Page 2 of 5
Re: Houdini 1.02
Posted: Sat Jun 19, 2010 11:09 pm
by beram
second 50 games match ended and this time DR4 won by small margin
Code: Select all
AMD Phenom II X6 1090T 3200 Mhz,(Fritzmark 23,1)
TC Blitz 2m+2s, 512 Mb Hash, 50 games Nunn2 positions
1 Deep Rybka 4 SSE42 x64 +10/=34/-6 54.00% 27.0/50
2 Houdini 1.02 x64 POPCNT 8_CPU +6/=34/-10 46.00% 23.0/50
so after 100 games they are almost at 50 - 50
Code: Select all
AMD Phenom II X6 1090T 3200 Mhz,(Fritzmark 23,1)
TC Blitz 2m+2s, 512 Mb Hash, 100 games Nunn2 and privatbook positions
1 Deep Rybka 4 SSE42 x64 +19/=63/-18 50.50% 50.5/100
2 Houdini 1.02 x64 POPCNT 8_CPU +18/=63/-19 49.50% 49.5/100
Who wants to play another 1900 games so we can who really know is better under this conditions ?
kind regards Bram
Re: Houdini 1.02
Posted: Sat Jun 19, 2010 11:28 pm
by 64x
Quick question. I thought SSE4.2 and POPCNT were only supported on Intel chips. How do these versions compare in performance with the standard 64bit versions on your Phenom?
Re: Houdini 1.02
Posted: Sun Jun 20, 2010 12:20 am
by Peter C
64x wrote:Quick question. I thought SSE4.2 and POPCNT were only supported on Intel chips. How do these versions compare in performance with the standard 64bit versions on your Phenom?
Phenom has SEE4a which includes POPCNT.
Peter
Re: Houdini 1.02
Posted: Sun Jun 20, 2010 12:30 am
by 64x
Thank you Peter. Is the instruction set on SSE4a also compatible with SSE4.2 at least on what these engines need (Rybka SSE42)?
Re: Houdini 1.02
Posted: Sun Jun 20, 2010 1:04 am
by Peter C
64x wrote:Thank you Peter. Is the instruction set on SSE4a also compatible with SSE4.2 at least on what these engines need (Rybka SSE42)?
Yeah. I know they run, but AFAIK the only thing special about Rybka 4 SEE42 is POPCNT.
Peter
Re: Houdini 1.02
Posted: Sun Jun 20, 2010 3:26 am
by JackStraw
4'2"/ i7-940 @ 3.75 /3 cores ea / 512 mb / ponder off / klo openings / R4 default settings , no large pages , using 3-4-5 egtb /
Houdini 1.02 POPCNT 4_cpu............51.5.........+22,=59,-19
Rybka 4 SSE42..........................48.5.........+19,=59,-22
I tried to put this under the " Tournaments and Matches " / Houdini thread , but every time I click on the thread I get a pop up box that says " Error on ply 27...e3 " , and it locks my computer up.........
Re: Houdini 1.02
Posted: Sun Jun 20, 2010 3:53 am
by kingliveson
JackStraw wrote:
.
.
.
I tried to put this under the " Tournaments and Matches " / Houdini thread , but every time I click on the thread I get a pop up box that says " Error on ply 27...e3 " , and it locks my computer up.........
It should be fixed now.
Re: Houdini 1.02
Posted: Sun Jun 20, 2010 4:42 am
by yanquis1972
surprised no one has mentioned houdini is now being 'officially' tested by ingo bauer --
http://www.inwoba.de/index.html
currently in the #2 spot trailing rybka by 11 points, which is roughly what i'd expect. more games to go.
Re: Houdini 1.02
Posted: Sun Jun 20, 2010 11:44 am
by BTO7
yanquis1972 wrote:surprised no one has mentioned houdini is now being 'officially' tested by ingo bauer --
http://www.inwoba.de/index.html
currently in the #2 spot trailing rybka by 11 points, which is roughly what i'd expect. more games to go.
Thanks for pointing this site out. Finally one testing and IPPO family. The other rating sites better wake up ...or nobody will be visiting their sites once more like this one catch on and test ALL engines.
Regards
BT
Re: Houdini 1.02
Posted: Sun Jun 20, 2010 1:37 pm
by John Blake
BTO7 wrote:yanquis1972 wrote:surprised no one has mentioned houdini is now being 'officially' tested by ingo bauer --
http://www.inwoba.de/index.html
currently in the #2 spot trailing rybka by 11 points, which is roughly what i'd expect. more games to go.
Thanks for pointing this site out. Finally one testing and IPPO family. The other rating sites better wake up ...or nobody will be visiting their sites once more like this one catch on and test ALL engines.
Regards
BT
Thats great news, lets hope this is the start of something good for computer chess.