Fractint

Q21a: What is Fractint?
A21a: Fractint is a very popular freeware (not public domain) fractal
generator.  There are DOS, Windows, OS/2, and Unix/X versions.  The DOS
version is the original version, and is the most up-to-date.

Please note: sci.fractals is not a product support newsgroup for Fractint.
Bugs in Fractint/Xfractint should usually go to the authors rather than being
posted.

Fractint is on many ftp sites.  For example:
DOS: ftp from wuarchive.wustl.edu:/mirrors/msdos/graphics [128.252.135.4].
    The source is in the file frasr182.zip.  The executable is in the file
    frain182.zip.  (The suffix 182 will change as new versions are released.)
    Fractint is available on Compuserve: GO GRAPHDEV and look for FRAINT.EXE
    and FRASRC.EXE in LIB 4.
There is a collection of map, parameter, etc. files for Fractint, called
    FracXtra.  Ftp from wuarchive.wustl.edu:/pub/MSDOS_UPLOADS/graphics.  File
    is fracxtr5.zip.
Windows: ftp to wuarchive.wustl.edu:/mirrors/msdos/window3 .  The source is in
    the file winsr173.zip.  The executable is in the file winfr173.zip.
OS/2: available on Compuserve in its GRAPHDEV forum.  The files are PM*.ZIP.
    These files are also available by ftp:
    ftp-os2.nmsu.edu:/pub/os2/2.0/graphics in pmfra2.zip.
Unix: ftp to sprite.berkeley.edu [128.32.150.27].  The source is in the file
    xfract203.shar.Z.  Note: sprite is an unreliable machine; if you can't
    connect to it, try again in a few hours, or try hijack.berkeley.edu.
    Xfractint is also available in LIB 4 of Compuserve's GO GRAPHDEV forum in
    XFRACT.ZIP.
Macintosh: there is no Macintosh version of Fractint, although there are
    several people working on a port. It is possible to run Fractint on the
    Macintosh if you use Insignia Software's SoftAT, which is a PC AT
    emulator.

For European users, these files are available from ftp.uni-koeln.de.  If you
can't use ftp, see the mail server information below.

Q21b: How does Fractint achieve its speed?
A21b: Fractint's speed (such as it is) is due to a combination of:

1. Using fixed point math rather than floating point where possible (huge
improvement for non-coprocessor machine, small for 486's).

2. Exploiting symmetry of the fractal.

3. Detecting nearly repeating orbits, avoid useless iteration (e.g. repeatedly
iterating 0^2+0 etc. etc.).

4. Reducing computation by guessing solid areas (especially the "lake" area).

5. Using hand-coded assembler in many places.

6. Obtaining both sin and cos from one 387 math coprocessor instruction.

7. Using good direct memory graphics writing in 256-color modes.

The first four are probably the most important. Some of these introduce
errors, usually quite acceptable.