Author: leonid
Date: 03:39:35 04/19/01
Go up one level in this thread
On April 19, 2001 at 03:53:52, Tim Foden wrote: >On April 18, 2001 at 07:43:54, leonid wrote: > >>On April 18, 2001 at 07:40:53, leonid wrote: >> >>>Hello! >>> >>>This mate can be found with every program. If you like, your can solve it. >>> >>>[D]1knb1N1Q/1rp1r1r1/Pp1PPN2/nPQ2p2/1N1qBBp1/1n6/R5b1/2R1K3 w - - >>> >>>Please indicate your result. >>> >>>If you happened to use Genius, Chess Master or Rebel, probably you could help >>>me. I know those programs as fearly correct in solving mate problems and I am >>>curious to see how much effective they are in reducing search time by using >>>hash table. If you could say me time for search, with hash and without, for >>>every of those programs it will nice. The best will be to indicate the time for >>>search by brute force for 4 and 5 moves deep. It will be enough. >>> >>>Thanks for your help, >>> >>>Leonid. > >Well, I haven't got Genius, CM or Rebel, but I thought I'd give it a go anyway, >even though GreenLight was confused by your previous mate problem. :) > >GLC says that N8d7+ is mate in 9 after 10 seconds. > >Cheers, Tim. > >Analysis (v2.14-pr7, 880MHz Duron, 48M hash): > >>anal > Game stage: Opening > Current eval: 5.67 > Ply Time Score Nodes Principal Variation > 3 0.04 ++ 5888 Qxd4 (a=11.56 b=12.56 e=12.56) > 3 0.05 13.316 10153 Qxd4 Nxd4 axb7 Bxe4 dxe7 Nf3+ Kd1 > 3 0.11 13.316 25410 Qxd4 Nxd4 axb7 Bxe4 dxe7 Nf3+ Kd1 > 4 0.15 12.987 32823 Qxd4 Nxd6 N6d7+ Rxd7 Nxd7+ Rxd7 > 4 0.27 12.987 85374 Qxd4 Nxd6 N6d7+ Rxd7 Nxd7+ Rxd7 > 5 0.33 ++ 101761 Qxd4 (a=12.49 b=13.49 e=13.49) > 5 1.72 -- 740428 Qxd4 (a=13.49 b=321.00 e=13.49) > 5 1.74 13.956 747753 Qxd4 Nxd6 N6d7+ Rxd7 Nxd7+ Rxd7 exd7 Nxd4 Bxb7 > 5 1.75 13.956 747876 Qxd4 Nxd6 N6d7+ Rxd7 Nxd7+ Rxd7 exd7 Nxd4 Bxb7 > 6 1.95 ++ 816353 Qxd4 (a=13.46 b=14.46 e=14.46) > 6 2.22 15.561 921922 Qxd4 Nxd6 N8d7+ Rxd7 exd7 Nxd4 Qxd8+ Ka7 axb7 Nxb7 > Bxb7 Bxb7 Bxc7 > 6 7.63 20.185 3014468 N8d7+ Rxd7 Nxd7+ Rxd7 dxc7+ Rdxc7 Bxc7+ Ka7 Nc6+ > Ka8 axb7+ Kxb7 Nxd4+ Bxe4 > 6 8.43 20.185 3483264 N8d7+ Rxd7 Nxd7+ Rxd7 dxc7+ Rdxc7 Bxc7+ Ka7 Nc6+ > Ka8 axb7+ Kxb7 Nxd4+ Bxe4 > 7 8.44 ++ 3486588 N8d7+ (a=19.68 b=20.68 e=20.68) > 7 10.15 Mate09 4967198 N8d7+ Rxd7 Nxd7+ Rxd7 dxc7+ Rdxc7 Bxc7+ Ka7 Nc6+ > Nxc6 axb7+ Qa4 bxc8=N+ Kb7 Qxc6+ Kxc8 Qxd8# <ht> > 7 14.15 Mate09 8350544 N8d7+ Rxd7 Nxd7+ Rxd7 dxc7+ Rdxc7 Bxc7+ Ka7 Nc6+ > Nxc6 axb7+ Qa4 bxc8=N+ Kb7 Qxc6+ Kxc8 Qxd8# <ht> > 8 18.20 Mate09 11879k N8d7+ Rxd7 Nxd7+ Rxd7 dxc7+ Rdxc7 Bxc7+ Ka7 Nc6+ > Nxc6 axb7+ Qa4 bxc8=N+ Kb7 Qxc6+ Kxc8 Qxd8# > 8 40.41 Mate09 30918k N8d7+ Rxd7 Nxd7+ Rxd7 dxc7+ Rdxc7 Bxc7+ Ka7 Nc6+ > Nxc6 axb7+ Qa4 bxc8=N+ Kb7 Qxc6+ Kxc8 Qxd8# > 9 52.28 Mate09 41043k N8d7+ Rxd7 Nxd7+ Rxd7 dxc7+ Rdxc7 Bxc7+ Ka7 Nc6+ > Nxc6 axb7+ Qa4 bxc8=N+ Kb7 Qxc6+ Kxc8 Qxd8# <ht> > 9 1:37 Mate09 78904k N8d7+ Rxd7 Nxd7+ Rxd7 dxc7+ Rdxc7 Bxc7+ Ka7 Nc6+ > Nxc6 axb7+ Qa4 bxc8=N+ Kb7 Qxc6+ Kxc8 Qxd8# <ht> > 10 2:25 Mate09 120579k N8d7+ Rxd7 Nxd7+ Rxd7 dxc7+ Rdxc7 Bxc7+ Ka7 Nc6+ > Nxc6 axb7+ Qa4 bxc8=N+ Kb7 Qxc6+ Kxc8 Qxd8# > 10 5:55 Mate09 303108k N8d7+ Rxd7 Nxd7+ Rxd7 dxc7+ Rdxc7 Bxc7+ Ka7 Nc6+ > Nxc6 axb7+ Qa4 bxc8=N+ Kb7 Qxc6+ Kxc8 Qxd8# > 11 8:09 Mate09 423134k N8d7+ Rxd7 Nxd7+ Rxd7 dxc7+ Rdxc7 Bxc7+ Ka7 Nc6+ > Nxc6 axb7+ Qa4 bxc8=N+ Kb7 Qxc6+ Kxc8 Qxd8# <ht> > 11 10:20 1/71 dxc7+ >exit > local: t=10:27 nps=864854.5 n=542742056 (f=111433729 q=431308327) > total: t=10:27 nps=864854.5 n=542742056 draws=72102 > trans: probes=67195976 hits=4952373 (7.37%) draft=2841268 (4.23%) > tcuts: exact=197 (0.00%) upper=928466 (1.38%) lower=1889721 (2.81%) > tstor: exact=813 (0.00%) upper=18743943 (63.97%) lower=10555328 (36.02%) > ext: check=40868504 recap=827260 ppush=9996189 1rep=2046913 thrt=0 > q-moves: gen=8955270 tested=1676869 made/un=968733 max-dep=11 > max eval diff: part-1=5.356 part-2=3.325 >> Thanks, Tim! I hope now that this position is really 9 moves deep. I solved it only by selective since my brute force (600Mhz, no hash) took already 7 hours and 1 min 36 sec to say no. Could not permit to search 8 moves deep. Was able finally to find the time and how (have a new computer and this make me confuse sometime) my Rebel 10 must use hash without making too much use of hard disk. His hash is very effective. For 5 moves brute force and 10M hash (was afraid to give more hash) his time search went down for good 4 times (from 45 min 1 sec to 11 min 29 sec). Now I can see better how much advantage could be expected from big hash and much deeper search. Cheers, Leonid.
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.