256x256 Rubik's cube solve - World record attempt

[Multicubing]

This looks like a LOT of fun, but unfortunately I do not have the time for it. Nor do I have the patience at this time to learn computer cubing. Computer cubing a little frustrating. Right now I'll stick with physical large cubes... but enjoying watching this progress!

 

[soup]

Small update showing off current progress at ~16 hours into the solve:

 

[abunickabhi]

There's not really any point in having lots of random people join if they aren't already decent at big cubes (say sub 40 minutes on 20x20). We also decided that each person must do at least 5 hours of the solve to be counted as a contributor

Interesting rules for team work for solving.

I hope you guys find enough volunteers. this seems like an interesting project.

Will the solution be recorded somewhere, or any end goal with this project, like a video or something?

 

[Ben Whitmore]

Interesting rules for team work for solving.

I hope you guys find enough volunteers. this seems like an interesting project.

Will the solution be recorded somewhere, or any end goal with this project, like a video or something?

The program records the reconstruction and allows exporting a video of the solve, so yes

 

[abunickabhi]

The program records the reconstruction and allows exporting a video of the solve, so yes

Do link up the video on this thread, once you export and decide to make it public.

Also the exec has to be sped up to make the video shorter

 

[soup]

The 1st center has been solved at about 24:19:00. Short video showing off the solved center:

 

[Scollier]

The 1st center has been solved at about 24:19:00. Short video showing off the solved center:

Wow! A whole day's worth just to solve one center. I hope it gets easier for you as there will be less sorting through colors, since green is elliminated, but a lot more breaking and restoring. Best of luck!

 

[qqwref]

A few things on the method:
- We're not coming up with this on the fly or anything - the method is basically the same one that has been used for big computer cubes for many years, with a couple of smaller timesaves discovered along the way. I explain the basic form of it in this youtube video, if anyone wants to try it themselves.
- The idea of the "clearing" or "cleaning" step is to solve a lot of pieces very efficiently (only a little over 1 move per piece) but at random. After that the remaining pieces are solved one at a time with a 4-move algorithm. We do this approach for each of the first 3 centers, and then for the remaining centers we have a specialized method to do it all in one step.
- The edges and 3x3x3 will be solved after the first two centers (green and blue). I normally start with white and yellow but we're using green and blue for now because that's what soup uses. The current version of the simulator allows us to change the color scheme as we go, so we might take advantage of that at some point.

 

[Ben Whitmore]

Wow! A whole day's worth just to solve one center. I hope it gets easier for you as there will be less sorting through colors, since green is elliminated, but a lot more breaking and restoring. Best of luck!

Typically the second center takes about the same amount of time as the first center, but it starts speeding up a lot after that. We usually find that first 2 centers + edges is around 50% of the solve.

 

[abunickabhi]

Typically the second center takes about the same amount of time as the first center, but it starts speeding up a lot after that. We usually find that first 2 centers + edges is around 50% of the solve.

Interesting statistics. How did you arrive at this stat?

 

[Ben Whitmore]

Interesting statistics. How did you arrive at this stat?

By doing solves and observing that it was the case, also the fact that log(24!/(4!^2 16!)) / log(24!/4!^6) ~= 0.497

 

[abunickabhi]

By doing solves and observing that it was the case, also the fact that log(24!/(4!^2 16!)) / log(24!/4!^6) ~= 0.497

Ahh thanks for the info!

 

[Neon]

Hi there. I'd like to help with 256x256, and I have some small method ideas, some of witch might be useful and not known.
It's Igor Tarasov(don't know if it's possible to change my name here), besides awful 17 hour 100x100 solve I have near sub9 10x10 single and something like 36 on 20x20

 

[effperm]

you can change your name if you are premium

also the clearing of the second center is finished!

 

[qqwref]

The second center is now done. When the edges and 3x3 (the next step) are finished, we'll be about halfway there.

 

[Ben Whitmore]

About 2 hours after the second center was finished, edges and 3x3 are now solved:

 

[effperm]

today is the 1 month anniversary of the 256x256x256 solve

woohoo

 

[Ben Whitmore]

qqwref just finished clearing the third center, total time just under 55 hours:

 

[joshsailscga]

I see you did end up using that color scheme swap option you mentioned earlier. Must be neat to have that ability in the software, and handy for your different solving orders.

 

[kbrune]

A few things on the method:
- We're not coming up with this on the fly or anything - the method is basically the same one that has been used for big computer cubes for many years, with a couple of smaller timesaves discovered along the way. I explain the basic form of it in this youtube video, if anyone wants to try it themselves.

This is so cool. Wish I had the skills to get in on this action. I've never seen or tried computer solving. I watched a bit of the video you linked but i'm still a bit confused. Learning the commutator stuff you talk about isn't a big problem though. However I know less then nothing about solving on a comp. Have you or anyone posted a more beginner video on things like: software to use, user interface explanations, stuff like that? Or is it trial and error friendly?

Where do you dload a computer solving program?