Ada 2022 introduces support for Big Numbers into the standard library, however the best available implementation of this standard (GNAT >= 12.1.0) suffers from slow performance as numbers grow larger.
In its current implementation, GNAT enforces a restriction on the length of Big_Integers, preventing them from exceeding 200 32-bit digits. If this restriction is removed, however, it becomes apparent that Big_Integers is not currently optimised for the case of exceedingly large numbers, and as a result tests (like the enclosed pi_digits benchmark) run very slowly, if at all.
To remedy this problem, this repository contains a small (hopefully) standards-compliant
wrapper for Ada.Numerics.Big_Numbers.Big_Integers that uses libgmp behind the scenes to
drastically improve performance. A direct mapping to a subset of GMP's API is also included
in gmp.ads, however this is primarily an implementation detail.
I haven't given it much thought yet, but if there is interest, I would be willing to write a compliant implementation of Big_Reals as well that uses GMP and then submit it to GNAT upstream for consideration. At the very least, I hope this proves useful for people using Ada for mathematics.
This library is still in development, and there may be bugs. I have done my best to ensure it functions as required by the Ada Reference Manual (A.5.5), however since Ada 2022 is still a draft things may change, or I may have simply made a mistake.
This library is licensed under the LGPL Version 3 or Higher.