Computing some burrs

Have a project you're currently working on? Post it here!
Post Reply
flawr
Posts: 48
Joined: June 12th, 2012, 8:05 am
Contact:

Computing some burrs

Post by flawr » October 12th, 2012, 7:01 pm

Today I tried to develop a 6 piece burr with a special move, but I dindn't succeed. Insted I ended up in creating a set of about 400 pieces out of all about 800 possible ones. First I let burrtools create all possible pieces, then i started deleting the ones with more than 2-3 internal corners and all others who seem to be difficult to craft. Now I started computing possible burrs with each 6 different of thos 400 remaining pieces, and I guess I have to keep the computer running for two or thee days=) Now the program is running for two and a half hours, it found 8017 possible Solutions and checked almost tree million assemblies. I already found some pretty interesting burrs among a vast amount of rubbish=) When I get to find anything interesting I will post it here. It is the first time I use "brute force" for such a big number of pieces, so I am pretty curious wheater there wil be any interesting results (and wheater the computer or burrtools does crash or not^^) It would be a great feature in burrtools if you could use more than one computer to work on the same problem. Or even if you could use more than one core of multicore processors=P
Have a nice weekend=)
update:
After 4 hours, about 13'000'000 assemblies were checked, I just estimated the total number of assemblies by the following formula: (411 pieces)
411!/(411-6)!*2^5 which is ~1.5*10^17 which means I need some more time...=P
-NO, I now stopped the calculation, but thanks to burrtools I could resume whenever I want-but it is enough for now, because I already found some special burrs. I now realised, that I prefer to design burrs in a "constructive" way, because that "brute force" method is just quite boring, picking some nice ones out of thousends of random ones.... However it is a special experience.

User avatar
Pio2001
Posts: 218
Joined: August 10th, 2011, 9:22 pm
Location: France
Contact:

Re: Computing some burrs

Post by Pio2001 » October 13th, 2012, 5:41 pm

Hello Flawr,
Will you publish your designs somewhere ? I'd like to see them. I could try to solve them with Live cubes. ;)

flawr
Posts: 48
Joined: June 12th, 2012, 8:05 am
Contact:

Re: Computing some burrs

Post by flawr » October 16th, 2012, 8:08 pm

Hello Pio,
thanks for asking! That just motivated me to make a little site for sharing my puzzles.
*drumms and fanfare*
http://asdf.bplaced.net/puzzle/
here it is=) and let me know what you think of the newest one=)

User avatar
Pio2001
Posts: 218
Joined: August 10th, 2011, 9:22 pm
Location: France
Contact:

Re: Computing some burrs

Post by Pio2001 » October 16th, 2012, 9:44 pm

Hi,
27 minutes to assemble the pieces together !
I tried nearly all the wrong assemblies before finding the right one.

It is level 6 isn't it ? I didn't make its Burrtools file.
No ambiguous piece. A handful of wrong assemblies. A small dead end at the start.

The disassembling sequence is quite tortuous.
It is of the same kind as Abad's level 9.2, but without the abiguous piece, with a simpler disassembling sequence, with more false assemblies, and 2 internal corners instead of 5.

Quite a nice design.

flawr
Posts: 48
Joined: June 12th, 2012, 8:05 am
Contact:

Re: Computing some burrs

Post by flawr » October 17th, 2012, 1:38 pm

oh, great, I am lookng forward to making one myself out of wood, because it is my first design for a holey 6piece burr=) According to burrtools there are 4 assemblies (3 invalid ones) and it's level is maximum 5 (because burrtools tells me it is 5, which i can reproduce with the motions) but I do not know the convention of how to count moves that involve more than one piece. And I just noticed that you cannot find any major similarities between that one and the one I started the developing of that one with. So those random calculated burrs seem more an inspiration or a creativity-seed than a finished burr, which again tells someting about the usefullness and reasonability of brute-force-calculations, because for getting a nice puzzle, it still needs some 'hard' work - on the one hand you'd need to search throu thousands of 'possible' puzzles until you could find a 'good' one, on the other hand you can just take one that is 'not so good' and you have to change it over and over again until you get a nice puzzle. But thats just my point of view NOW after that experience=)

User avatar
Pio2001
Posts: 218
Joined: August 10th, 2011, 9:22 pm
Location: France
Contact:

Re: Computing some burrs

Post by Pio2001 » October 17th, 2012, 6:17 pm

I made a mistake. I didn't see the voxel on the pink piece. Without it, the puzzle has 19 assemblies and 5 solutions.

...but I nonetheless found, by luck, the 5-moves one enforced by the extra voxel.

Post Reply