Stockfish 2.3 update available

Discussion of anything and everything relating to chess playing software and machines.

Moderators: bob, hgm, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
JManion
Posts: 187
Joined: Wed Dec 23, 2009 7:53 am

Re: Stockfish 2.3 update available

Post by JManion » Sun Sep 16, 2012 6:11 am

What is the difference between GCC 4.2 sse and the regular 4.2 sse?

Thanks

Zlatin
Posts: 7
Joined: Wed Dec 07, 2011 7:38 pm

Re: Stockfish 2.3 update available

Post by Zlatin » Sun Sep 16, 2012 7:02 am

Bram Visser wrote:Great! Waiting for the new Houdini and we get a new Stockfish !
Yes!
Also new Rybka 5 (maybe) :D
This is Top 3 !

IWB
Posts: 1539
Joined: Thu Mar 09, 2006 1:02 pm

Re: Stockfish 2.3 update available

Post by IWB » Sun Sep 16, 2012 7:41 am

Hello Jim,

AsI had some bad experiences with the last releases and very fast bugfixes I will wait a while.

I am a bit sceptical about the huge diverity of special compiles for this or that.

Here is an example:

This is the special i5-i7 compile ON A AMD-PHENOM2:

Code: Select all


Engine: Stockfish 2.3 JA 64bit SSE4.2 INTEL (256 MB)
by Tord Romstad, Marco Costalba and Joona
 15/33	 0:01 	-0.60 	38...Nf6 39.d5 Nb6 40.Nc3 Bb8 41.Rd1 Re8 42.Kf1 Kh7 43.Kg1 Kg8 44.Qa2 Re5 45.Ra1 Nfd7 46.Rd1 Kf8 47.Rd4 Nc5 (2.631.468) 1414
 16/33	 0:01 	-0.60 	38...Nf6 39.d5 Nb6 40.Nc3 Bb8 41.Rd1 Re8 42.Kf1 Kh7 43.Kg1 Kg8 44.Qa2 Re5 45.Ra1 Nfd7 46.Rd1 Kf8 47.Rd4 Nc5 (2.690.563) 1410
 17/33	 0:02 	-0.52--	38...Nf6 39.d5 Nb6 40.Nc3 Bb8 41.Rd1 Re8 42.Kf1 Kh7 43.Ne2 Nfxd5 44.Nf4 c6 45.Kg1 Kg8 46.bxc6 bxc6 47.Nxd5 Nxd5 48.Bf1 Qa7 49.Bb5 (3.022.727) 1401
 17/33	 0:02 	-0.68++	38...Nf6 39.d5 Bb6 40.Bc3 Rh7 41.Rd1 Nd7 42.Qa2 Ne5 43.Bxe5 dxe5 44.Qa1 Qf6 45.Nxb6 Nxb6 46.Qa7 Rd7 47.Qxb7 Qd6 (3.753.200) 1380
