arnaucube
arnaucube/poulpyPublic Self-hosted repository
updated README
Pro7ech jeanphilippe.bossuat@gmail.com
Mon, 25 Aug 2025 12:33:53 +0200
commit

99b9e3e10e41840a1881f35b621734a4de80cc95

parent

39062462bcd1abae60370351d603f2f76e7fe1db

1 files changed, 1 insertions(+), 1 deletions(-)

jump to
M README.md → README.md

@@ -7,7 +7,7 @@ </p>

[![CI](https://github.com/phantomzone-org/poulpy/actions/workflows/ci.yml/badge.svg?branch=main)](https://github.com/phantomzone-org/poulpy/actions/workflows/ci.yml) -**Poulpy** is a fast & modular FHE library that implements Ring-Learning-With-Errors based homomorphic encryption. It adopts the bivariate polynomial representation proposed in [Revisiting Key Decomposition Techniques for FHE: Simpler, Faster and More Generic](https://eprint.iacr.org/2023/771). In addition to simpler and more efficient arithmetic than the residue number system (RNS), this representation provides a common plaintext space for all schemes and native bridges between any two schemes. Poulpy also decouples the schemes implementations from the polynomial arithmetic backend by being built from the ground up around a **hardware abstraction layer** that closely matches the API of [spqlios-arithmetic](https://github.com/tfhe/spqlios-arithmetic). Leveraging HAL library users can develop applications while being generic over the backend and choose any backend of choice at runtime. +**Poulpy** is a **fast & modular** FHE library that implements Ring-Learning-With-Errors based homomorphic encryption over the Torus. It adopts the bivariate polynomial representation proposed in [Revisiting Key Decomposition Techniques for FHE: Simpler, Faster and More Generic](https://eprint.iacr.org/2023/771) to represent Torus polynomials. In addition to simpler and more efficient arithmetic than the residue number system (RNS), this representation provides a **common plaintext space** for all schemes and native bridges between any two schemes. Poulpy also decouples the schemes implementations from the polynomial arithmetic backend by being built from the ground up around a **hardware abstraction layer** that closely matches the API of [spqlios-arithmetic](https://github.com/tfhe/spqlios-arithmetic). Leveraging HAL library users can develop applications while being generic over the backend and choose any backend of choice at runtime. <p align="center"> <img src="docs/lib_diagram.png" />