
gansu2.finalize()

Or from the command line (the gansu2 command is installed by the wheel):

gansu2 -x H2O.xyz -g sto-3g                                   # RHF
gansu2 -x H2O.xyz -g cc-pvdz --post_hf_method ccsd            # CCSD
gansu2 -x H2O.xyz -g cc-pvdz --method rks --functional pbe    # DFT (RKS / PBE)
gansu2 -x H2O.xyz -g cc-pvdz -r optimize                      # geometry optimization
gansu2 --list-basis                                           # bundled basis sets

License key

Small calculations run without any license. Larger ones need a free license key, which grants full, unlimited use (commercial use included):

  1. gansu2-license -s prints an 8-character sign-up code.
  2. Create an account at the User Portal, enter the sign-up code, and press Start. Your key is emailed to you.
  3. gansu2-license -k <YOUR-KEY> -a registers the key on the machine.

The key is a floating license: it is not tied to a machine. See the installation guide and the Terms and Conditions.

Offline / air-gapped use

Download the wheel and the matching libgansu2.so from the release page on a connected machine, then on the target machine:

pip install ./gansu2-<version>-py3-none-manylinux_2_28_x86_64.whl
export GANSU2_LIB=/path/to/libgansu2-<version>-linux-x86_64.so
python -c "import gansu2; gansu2.init()"

GANSU2_LIB makes the loader use that file instead of downloading.

Documentation

Acknowledgements and license

GANSU2 is developed by Hiroshima University and Fujitsu Limited. It is derived from the open-source GANSU project (BSD-3-Clause, https://github.com/Yasuaki-Ito/GANSU); the BSD-3-Clause notice of GANSU and the third-party components are listed in the NOTICE file shipped with the wheel. The GANSU2 binaries themselves are not BSD-licensed: their use is governed by the Terms and Conditions, shipped as the wheel's LICENSE file.

Release files for gansu2 2026.9.25

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Built distribution (wheel)

Table of built distributions (wheels) for gansu2 2026.9.25
File Interpreter ABI Platform
gansu2-2026.9.25-py3-none-manylinux_2_28_x86_64.whl Python 3 none Linux glibc 2.28+ x86-64 Details

Release files / gansu2-2026.9.25-py3-none-manylinux_2_28_x86_64.whl

Download URL gansu2-2026.9.25-py3-none-manylinux_2_28_x86_64.whl
Size 6.6 MB
Tags Linux glibc 2.28+ x86-64 Python 3
SHA-256 checksum
How to use checksums
ee72b3bd8e396d592a5a655dc3d51c9119017e654bc5db1b603fa5daea63cf0f
BLAKE2b-256 checksum
How to use checksums
f68c8f41628781707614124c2e9a5184d0846a5f4acab60c22c5d07a108ce0ff
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 25, 2026.

Transparency log

Release history Release notifications | RSS feed

This release

2026.9.25 This release

1 release file

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page