MatPlus.Net

 Website founded by
Milan Velimirović
in 2006

23:25 UTC
ISC 2024
 
  Forum*
 
 
 
 

Username:

Password:

Remember me

 
Forgot your
password?
Click here!
SIGN IN
to create your account if you don't already have one.
CHESS
SOLVING

Tournaments
Rating lists
1-Oct-2024

B P C F





 
 
MatPlus.Net Forum Internet and Computing Popeye hint #1
 
You can only view this page!
(1) Posted by Geoff Foster [Thursday, Aug 14, 2008 06:01]

Popeye hint #1


When Popeye solves a series-mover in N moves, it first looks for a solution in 1 move, then 2 moves, right up to N-1 moves, before looking for a solution in N moves. This can be irritating if one already knows that there are no solutions in fewer than N moves. A way to get around this is to use the "StartMoveNumber" option. This option is supposed to be used to re-start solving after an interruption, but the following text will trick Popeye into solving a series-mover without looking for shorter solutions:

option intelligent StartMoveNumber 1

In German this is "option intelligent StartZugnummer 1", in French it is "option intelligent Enroute 1".
 
(Read Only)pid=2614
(2) Posted by Joost de Heer [Thursday, Aug 14, 2008 19:06]

The same can be reached with stipulation 'exact-[stipulation]' instead of '[stipulation]'.
 
 
(Read Only)pid=2617
(3) Posted by Geoff Foster [Thursday, Aug 14, 2008 23:46]

True, but I found the "exact" stipulations were extremely slow.
 
   
(Read Only)pid=2618
(4) Posted by Eric Huber [Friday, Aug 15, 2008 00:10]

Thank you for the tip, Geoff.
 
 
(Read Only)pid=2620
(5) Posted by Thomas Maeder [Friday, Aug 15, 2008 23:00]

 QUOTE 
True, but I found the "exact" stipulations were extremely slow.

Care to post an example to the bugs tracker?
 
   
(Read Only)pid=2622
(6) Posted by Geoff Foster [Sunday, Aug 17, 2008 07:26]

I ran a test using the following problem.

T R Dawson
Fairy Chess Review, 1947
(= 2+8 )

series h#8
1.Kc4 2.Qb4 3.Rb3 4.Bc3 5.d4 6.Rb5 7.c5 8.Bd5, d3#

It took 3 hours 13 minutes (3:13:06) to solve. I tried "startmove 1" but it didn't help. I tried "exact-ser-h#8" and it was even slower (4:15:22). I then tried "option intelligent" and it took 0.28 seconds (0:0:0.281)!

This was such an amazing improvement that I had to find another problem to test using intelligent mode, so I used a much more difficult ser-h=9. This took 6 minutes 33 seconds (0:6:33.125) in intelligent mode, taking 0:3:46 for the first 8 moves and 0:2:47 for the final move. I also tried "startmove 1" and "exact" using intelligent mode, and they each took 0:2:47, which is not surprising, because this is the same time as it took before to do the final move.

So in summary, always use "option intelligent", and to make it even faster (more than twice as quick) use the "exact" stipulation or "startmove 1", which are the same thing. By the way, don't use "startmove 1" if you are not using intelligent mode, because it misses out the first move. Also, don't use "exact" if you are not using intelligent mode, because it is slower!
 
   
(Read Only)pid=2626
(7) Posted by Joost de Heer [Sunday, Aug 17, 2008 10:58]; edited by Joost de Heer [08-08-17]

That's what I meant: Use 'exact-[stipulation]' in combination with intelligent mode, it has the same effect as 'startmove 1'.
 
 
(Read Only)pid=2628
(8) Posted by Torsten Linß [Sunday, Aug 31, 2008 00:25]

Popeye uses hash tables. Therefore knowing that certain positions do not have a solution in (n-1) or (n-2) etc is advantageous. This explains why ser-h#8 is faster than exact-ser-h#8...
 
 
(Read Only)pid=2678
(9) Posted by Torsten Linß [Sunday, Aug 31, 2008 00:40]

Quoting Geoff Foster

 QUOTE 

It took 3 hours 13 minutes (3:13:06) to solve. I tried "startmove 1" but it didn't help. I tried "exact-ser-h#8" and it was even slower (4:15:22). I then tried "option intelligent" and it took 0.28 seconds (0:0:0.281)!

This was such an amazing improvement that I had to find another problem to test using intelligent mode, so I used a much more difficult ser-h=9. This took 6 minutes 33 seconds (0:6:33.125) in intelligent mode, taking 0:3:46 for the first 8 moves and 0:2:47 for the final move. I also tried "startmove 1" and "exact" using intelligent mode, and they each took 0:2:47, which is not surprising, because this is the same time as it took before to do the final move.


... and whenever using Popeye's intelligent option don't forget to send me a tipp ;-)
 
 
(Read Only)pid=2679

No more posts


MatPlus.Net Forum Internet and Computing Popeye hint #1