qiboteam / quantum-search-scaled-hash-preimages

Code for a quantum implementation of Grover oracles for finding scaled hash function preimages.

Quantum Search for Scaled Hash Function Preimages

This repository contains code to reproduce the result presented in the paper: Quantum Search for Scaled Hash Function Preimages.

The scripts presented use Qibo, a Python library for classical simulation of quantum circuits. In order to properly run the scripts, please start by installing Qibo to your Python enviroment using:

pip install qibo

Run the main.py file to find a preimage for hash value 10100011 without knowing the total amount of preimages. This code is based on the Toy Sponge Hash implementation described in the paper.

The following arguments can alter the program output:

• --hash (int) hash value whose preimages will be found. Has to be represented by less than 8 bits, and is 163 by default.
• --bits (int) minimum amount of bits to use when defining the hash value. Default value is 7.
• --collisions (int) if known, the number of preimages of target hash, to perform direct Grover over the iterative method. Hash 163 has 2 collisions, but is set to None by default.
• --noise perform a noisy simulation with Pauli errors after every gate and return a graph of success probability for increasing error.

The program returns:

• The target hash value used in the program.
• If as successful solution has been found.
  • If number of collisions is not given: a preimage of the function.
  • If number of collisions is given: all preimages of the function.
• Total Grover iterations taken to find a primage.

If --noise was given:

• grover_bitphase.png image detailing the success probability of runing the full, or half Grover algorithm under increasing amounts of noise.

Circuit specifications and gate counts for a single Grover step for both toy models and real implementations of the Sponge Hash construction as well as the Blake2 hash construction are found in the circuit-specifications folder.