#elliptic-curve #factorization #modular-arithmetic #128-bit #prime-factors

bin+lib prime_factorization

Prime factorization up to 128 bit integers

5 stable releases

1.0.4 Jun 10, 2023
1.0.3 Nov 19, 2022
1.0.2 Oct 13, 2022
1.0.1 Aug 12, 2022
1.0.0 Aug 10, 2022

#141 in Math

Download history 104/week @ 2023-12-11 126/week @ 2023-12-18 76/week @ 2023-12-25 6/week @ 2024-01-01 3/week @ 2024-01-08 6/week @ 2024-02-12 7/week @ 2024-02-19 116/week @ 2024-02-26 21/week @ 2024-03-04 28/week @ 2024-03-11 23/week @ 2024-03-18 26/week @ 2024-03-25

101 downloads per month
Used in 3 crates

CC0 license

100KB
2K SLoC

Prime factorization

main crate

Program to decompose a natural number N, up to u128::MAX, into a product of its prime factors. Based on the fundamental theorem of arithmetic every natural number larger than one is either a prime itself or can be represented as a product of primes that is unique up to the order of these prime numbers.

The whole factorization algorithm consists of trial division with the first one-thousand primes, Fermat's factorization method and Lenstra elliptic-curve factorization using projective coordinates with Suyama's parametrization. After Fermat's and before advancing to elliptic-curve factorization step, possible primality of the number is checked and this is conducted either with Miller-Rabin or strong Baillie-PSW primality test depending on the magnitude of the number. Latter test is not deterministic in the number range it's used here (up to 128 bits) but there aren't known counterexamples.

Install

To install as a dependency (library target) for some other program, add the following to your Cargo.toml

[dependencies]
prime_factorization = "1.0.4"

For the binary target, run command cargo install prime_factorization and make sure that the installation location is in PATH (i.e., Rust toolchain properly configured).

Use

Use the library as follows

use prime_factorization::Factorization;

// Factorize following semiprime
let num: u128 = 3_746_238_285_234_848_709_827;

let factor_repr = Factorization::run(num);

// Check that the returned factors are correct
assert_eq!(factor_repr.factors, vec![103_979, 36_028_797_018_963_913]);

Notice that all integers from 2 to 2^128 - 1 can be factorized but the used integer type must implement (alongside few others) trait From<u32>.

Sometimes it might be enough to check whether a particular number is a prime

use prime_factorization::Factorization;

let num: u128 = 332_306_998_946_228_968_225_951_765_070_086_139;

// Use the `is_prime` convenience field
assert_eq!(Factorization::run(num).is_prime, true);

If the binary target was installed, CLI can be used as follows

prime_factorization num [-p | --pretty]

where argument num is the mandatory natural number and option -p or --pretty is a print flag which, when given, causes the output to be in the proper factor representation format $$p_1^{k_1} * ... * p_m^{k_m}$$ Without the flag, output only lists all the prime factors from the smallest to largest.

Remarks

  • Elliptic-curve factorization must use OS threads to be efficient. The thread count should be set to a value of at least two and preferably below the number of CPU cores to optimize performance. In terms of performance, lower value (2-5) seems to be the best but large 128 bit semiprimes could be factorized faster with larger thread count based on benchmarking. Thread count can be changed by the MAX_THREADS_ constants in the factor module.

  • Miller-Rabin and Baillie-PSW primality tests are probabilistic but do not contain counterexamples in the number range this program uses. Elliptic-curve factorization uses random initial points on the curves which can cause some deviation to execution times.

License

This program is licensed under the CC0v1.

Dependencies

~1MB
~22K SLoC