Number-Theory's end goal is to be among the pre-eminent elementary number theory libraries of any language
While it is quite fast for machine-sized words (possibly the fastest in any language). It slacks in arbitrary precision arithmetic, with exponentiation in particular being about 60% slower than num-bigint.
In order of priority
- Faster exponentiation. Look into preallocation.
- Specialized Squaring operation
- Faster euclidean division, possibly recursive division
- Fast base-2 strong Fermat check
- Faster factorization needed, either in the form of elliptic curve factorization and/or quadratic sieve. Eventually the GNFS is planned to be implemented.
- Schonhage-Strassen multiplication using Fermat Transforms
Number-Theory's end goal is to be among the pre-eminent elementary number theory libraries of any language
While it is quite fast for machine-sized words (possibly the fastest in any language). It slacks in arbitrary precision arithmetic, with exponentiation in particular being about 60% slower than num-bigint.
In order of priority