Bug#775867: ITP: libtomsfastmath -- Fast multiple-precision integer library.
Owner: Sebastian Andrzej Siewior <email@example.com>
* Package name : libtomsfastmath0
Version : 0.12
Upstream Author : Tom St Denis
* URL : http://www.libtom.net/?page=features&newsitems=5&whatfile=tfm
* License : Public Domain
Programming Lang: C
Description : TomsFastMath is a fast large integer arithmetic library.
It is a port of LibTomMath with optional support for inline assembler multipliers.
It has the following features:
- Public domain, open source and freely accessible
- Builds with GCC [preferably v3.3.x or higher] on any 32-bit/64-bit box
- Includes ASM optimizations (inline) for
- x86-32 (e.g. i386 and SSE2)
- x86-64 (e.g. AMD64)
- ARM (ARMv4 and higher)
- PPC32 (e.g. G3/G4)
- PPC64 (e.g. G5)
- AVR32 (Atmel 32-bit RISC processor)
- MIPS (32-bit)
- Includes fairly complete API (provides most of what LibTomMath provides)
- Uses similar API as LibTomMath makes porting easy
- Competitive speeds (and getting higher).
- Faster than OpenSSL on the AMD64
- Still room for improvements
- No heap usage (malloc/calloc/free)
- Uses only inline asm, simplifies build process, simple to add new ports
This library is already used by a few application in the Debian archive
which use a copy of this library within their source tree. I noticed
this in the clamav package. Debian's code search  reveals other