 17/33	 0:03 	-0.72 	38...Bb6 39.Qc4 Re8 40.Qb3 Rg8 41.Bf1 Rh8 42.Bg2 Nce7 43.Qc4 Ba5 44.Nc5 Bb4 45.Nxb7 Qh7 46.Na5 Qh2+ 47.Kf1 Kg7 (5.281.540) 1379
 18/33	 0:04 	-0.80++	38...Nf6 39.Qxf7+ Kxf7 40.Bf1 Bb6 41.Bc3 d5 42.Bb4 Ke6 43.Nc5+ Bxc5 44.dxc5 Nd7 45.Bc3 Ne5 46.Bg2 b6 47.c6 Ne7 (5.687.417) 1373
 18/33	 0:04 	-0.64--	38...Nf6 39.Qxf7+ Kxf7 40.Bf1 Bb6 41.Bc3 d5 42.Bb4 Ke6 43.Kg2 Nd6 44.Nxb6 cxb6 45.Ra7 Nc4 46.Rxb7 Nd7 47.Bxc4 dxc4 48.Rc7 (5.923.031) 1378
 18/33	 0:04 	-0.52--	38...Nf6 39.Qxf7+ Kxf7 40.Bf1 Bb6 41.Bc3 d5 42.Bb4 Ke6 43.Kg2 Nd6 44.Nxb6 cxb6 45.Be2 Nc4 46.Rc1 Rc8 47.Ra1 Nd6 48.Ra7 (6.695.845) 1382
 18/33	 0:04 	-0.52 	38...Nf6 39.Qxf7+ Kxf7 40.Bf1 Bb6 41.Bc3 d5 42.Bb4 Ke6 43.Kg2 Nd6 44.Nxb6 cxb6 45.Be2 Nc4 46.Ra7 Rb8 47.Bd1 Nd6 48.Be2 Nd7 49.Ra1 Rc8 50.Rh1 Nf6 (6.768.778) 1383
 19/34	 0:05 	-0.60++	38...Nf6 39.Qxf7+ Kxf7 40.Nc3 Bb8 41.Ne2 Nd5 42.Nc3 Ncb6 (7.293.834) 1376
 19/34	 0:05 	-0.68++	38...Nf6 39.Qxf7+ Kxf7 40.Nc3 Bb8 41.Ne2 Nd5 42.Nc3 Ncb6 (7.389.918) 1378
 19/34	 0:05 	-0.64 	38...Nf6 39.Qxf7+ Kxf7 40.Bc3 Nb6 41.Bf1 Nfd5 42.Nxb6 Bxb6 43.Bc4 Ke6 44.Kg2 c6 45.bxc6 bxc6 46.Ra6 Rb8 47.Kf1 Kd7 48.Bxd5 cxd5 49.Ke2 Bd8 50.Bd2 (8.042.660) 1387
 20/34	 0:07 	-0.56--	38...Nf6 39.Qc2 Nb6 40.Nc3 Bb8 41.Ne2 Kg7 42.Nf4 Qc4 43.Qxc4 Nxc4 44.Bc3 Kf7 45.Bf1 Nb6 46.Bd2 Nbd5 47.Nxd5 Nxd5 48.Bc4 (9.891.217) 1391
 20/34	 0:07 	-0.72++	38...Nf6 39.Qc2 Nb6 40.Nc3 Bb8 41.Ne2 Nfd5 42.Kf1 Kf6 43.Ra5 Ke7 (10.856.867) 1386
 20/34	 0:08 	-0.68 	38...Nf6 39.Qc2 Nb6 40.Nc3 Bb8 41.Ne2 Nfd5 42.Bc3 Nxc3 43.Nf4+ Kf6 44.Qxc3 Nd5 45.Qb2 Ke7 46.Nxd5+ Qxd5 47.Qb4 Kf7 48.b6 c6 49.Qc3 Kg6 50.Bf1 Kf6 51.Bc4 (11.708.912) 1385
