apfloat v. 1.20

Last updated: July 4th, 1996

Download the package here:

The common source files version 1.20 required to compile the program.

In addition you need one of these depending on your system:

32-bit version for Linux, optimized for 386/486.

32-bit version for Linux, optimized for the Pentium.

32-bit version for djgpp v.1/v.2, optimized for 386/486.

32-bit version for djgpp v.1, optimized for the Pentium.

32-bit version for djgpp v.2, optimized for the Pentium.

32-bit version for bcc32, not optimized at all.

32-bit version for general UNIX gcc, allows slightly greater precision.

64-bit version for general UNIX gcc.

64-bit version for DEC Alpha gcc.

64-bit version for MIPS R4000 (and greater) processors' gcc. NOTE: This version is untested. If you have a chance to test it, please let me know.

floating-point version (doubles) for general UNIX gcc (or any compiler compatible with the ANSI C++ draft).

floating-point version (floats) for general UNIX gcc (or any compiler compatible with the ANSI C++ draft).

Unzipping instructions:

Then unzip the version-specific package (like 32_unix.zip) with the -o switch to overwrite the files that are different from the common package.

Unix and Linux users must unzip the zip files with the -a and -L options.

Read the files LEGAL and README.1ST in the source package, especially if you don't read the full documentation.

If you use MS-DOS and don't have djgpp, get it here .