Jump to content

  • Log In with Google      Sign In   
  • Create Account

#Actualpatishi

Posted 03 July 2013 - 06:19 AM

EDIT: I rewrote this post, i've  managed to work it!  now i have a zobrist hashing and the speed gain is very nice.    thanks a lot Alvaro for the help smile.png

      But i have a another question related to the transposition table management.  would you recommend using some sort of replacement scheme for same positions with different depth?   say if the new position has a larger depth than the saved one than i should replace it  with the new one.        casue right now i am not doing any replacement,i am just inserting the positions blindly to the table.     

I am handling collisions by using a LinkedList in every table index,so if two board positions fall at the same index, i just add them to the list.   

And what size of Transposition table do you recommend using?  right now i have table at size 500009 .  but when i measure the amount of insertions to the table,I get more than 3 million ! (depth 15)


#12patishi

Posted 03 July 2013 - 06:19 AM

EDIT: I rewrote this post, i've  managed to work it!  now i have a zobrist hashing and the speed gain is very nice.    thanks a lot Alvaro for the help smile.png

      But i have a another question related to the transposition table management.  would you recommend using some sort of replacement scheme for same positions with different depth?   say if the new position has a larger depth than the saved one than i should replace it  with the new one.        casue right now i am not doing any replacement,i am just inserting the positions blindly to the table.     

I am handling collisions by using a LinkedList in every table index,so if two board positions fall at the same index, i just add them to the list.   

And what size of Transposition table do you recommend using?  right now i have table at size 500009 .  but when i measure the amount of insertions to the table,I get more than 3 million ! (at depth 15)


#11patishi

Posted 03 July 2013 - 06:01 AM

EDIT: I rewrote this post, i've  managed to work it!  now i have a zobrist hashing and the speed gain is very nice.    thanks a lot Alvaro for the help smile.png

      But i have a another question related to the transposition table management.  would you recommend using some sort of replacement scheme for same positions with different depth?   say if the new position has a larger depth than the saved one than i should replace it  with the new one.        casue right now i am not doing any replacement,i am just inserting the positions blindly to the table.     

I am handling collisions by using a List in every table index,so if two board positions fall at the same index, i just add them to the list.   

And what size of Transposition table do you recommend using?  right now i have table at size 500009 .  but when i measure the amount of insertions to the table,I get more than 3 million !


#10patishi

Posted 03 July 2013 - 05:53 AM

EDIT: I rewrote this post, i've  managed to work it!  now i have a zobrist hashing and the speed gain is very nice.    thanks a lot Alvaro for the help smile.png

      But i have a another question related to the transposition table management.  would you recommend using some sort of replacement scheme for same positions with different depth?   say if the new position has a larger depth than the saved one than i should replace it  with the new one.        casue right now i am not doing any replacement,i am just inserting the positions blindly to the table.     

I am handling collisions by using a List in every table index,so if two board positions fall at the same index, i just add them to the list.   

And what size of Transposition table do you recommend using?  right now i have table at size 500009 .  but when i measure the amount of retrievals from the table (EXACT values or Alpha Beta cut offs after updating alpha and beta) i get much more than a million returnes.


#9patishi

Posted 03 July 2013 - 05:43 AM

EDIT: I rewrote this post, i've  managed to work it!  now i have a zobrist hashing and the speed gain is very nice.    thanks a lot Alvaro for the help smile.png

      But i have a another question related to the transposition table management.  would you recommend using some sort of replacement scheme for same positions with different depth?   say if the new position has a larger depth than the saved one than i should replace it  with the new one.        casue right now i am not doing any replacement,i am just inserting the positions blindly to the table.     

I am handling collisions by using a List in every table index,so if two board positions fall at the same index, i just add them to the list.   


#8patishi

Posted 03 July 2013 - 05:43 AM

EDIT: I rewrote this post, i've  managed to work it!  now i have a zobrist hashing and the speed gain is very nice.    thanks a lot Alvaro for the help smile.png

      But i have a another question related to the transposition table management.  would you recommend using some sort of replacement scheme for same positions with different depth?   say if the new position has a larger depth than the saved one than should it replace with the new one.        casue right now i am not doing any replacement,i am just inserting the positions blindly to the table.     

I am handling collisions by using a List in every table index,so if two board positions fall at the same index, i just add them to the list.   


PARTNERS