 21/36	 0:09 	-0.76++	38...Nf6 39.Qc2 Nb6 40.Nc3 Bb8 41.Ne2 Nfd5 42.Bc3 Nxc3 43.Nf4+ Kg7 44.Qxc3 Nd5 45.Nxd5 Qxd5 46.Qb2 Kg6 47.b6 c6 48.Qc3 Qg8 (12.910.161) 1386
 21/36	 0:09 	-0.72 	38...Nf6 39.Qc2 Nb6 40.Nc3 Bb8 41.Ne2 Nfd5 42.Bc3 Nxc3 43.Nf4+ Kg7 44.Qxc3 Nd5 45.Nxd5 Qxd5 46.b6 cxb6 47.Qb4 Bc7 48.Rc1 Qf7 49.d5 Rc8 50.Qc4 Kg6 51.Kf1 Kg5 (13.857.207) 1390
 22/36	 0:10 	-0.72 	38...Nf6 39.Qc2 Nb6 40.Nc3 Bb8 41.Ne2 Nfd5 42.Bc3 Nxc3 43.Nf4+ Kg7 44.Qxc3 Nd5 45.Nxd5 Qxd5 46.b6 cxb6 47.Qb4 Bc7 48.Rc1 Qf7 49.d5 Rc8 50.Qc4 Kg6 51.Kf1 Kg5 (14.346.501) 1388
 23/36	 0:10 	-0.64--	38...Nf6 39.Qc2 Nb6 40.Nc3 Bb8 41.Ne2 Nfd5 42.Bc3 Nxc3 43.Nf4+ Kg7 44.Qxc3 Nd5 45.Nxd5 Qxd5 46.b6 cxb6 47.Rb1 Qc6 48.Qb3 d5 49.Qxb6 Qxb6 50.Rxb6 (15.215.441) 1387
 23/38	 0:12 	-0.64 	38...Nf6 39.Qc2 Nb6 40.Nc3 Bb8 41.Ne2 Nfd5 42.Bc3 Nxc3 43.Nf4+ Kf6 44.Qxc3 Nd5 45.Nxd5+ Qxd5 46.b6 cxb6 47.Rb1 Qc6 48.Qxc6 bxc6 49.Rxb6 Rc8 50.Bf1 Kg5 51.Ba6 Ba7 (17.454.865) 1385
 24/44	 0:13 	-0.72++	38...Nf6 39.Qc2 Nb6 40.Nc3 Bb8 41.Ne2 Nfd5 42.Bc3 Nxc3 43.Nf4+ Kf6 44.Qxc3 Nd5 45.Nxd5+ Qxd5 46.b6 cxb6 47.Rb1 Qc6 48.Qxc6 bxc6 49.Rxb6 Rc8 50.Bf1 Kg5 51.Ba6 Ba7 (19.331.095) 1390
 24/44	 0:16 	-0.68 	38...Nf6 39.Qc2 Nb6 40.Nc3 Bb8 41.Ne2 Nfd5 42.Bc3 Nxc3 43.Nf4+ Kf6 44.Qxc3 Nd5 45.Nxd5+ Qxd5 46.b6 cxb6 47.Rb1 Qc6 48.Qxc6 bxc6 49.Rxb6 Rc8 50.Bf1 Kg5 51.Ba6 Ba7 (22.684.579) 1385
 25/44	 0:19 	-0.60--	38...Nf6 39.Qc2 Nb6 40.Nc3 Bb8 41.Ne2 Nfd5 42.Bc3 Nxc3 43.Nf4+ Kf6 44.Qxc3 Nd5 45.Nxd5+ Qxd5 46.b6 cxb6 47.Rb1 Qc6 48.Qxc6 bxc6 49.Rxb6 Rc8 50.Bf1 Kg5 51.Ba6 Ba7 (27.599.850) 1383
 25/44	 0:50 	-0.76++	38...Bb6 (69.117.362) 1369
best move: Ba7-b6 time: 1:02.109 min  n/s: 1.369.908  nodes: 69.117.362 
This is the NORMAL SSE42 compile on an AMD PHENOM2:

