*Where can I find resource on parallel search algorithms *(not just "difficult & painful to read" research papers) ?
*and has there been no improvement in this field -"I understand it's difficult to parallelizing this search"- in the past 20 years (according to chess programming wiki DTS(1994) has the best known scalability) ?
*what is the best open source engine to study it's code to get a clue of what Parallel search looks like (PS: the likes of stock fish & Crafty are fairly difficult to read "the search part") ?
Parallel Search algorithms resources ?
Moderator: Ras
-
MahmoudUthman
- Posts: 237
- Joined: Sat Jan 17, 2015 11:54 pm
-
jdart
- Posts: 4429
- Joined: Fri Mar 10, 2006 5:23 am
- Location: http://www.arasanchess.org
Re: Parallel Search algorithms resources ?
Do you want something easy to understand and implement, or something that is high performing? It is hard to get both.
--Jon
--Jon
-
bob
- Posts: 20943
- Joined: Mon Feb 27, 2006 7:30 pm
- Location: Birmingham, AL
Re: Parallel Search algorithms resources ?
I think it is hard to get EITHER.jdart wrote:Do you want something easy to understand and implement, or something that is high performing? It is hard to get both.
--Jon
-
Edsel Apostol
- Posts: 803
- Joined: Mon Jul 17, 2006 5:53 am
- Full name: Edsel Apostol
Re: Parallel Search algorithms resources ?
Viper is one of the simple YBWC implementation out there. It's a simplified Glaurung (a pre-cursor to Stockfish).MahmoudUthman wrote:*Where can I find resource on parallel search algorithms *(not just "difficult & painful to read" research papers) ?
*and has there been no improvement in this field -"I understand it's difficult to parallelizing this search"- in the past 20 years (according to chess programming wiki DTS(1994) has the best known scalability) ?
*what is the best open source engine to study it's code to get a clue of what Parallel search looks like (PS: the likes of stock fish & Crafty are fairly difficult to read "the search part") ?
-
mar
- Posts: 2683
- Joined: Fri Nov 26, 2010 2:00 pm
- Location: Czech Republic
- Full name: Martin Sedlak
Re: Parallel Search algorithms resources ?
If 100 elo (4 cores vs 1) is enough, then lazy smp is worth trying as it has the best performance/price ratio.
-
MahmoudUthman
- Posts: 237
- Joined: Sat Jan 17, 2015 11:54 pm
Re: Parallel Search algorithms resources ?
Jon Dart,Robert Hyatt,Edsel Apostol & Martin Sedlak ,thanks for replying .
.
easy to understand would be nicejdart wrote:Do you want something easy to understand and implement, or something that is high performing? It is hard to get both.
--Jon
.Edsel Apostol wrote:Viper is one of the simple YBWC implementation out there. It's a simplified Glaurung (a pre-cursor to Stockfish).
thank you I'll check both.mar wrote:If 100 elo (4 cores vs 1) is enough, then lazy smp is worth trying as it has the best performance/price ratio.