Ripple: Accelerating Programmable Bootstraps for FHE with Wavelet Approximations

How to cite this work

Ripple will appear in the proceedings of the 27th Information Security Conference (ISC), 2024. The preprint can be accessed here; you can cite this work as follows:

@InProceedings{ISC:GUMVT24,
  author =      "Charles Gouert and
                 Mehmet Ugurbil and
                 Dimitris Mouris and
                 Miguel de Vega and
                 Nektarios Georgios Tsoutsos",
  title =       "{Ripple: Accelerating Programmable Bootstraps for FHE with Wavelet Approximations}",
  booktitle =   "{International Conference on Information Security}",
  year =        2024,
  publisher =   "Springer",
  pages =       "1--20",
}

Building & Running

This repository comprises multiple binaries located at the src directory. For each application, we have different variants, namely: plaintext, Haar DWT, Biorthogonal DWT, and quantization.

Building

For x86_64-based machines running Unix-like OSes:

❯❯ cargo b --release --features x86

For Apple Silicon or aarch64-based machines running Unix-like OSes:

❯❯ cargo b --release --features aarch64

Running

Example:

❯❯ cargo run --release --bin lr_haar

Disclaimer

This is software for a research prototype and not production-ready code. This repository builds upon TFHE-rs and DWT.