Code: Select all

Engine: Stockfish 2.3 JA 64bit SSE4.2 AMD (256 MB)
by Tord Romstad, Marco Costalba and Joona
 16/28	 0:01 	-0.68 	38...Bb6 39.Bc3 Nxc3 40.Qxc3 Qd7 41.Qb3 Kg7 42.Nc3 Ne7 43.Na4 d5 44.Nxb6 cxb6 45.Bf1 Qc7 46.Qb4 Ng6 47.Qb3 Ne7 (1.522.495) 1352
 17/29	 0:02 	-0.60--	38...Bb6 39.Qc4 Ba5 40.Nc5 Ncb6 41.Qc2 Bb4 42.Nxb7 Qh7 43.Na5 Bxa5 44.Rxa5 Nb4 45.Qb3 N6d5 46.Ba1 (3.367.590) 1338
 17/30	 0:03 	-0.52--	38...Bb6 39.Qc4 Nf6 40.Qc3 Nd5 41.Qc4 (4.118.735) 1344
 17/30	 0:03 	-0.52 	38...Bb6 39.Qc4 Nf6 40.Qc3 Rh6 41.Qd2 Kg7 42.Qb4 Nd5 43.Qb3 Nce7 44.Kf1 Kg8 45.Kg1 Ba5 46.Nc5 b6 47.Nb7 Bb4 48.Ra8+ Kg7 (5.303.510) 1336
 18/30	 0:04 	-0.60++	38...Nf6 39.d5 Bb6 40.Qd1 Rh5 41.Nxb6 Nxb6 (6.250.573) 1333
 18/30	 0:04 	-0.68++	38...Nf6 39.d5 Bb6 40.Qd1 Rh5 41.Nxb6 Nxb6 (6.496.526) 1336
 18/30	 0:05 	-0.80++	38...Nf6 39.d5 Bb6 40.Qd1 Rh5 41.Nxb6 Nxb6 42.Qd4 Nbd7 43.Ra7 Qxd5 44.Qxd5 Nxd5 45.Rxb7 Nc5 (7.301.911) 1342
 18/30	 0:05 	-0.72 	38...Nf6 39.d5 Bb6 40.Qc3 Rh5 41.Rd1 Rh7 42.Ba1 Nd7 43.Nxb6 Ndxb6 44.Bf1 Kh5 45.Qa5 Qg8 46.Be2 Kh6 47.Bf6 Ne7 48.Kg2 Nexd5 49.Rh1+ Kg6 (8.008.614) 1348
 19/30	 0:06 	-0.64--	38...Nf6 39.Qc2 Nb6 40.Nc3 Bb8 41.Ne2 Qc4 42.Nf4+ Kf7 43.Qxc4+ Nxc4 44.Bc3 c6 45.bxc6 bxc6 46.Rb1 Bc7 47.d5 (9.161.527) 1344
 19/30	 0:07 	-0.76 	38...Nf6 39.Qc2 Nb6 40.Nc3 Bb8 41.Ne2 Nfd5 42.Qb3 Kg7 43.Nf4 Rh6 44.Nxd5 Qxd5 45.Qc3 Qc4 46.Bf1 Qxc3 47.Bxc3 Kf6 48.Ra2 Nd5 (9.562.970) 1341
 20/30	 0:07 	-0.68--	38...Nf6 39.Qc2 Nb6 40.Nc3 Bb8 41.Ne2 Nfd5 42.Bc1 Nb4 43.Nf4+ Kf6 44.Qb2 N4d5 45.Nxd5+ Qxd5 46.Bd2 Nc4 47.Qb3 Nb6 48.Qb2 (10.396.214) 1341
 20/30	 0:07 	-0.60--	38...Nf6 39.Qc2 Nb6 40.Nc3 Bb8 41.Ne2 Nfd5 42.Bc1 Nb4 43.Nf4+ Kf6 44.Qb2 N4d5 45.Nxd5+ Qxd5 46.Bd2 Nc4 47.Qb3 Nb6 48.Qb2 (10.581.321) 1343
