Rudy Velthuis is the orignal author of this library. He unfortunately passed away a while back. In an effort to keep this valuable library alive we've done the following:
- Pulled changes from all the other branches into this repository
- Created Wiki articles based on his original documentation (with some updates)
- Opened up for issues
- Fixed some other issues and updated for Delphi 11.3
Pull requests welcome. If you want to help maintain this library, please let us know.
These are implementations of the multi-precision BigInteger, BigDecimal and BigRational types, built from scratch by Rudy Velthuis and now maintained by TurboPack.
BigInteger is a multi-precision integer. Its size is only limited by available memory.
BigInteger is built for ease of use, speed and reliability. It is written in plain Object Pascal and x86-32/x86-64 assembler, but every assembler function has a so called "pure Pascal" equivalent as well. It is modelled after the BigInteger type in .NET, but is far more optimized than that and provides an interface that is more in line with Delphi. It uses higher level algorithms like Burnikel-Ziegler, Karatsuba, Toom-Cook, etc. to make things fast even for very large integers. It offers overloaded operators and all the usual functions. More information can be found on the BigIntegers unit wiki page.
BigDecimal is a multi-precision decimal floating point type. It can have an almost unlimited precision.
BigDecimal is equally built for ease of use and reliability. It builds on top of BigInteger: the internal representation is a BigInteger for the significant digits, and a scale to indicate the decimals. It also offers overloaded operators and all the usual functions. This is modelled after the BigDecimal type in Java, but the interface is more in line with Delphi. More information about this type can be found on the BigDecimals unit wiki page.
A type that holds a number as fraction (ratio) of two BigIntegers, a numerator and a denominator, i.e. 1/7 or 100/3.
This type is very good at simple arithmetic (+, -, *, /), since it doesn't lose precision or need any rounding. Still a work in progress...
The newest version of BigIntegers has additional overloaded operators and additional constructors that are compatible with C++Builder. So now you simply include:
#include "Velthuis.BigIntegers.hpp"
and then you can do things like:
BigInteger a = 17;
BigInteger b = "123";
BigInteger c = a + b;
RudysBigNumbers
Source --- Sources for units and for bases.inc
Tests
BigDecimals /... --- Sources for DUnit tests for BigDecimals
BigIntegers /... --- Sources for DUnit tests for BigIntegers
BigRationals /... --- Sources for DUnit tests for BigRationals
Visualizers --- Sources for IDE debug visualizer DLL and packages for BigInteger and BigDecimal
Samples --- A Few samples of Big Math including Pi and Euler's number
The data generators are now in their own repository.