Rug, any way to set max exponent value?

help
1

I am coding an algorithm to compute large Bernoulli numbers. This algorithm requires factorials of very large values in floating point. The algorithm works for up to about Bernoulli(44_000_000) but fails for 45_000_000. I have found that the issue is the conversion of Integer to Float fails if the conversion results in an exponent of 2^30-1 or greater (results in an infinity). This does not look like a max precision problem but a max exponent problem. I am fairly new to Rust and do not know how to report issues to crate owners. Some test code is below:

use rug::{Float, Complete};
use rug::{Integer};
use rug::integer::IntegerExt64;

fn main() {
    let fact = Integer::factorial(44_000_000).complete();
    let sb = fact.significant_bits_64();
    let fact_float = Float::with_val_64(sb, fact);
    println!("fact_float 44m finite? {}", fact_float.is_finite());
    let fact = Integer::factorial(45_000_000).complete();
    let sb = fact.significant_bits_64();
    let fact_float = Float::with_val_64(sb, fact);
    println!("fact_float 45m finite? {}", fact_float.is_finite());
}

I have also tried floating point factorial function but it so slow that I can't test large values, but they probably will fail any way.

2

After do further looking it looks like the max exponent is limited to i32. Shame since I have several algorithms that need larger exponents.

3

Does it really? Are you sure you can't phrase it in terms of https://docs.rs/rug/latest/rug/struct.Float.html#method.ln_gamma or something?

4

Using gamma to compute factorial would still result in the same large numbers.

5

There's a reason I said logarithm of gamma.

6

I realize that log gamma would generate a smaller number but what I need is the large number for the Bernoulli calculation. The algorithm chosen produces a fraction (num/den). Bernoulli numerators grow dramatically as the Bernoulli numbers increase. Algorithm used is a Zeta and Staudt_Claussen algorithm that computes the numerator in floating point and truncates the final answer to an integer. The factorial is one component of the numerator computation.

7

For the record, the typical way to report an issue in a crate is:

  • Go to the crate's page on crates.io (in this case, https://crates.io/crates/rug)
  • Click the link under "Repository" on the right-hand side to go to the source repository for the crate on GitHub, GitLab, Codeberg, or the like (in this case, https://gitlab.com/tspiteri/rug)
  • Create an issue in the web interface there. For GitLab, issues are listed under "Plan" > "Work Items" in the menu on the left-hand side.
8

Thanks, jwodder, found the issue page for rug and made a request.