 20/30	 0:11 	-0.64 	38...Bb6 39.Qc4 Re8 40.Qb3 Rg8 41.Bc3 Nxc3 42.Qxc3 Rh8 43.Rc1 Rh7 44.Qb4 Kf6 45.Ra1 Rh8 46.Qc3 d5 47.Qb3 Nd6 48.Nc3 Nc4 49.Ra4 (15.280.696) 1341
 21/30	 0:12 	-0.60 	38...Bb6 39.Qc4 Re8 40.Qb3 Rg8 41.Bc3 Nxc3 42.Qxc3 Rh8 43.Rc1 Rh7 44.Qb4 Kf6 45.Ra1 Rh8 46.Rc1 d5 47.Nc3 Kg7 48.Qb3 Ne7 49.Ra1 Qe6 50.Kf1 (16.506.994) 1343
 22/30	 0:13 	-0.60 	38...Bb6 39.Qc4 Re8 40.Qb3 Rg8 41.Bc3 Nxc3 42.Qxc3 Rh8 43.Rc1 Rh7 44.Qb4 Kf6 45.Ra1 Rh8 46.Rc1 d5 47.Nc3 Kg7 48.Qb3 Ne7 49.Ra1 Qe6 50.Kf1 (17.722.324) 1350
 23/32	 0:17 	-0.68++	38...Bb6 39.Qc4 Rd8 40.Qc1 Nb4 41.Bc3 Nd3 42.Qc2 Rh8 43.Nb2 Nxb2 44.Qxb2 d5 45.Ra8 Re8 46.Qb3 Nd6 47.Rxe8 Nxe8 48.Bb4 Qd7 49.Ba3 Ba5 (24.031.477) 1351
 23/33	 0:21 	-0.64 	38...Bb6 39.Qc4 Rd8 40.Bc3 Nxc3 41.Qxc3 Rh8 42.Rc1 Qd7 43.Qb4 Ne7 44.Nxb6 cxb6 45.Qb3 Kg7 46.Ra1 d5 47.Qb4 Rd8 48.Bf1 Rc8 49.Qb3 Rh8 50.Qb4 Qc7 (29.185.585) 1353
 24/36	 0:25 	-0.72++	38...Nf6 39.d5 Bb6 40.Qc3 Rh7 41.Nxb6 Nxb6 42.Qd4 Nbxd5 43.Rd1 Nb6 44.Bf1 Nfd5 (35.042.314) 1353
 24/39	 0:33 	-0.60 	38...Nf6 39.Qxf7+ Kxf7 40.Rc1 Nd5 41.Bf1 Ncb6 42.Ra1 Nxa4 43.Rxa4 Bb6 44.Bc4 Ke6 45.Bc3 c6 46.bxc6 bxc6 47.Ra6 Rb8 48.Kf1 Kd7 49.Bxd5 cxd5 50.Ke2 Bc7 51.Kd2 Bd8 (45.427.260) 1367
 25/39	 0:34 	-0.68++	38...Nf6 39.Qxf7+ Kxf7 40.Rc1 Nd5 41.Bf1 Ncb6 42.Ra1 Nxa4 43.Rxa4 Bb6 44.Bc4 Ke6 45.Bc3 c6 46.bxc6 bxc6 47.Ra6 Rb8 48.Kf1 Kd7 49.Bxd5 cxd5 50.Ke2 Bd8 (47.525.815) 1370
 25/39	 0:38 	-0.68 	38...Nf6 39.Qc2 Nb6 40.Nc3 Bb8 41.Ne2 Nfd5 42.Bc3 Nxc3 43.Nf4+ Kf6 44.Qxc3 Nd5 45.Nxd5+ Qxd5 46.b6 cxb6 47.Rb1 Qc6 48.Qb2 d5 49.Qxb6 Bd6 50.Qxb7 Qxb7 51.Rxb7 Rc8 (52.117.008) 1370
 26/39	 0:41 	-0.60--	38...Nf6 39.Qc2 Nb6 40.Nc3 Bb8 41.Ne2 Nfd5 42.Bc3 Nxc3 43.Nf4+ Kf6 44.Qxc3 Nd5 45.Nxd5+ Qxd5 46.b6 cxb6 47.Rb1 Qc6 48.Qb2 d5 49.Qxb6 Bd6 50.Bf1 Ra8 51.Qb3 (56.390.039) 1364
best move: Nd5-f6 time: 1:02.109 min  n/s: 1.364.946  nodes: 56.390.039 
All on one Thread!


