Skip to content

applied-crypto-lab/floating_point_summation

Repository files navigation

Secure Floating-Point Numbers Summation

This is our implementation of the protocols and from our PETS 2023 paper "Secure and Accurate Summation of Many Floating-Point Numbers."

The full implementation will be updated by the end of March 2023.

Requirements

  1. gcc 9 or newer
  2. OpenSSL v1.1.1

Docker

The authors currently don't have access to a dev envirments that supports Docker. If a docker is preferred, please use the one provided by (https://github.com/anbaccar/RSS_ring_ppml) as we have the same system/library requirments.

Compile

For compile the binary:

make

Run

For rerun the experiments listed in the paper in a signle machine:

  1. Open a terminal, run bash run.sh 3 Y
  2. Open another terminal, run bash run.sh 2 Y
  3. Open one more terminal, run bash run.sh 1 Y

For running experiments in different machines, please specify the IP addess and port of each machine in runtime-config and run the above command from party 3 (X = 3) to party 1 (X = 1) on these machines sequentially.

Y represents the setting we used in the paper:

Y=0: Single & w = 16

Y=1: Single & w = 32

Y=2: Double & w = 16

Y=3: Double & w = 32

For running the experiments using different machines, please update the IP address of each machine in runtime-config accordingly.

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages