[Magictour] Re: pandiagonal

> > Since no closed magic tour on 12x12 is known to this date, I suggest
that
> we
> > put our efforts on enumerating all semi-regular closed 12x12 magic
tours.
>
> It is to inform you that hundreds of closed magic tours on 12x12 is known.
> Just go through 'The Games and Puzzles Journal' # 24.
>
Are these semi-regular ?

> I am interesred in the programme. Kindly send it to me.
>
I decided to release the source-code, so anybody could configure the program
to fit his needs.
Attached, the ZIPped source files.

How does it work ?
First, configure the #defines in GenASM2.C (the program is currently
configured to search for regular closed magic 12x12 knight tours, with 4x4
magic subsquares).

Then compile and run the program.
2 files will appear: ASM.H and ALL.ASM (59Mb with the configuration above).

Move these files in the Distributed folder.
Edit ASM.H and move the extern at the beginning of the source file, to
remove the warnings.

Then, open Distributed.dsp with Microsoft Visual Studio, and build the
project (Release build is recommended, but Debug should work).
With the current settings, you'll get a 13Mb EXEcutable.

The generator only searches for start-end tours, because of its internal
code generation.

The generator can be configured with the following options:

#define N
the size of the square board.

#define TORUS
to force search on torus boards, if set to 1

#define GRECOLATIN
to force graeco-latin tours

#define MAGIC
0=enumerates all tours
1=computes only magic squares, with lines and columns matching the magic
constant.
2=computes magic squares, with 4x4 subsquares matching a constant value.

#define FORCE_CLOSED
0=all tours
1=only closed tours

#define SEMIREGULAR
0=no constraint
1=semi-regular constraint

#define REGULAR
0=no constraint
1=regular constraint (can be combined with SEMIREGULAR constraint).

As Gunter noticed, perhaps the start-end approach is not the best one on the
12x12 search, but the code cannot be easily modified, and I have no time for
a full rewrite + tuning.

Anyway, the current program is able to find all 8x8 semi-regular tours in 3
minutes, all 8x8 regular in less time, and all 8x8 regular with 4x4
subsquares in a few seconds.

I let you test the different settings, and will be glad to help if more
constraints need to be added, or if a magic search is decided.
For the moment, I have no more time to put into this project, since I need
to write a program into dotNet for job reasons.

JC

Magictour mailing list
To unsubscribe: mailto:magictour-request@ml.free.fr?subject=unsubscribe
To mail to the mailing list: mailto:magictour@ml.free.fr