Author: Tord Romstad
Date: 15:20:10 09/27/05
Go up one level in this thread
Hi Steve! On September 27, 2005 at 17:29:54, Steve Maughan wrote: >Interesting! Is there any particular reason for switching back to MTD other >than curiosity? Yes. When I get some time for chess programming again (not any time soon, I'm afraid) I want to give parallell search a try. I think it will be easier to get a parallell search working if I start with a very simple and minimalistic search function. MTD is much easier to code than the other popular alpha beta variants, and can be implemented in very few lines of code. It also seems to be somewhat easier to parallellise, because all searches are done with a null window. This means that it is never necessary to adjust the bounds for other processors when one processor finds an improvement to alpha or beta. At least this is what I initially thought. Now I am no longer so sure. MTD is easier to implement, but on the other hand it tends to be harder to debug. When trying to write a parallell search, ease of debugging might be more important than ease of implementation. >What are your findings? Nothing interesting yet, except that it is possible to implement an MTD root driver, a search and a qsearch in less than 100 lines of code. >Are you going to stick with MTD or go back to PVS? I will almost certainly go back to PVS sooner or later. It is possible that there will be a public version with an MTD search first, though. As far as I know, there is no modern, open source chess engine of decent strength using MTD, which is a shame. There really should be an implementation for people to experiment with. >Have you started on your re-write of Glaurung? Yes, the MTD version is a complete re-write. It is still more than 100 points weaker than Glaurung Mainz, though (which isn't strange, considering that I have spent less than 10 hours working on it). >I haven't come across that much literature on double bound transposition tables. > I must say they intrigue me as I think the gap between the bounds could (in >some cases) be used to shape the tree i.e. extend or prune. I plan to play >around with double bounds with Monarch when I get the chance. Interesting, but in a PVS engine using two bounds I am afraid you usually won't find many nodes where both bounds are actually used. Tord
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.