Improving search speed.

Discussion of chess software programming and technical issues.

Moderators: hgm, Rebel, chrisw

ZirconiumX
Posts: 1334
Joined: Sun Jul 17, 2011 11:14 am

Re: Improving search speed.

Post by ZirconiumX »

konsolas wrote:Okay, my move ordering is now this:

Hash move (including the PV move if there is one
Winning captures (MVV/LVA)
Killer moves
Everything else ordered by history heuristic.

After searching around 5,000,000 nodes, My engine reaches depth 10 with a very reasonable PV (much better than before)

After comparing this to other engines, what I simply don't understand is how engines like Stockfish can search up to depth 10 in just 22466 nodes, whereas I need 5 million. What causes this huge difference?
Pruning. Lots and lots and lots of pruning.
Some believe in the almighty dollar.

I believe in the almighty printf statement.
User avatar
cdani
Posts: 2204
Joined: Sat Jan 18, 2014 10:24 am
Location: Andorra

Re: Improving search speed.

Post by cdani »

konsolas wrote:After comparing this to other engines, what I simply don't understand is how engines like Stockfish can search up to depth 10 in just 22466 nodes, whereas I need 5 million. What causes this huge difference?
Other search and ordering features like null move, lmr, various prunings... and tuning, tuning, tuning, tuning...
bob
Posts: 20943
Joined: Mon Feb 27, 2006 7:30 pm
Location: Birmingham, AL

Re: Improving search speed.

Post by bob »

konsolas wrote:Okay, my move ordering is now this:

Hash move (including the PV move if there is one
Winning captures (MVV/LVA)
Killer moves
Everything else ordered by history heuristic.

After searching around 5,000,000 nodes, My engine reaches depth 10 with a very reasonable PV (much better than before)

After comparing this to other engines, what I simply don't understand is how engines like Stockfish can search up to depth 10 in just 22466 nodes, whereas I need 5 million. What causes this huge difference?
Advanced LMR, NULL-move, forward pruning, etc...