I've been working on a new Big Integer library. Is anyone interested in testing it out?
I've tried to make it...
1 easy to use - minimal changes required to existing code
2 reasonably fast (EDIT: with latest optimisations, is now faster then the FNX library)
3 pure Pascal code, no assembly and no C, so portable and reliable
4 runs on 32 bit CPUs and 64 bit CPUs
5 small footprint, small code size
6 will actually compile straight out of the box with no messing about
7 unlicensed/public domain, I'm happy to give it away free of any restrictions
Why did I make this? I could not find any Big Integer library that met all the above criteria. The FNX library is very good, but is 64 bit only. EDIT: I've recently been looking at the MRArith library; it's very fast, but is rather difficult to use and requires substantial changes to existing code.
Disadvantages...
1 only whole integers, no support for other number types
2 only has basic arithmetic, logical and formatting operations
3 does not have extensive maths operations
4 does not have dynamically sized variables, only fixed size, but this can be changed at compiled time with a define
5 relies on Delphi mode for implicit operator overloading
I would welcome all criticism & comments, from anyone who has an interest in programming with big integers. I would especially welcome suggestions from speed freaks on how to make the code run faster
Thanks.