interesting human solvable puzzle

For fans of Killer Sudoku, Samurai Sudoku and other variants

Re: interesting human solvable puzzle

Postby urhegyi » Tue Sep 28, 2021 8:35 pm

1to9only wrote:New version (2021.9.28) of SukakuExplainer, see post here: http://forum.enjoysudoku.com/help-with-sudoku-explainer-t6677-270.html#p310348

I tested the posted puzzle above with rating 11.9 on the command line and get this:
Code: Select all
.4...1..7............2.......3.........4..6.9................5................... ED=20.0/0.0/2.8

What's getting wrong?(took 4.5 hours to find result)
contains diagonals
and custom pattern:
Code: Select all
.1.....2.11.1.2.22.1.1.2.2..111.222...........333.444..3.3.4.4.33.3.4.44.3.....4.
urhegyi
 
Posts: 612
Joined: 13 April 2020

Re: interesting human solvable puzzle

Postby 1to9only » Tue Sep 28, 2021 9:11 pm

1to9only wrote:
Code: Select all
java.exe -Xrs -Xmx1g -cp SukakuExplainer.jar diuf.sudoku.test.serate -U -X --input=custom.txt --output=custom.se

Did u use -X with -U?
The default (no options specified) is a vanilla sudoku.
The options describe the sudoku variant(s).
1to9only
 
Posts: 4039
Joined: 04 April 2018

Re: interesting human solvable puzzle

Postby urhegyi » Tue Sep 28, 2021 9:23 pm

This was the command I used:
Code: Select all
java.exe -Xrs -Xmx1g -cp C:\Users\Peter\Documents\SukakuExplainer\SukakuExplainer.jar diuf.sudoku.test.serate -U -X --input=C:\Users\Peter\Documents\SukakuExplainer\custom.txt --output=C:\Users\Peter\Documents\SukakuExplainer\custom.se

Tested with a SE 2.8 puzzle X-diagonals and custum regions like above
same error:
Code: Select all
...........3...8...86...21....2.6...............4.1....27...59...4...1........... ED=20.0/1.5/1.5
urhegyi
 
Posts: 612
Joined: 13 April 2020

Re: interesting human solvable puzzle

Postby 1to9only » Tue Sep 28, 2021 9:59 pm

Using the released SukakuExplainer.jar, I don't have a problem with the ED=2.8/1.2/1.2 Custom-X sudoku.

I'll assume you're using today's (2021.9.28) version - older versions will simply ignore the -U option!
And the SukakuExplainer.json has the correct custom (extra regions) layout:
Code: Select all
"custom":".1.....2.11.1.2.22.1.1.2.2..111.222...........333.444..3.3.4.4.33.3.4.44.3.....4.",
"count":"4",

I'm using hints to rate the ED=11.9 Custom-X - checking the output file, the program is chugging, has been for 50 mins now!
Nothing more I can suggest at the moment.

Edit: ED=11.9/2.9/2.6 took about 1 hour here.

Edit: Does your GUI work on the ED=2.8 sudoku?
Last edited by 1to9only on Tue Sep 28, 2021 10:07 pm, edited 1 time in total.
1to9only
 
Posts: 4039
Joined: 04 April 2018

Re: interesting human solvable puzzle

Postby urhegyi » Tue Sep 28, 2021 10:04 pm

1to9only wrote:Using the released SukakuExplainer.jar, I don't have a problem with the ED=2.8/1.2/1.2 Custom-X sudoku.

I'll assume you're using today's (2021.9.28) version - older versions will simply ignore the -U option!
And the SukakuExplainer.json has the correct custom (extra regions) layout:
Code: Select all
"custom":".1.....2.11.1.2.22.1.1.2.2..111.222...........333.444..3.3.4.4.33.3.4.44.3.....4.",
"count":"4",

I'm using hints to rate the ED=11.9 Custom-X - checking the output file, the program is chugging, has been for 50 mins now!
Nothing more I can suggest at the moment.

Edit: Does the GUI work on the ED=2.8 sudoku?

Deleted the JSON file. Created new one on restart
Now it works for SE2.8
Code: Select all
...........3...8...86...21....2.6...............4.1....27...59...4...1........... ED=2.8/1.2/1.2
urhegyi
 
Posts: 612
Joined: 13 April 2020

Re: 8-clues Custom-X

Postby 1to9only » Mon Oct 04, 2021 10:31 am

urhegyi wrote:Yesterday I tried to generate a hard one instead of a fiendish one. It stopped after a long time and a memory use over 4.2 GB.

When using the new 'Generate (NFC enabled, can use 100% CPU)...' option, the difficulty level should only be set to Diabolical. Setting difficulty to, say Fiendish (2.6-6.0), a valid grid will eventually be generated, AFTER rejecting many many grids that are 6.0+. Check the console window output.

About the high memory usage, I have also noticed this when generating/solving the very tough puzzles. This is not an issue if you have lots of RAM! I have some ideas about what may be causing this, so I may make some code changes in the next release to see if memory usage can be reduced.
1to9only
 
Posts: 4039
Joined: 04 April 2018

Re: interesting human solvable puzzle

Postby urhegyi » Sun Oct 24, 2021 6:39 am

After exploring the SE12.3 jigsaw found by, Mathimagics http://forum.enjoysudoku.com/hardest-sudoku-jigsaws-js-ed-11-x-t38196-30.html#p311596 I found this layout has some opportunities to generate some human solvable grids too, like thisone solvable with as hardest step a swordfish.
Image
Code: Select all
.6.....7..4.6.8.9....1.3..............8...7..............5.7....1.3.9.2..3.....8. 111111123144552223145522663445226633455266337456668397458888397488999997889977777

with sudokuexplainer as custom latin square:
Code: Select all
.6.....7..4.6.8.9....1.3..............8...7..............5.7....1.3.9.2..3.....8. ED=3.8/1.5/1.5

with jigsawexplainer:
Code: Select all
.6.....7..4.6.8.9....1.3..............8...7..............5.7....1.3.9.2..3.....8. 111111123144552223145522663445226633455266337456668397458888397488999997889977777 ED=3.8/1.2/1.2

Only difference is in rating singles
urhegyi
 
Posts: 612
Joined: 13 April 2020

Re: interesting human solvable puzzle

Postby urhegyi » Thu Oct 28, 2021 9:10 am

Image
urhegyi
 
Posts: 612
Joined: 13 April 2020

Previous

Return to Sudoku variants