Could the accepted answer for max solutions be wrong.

Everything about Sudoku that doesn't fit in one of the other sections

Re: Could the accepted answer for max solutions be wrong.

Postby champagne » Thu Mar 29, 2018 6:59 am

StrmCkr wrote:here's an idea/ method for generating all unique essentially different grids,
alternative methods have been done as well which takes around 2-3 weeks to generate in full IE all
5,472,730,538 essentially different grids


The process applied in the scan of ED solution grids to find 17 clues puzzles does it in less than one day as far as I remember

see here
Last edited by champagne on Sun Apr 01, 2018 11:04 am, edited 1 time in total.
champagne
2017 Supporter
 
Posts: 6046
Joined: 02 August 2007
Location: France Brittany

Re: Could the accepted answer for max solutions be wrong.

Postby StrmCkr » Thu Mar 29, 2018 3:33 pm

The process applied in the scan of ED solution grids to find 17 clues puzzles does it in less than one day as far as I remember

interesting time frame from what i remembered by gsf was approximately 2 weeks to generate from scratch.
but that was 12 years ago...

ps the link listed is broken. its missing a "L" @ html

probably a lot faster with todays tech and muti-threads
Some do, some teach, the rest look it up.
User avatar
StrmCkr
 
Posts: 759
Joined: 05 September 2006

Re: Could the accepted answer for max solutions be wrong.

Postby coloin » Sun Apr 01, 2018 9:07 am

No
PatmaxDaddy did the enumeration in 54.7 ms ....here
coloin
 
Posts: 1695
Joined: 05 May 2005

Re: Could the accepted answer for max solutions be wrong.

Postby Mathimagics » Sun Apr 01, 2018 9:53 am

Those very fast times are for counting the ED grids.

Actual enumeration (ie: producing a list of the 5.47 billion ED grids) would presumably slow things down just a trifle!

One only needs to do this job once, to produce a usable catalog of ED grids. gsf's sudoku tool can do this in one day or less (depending on CPU speed + multiple cores/threads availability), so the fast counting times reported are possibly moot.
User avatar
Mathimagics
2017 Supporter
 
Posts: 547
Joined: 27 May 2015

Re: Could the accepted answer for max solutions be wrong.

Postby champagne » Sun Apr 01, 2018 11:14 am

StrmCkr wrote:
The process applied in the scan of ED solution grids to find 17 clues puzzles does it in less than one day as far as I remember

interesting time frame from what i remembered by gsf was approximately 2 weeks to generate from scratch.
but that was 12 years ago...

ps the link listed is broken. its missing a "L" @ html

probably a lot faster with todays tech and muti-threads


Thanks to pointing on the url typo (post updated)

Part of the improvement is due to the overall performance of computers,
part comes from the use of a best performing brute force
and likely part derived of blue's handling of 2 bands auto morphing (I did not check what "gsf" did in that field).

The best performance is likely achieved using the minimal lexical form as canonical ED description. In the process shown in the link, the canonical form is slightly different (minimum number of valid solutions in a band with 6 clues)

And IMO writing the ED solution grids would not change that much the results with to-day computers

EDIT

just to answer to a remark here above

The file of ED solution grids is of interest if the planned task is to pick up randomly one solution grid. To work in sequence on a lot of grids, it's better to produce it in live.
champagne
2017 Supporter
 
Posts: 6046
Joined: 02 August 2007
Location: France Brittany

Previous

Return to General

cron