While the intel compile is marginal faster on the AMD than the normal one the nodes per depth are not identical and the selective depth is different as well.

That are basicaly two different engines.

Bye
Ingo

mcostalba
Posts: 2684
Joined: Sat Jun 14, 2008 7:17 pm

Re: Stockfish 2.3 update available

Post by mcostalba » Sun Sep 16, 2012 7:48 am

Hi Ingo,

could you please run

Code: Select all

./stockfish bench
from the command line and see what searched nodes you get? You should have 5416292.

Thanks

Modern Times
Posts: 2451
Joined: Thu Jun 07, 2012 9:02 pm

Re: Stockfish 2.3 update available

Post by Modern Times » Sun Sep 16, 2012 8:21 am

IWB wrote:Hello Jim,
This is the special i5-i7 compile ON A AMD-PHENOM2:
Well spotted, I simply did not try and run the special Intel compile on my AMD :D

Modern Times
Posts: 2451
Joined: Thu Jun 07, 2012 9:02 pm

Re: Stockfish 2.3 update available

Post by Modern Times » Sun Sep 16, 2012 8:29 am

mcostalba wrote:Hi Ingo,

could you please run

Code: Select all

./stockfish bench
from the command line and see what searched nodes you get? You should have 5416292.

Thanks
I tried both on my Phenom II X6

SSE42
Time 4671
Nodes searched 5416292
Nodes per second 1159557


GCC i5-i7 SSE42
Time 3611
Nodes searched 5416292
Nodes per second 1499942

The output of the bench command looks identical in both cases as regards the pv


info depth 12 seldepth 25 score cp 226 nodes 125927 nps 1040719 time 121 multipv 1 pv c5d5 d1d5 e6d5 c4e5 d8c7 f4d3 c7c3
a1c3 b4c3 e5d7 f6e7 b3b4 c3c2 d7c5 e7c5 b4c5 b7c6 g1f1 c6a4

info depth 12 seldepth 25 score cp 226 nodes 125927 nps 1773619 time 71 multipv 1 pv c5d5 d1d5 e6d5 c4e5 d8c7 f4d3 c7c3
a1c3 b4c3 e5d7 f6e7 b3b4 c3c2 d7c5 e7c5 b4c5 b7c6 g1f1 c6a4
Last edited by Modern Times on Sun Sep 16, 2012 8:42 am, edited 2 times in total.

IWB
Posts: 1539
Joined: Thu Mar 09, 2006 1:02 pm

Re: Stockfish 2.3 update available

Post by IWB » Sun Sep 16, 2012 8:34 am

Modern Times wrote:
IWB wrote:Hello Jim,
This is the special i5-i7 compile ON A AMD-PHENOM2:
Well spotted, I simply did not try and run the special Intel compile on my AMD :D
As the GCC is faster and runs on AMD and Intel. The non GCC compile is redundent.

I get different pv with both versions which means that one version is somehow "strange". Whichone is a different matter.

Bye
Ingo

Modern Times
Posts: 2451
Joined: Thu Jun 07, 2012 9:02 pm

Re: Stockfish 2.3 update available

Post by Modern Times » Sun Sep 16, 2012 8:36 am

IWB wrote:I get different pv with both versions which means that one version is somehow "strange". Whichone is a different matter.

Bye
Ingo
Did you try stockfish bench ? The PVs are identical for me. I'll try a GUI and see what happens

mcostalba
Posts: 2684
Joined: Sat Jun 14, 2008 7:17 pm

Re: Stockfish 2.3 update available

Post by mcostalba » Sun Sep 16, 2012 8:42 am

IWB wrote: I get different pv with both versions which means that one version is somehow "strange". Whichone is a different matter.
Since the 'bench' test reported same node counts I'd assume the 2 binaries are 100% equivalent. Neverthless becuase you reported differnt pv it is interesting to understand better. Could you please post what kind of test have you done?

1) From command line or from GUI ?
2) Number of threads set or let SF to choose automatically ?
3) What command did you give ? depth limited, time limited, infinite ?

Thanks
Marco
Last edited by mcostalba on Sun Sep 16, 2012 8:43 am, edited 1 time in total.

Zlatin
Posts: 7
Joined: Wed Dec 07, 2011 7:38 pm

Re: Stockfish 2.3 update available

Post by Zlatin » Sun Sep 16, 2012 8:42 am

mcostalba wrote:Hi Ingo,

could you please run

Code: Select all

./stockfish bench
from the command line and see what searched nodes you get? You should have 5416292.

Thanks
Hello Marco,

Where must type this command?

Best wishes

Post Reply