First match results 100games on my AMD x6 1090T
TC 3m2s,both 4cores, ponder off, privat book 25 lines
A small margin win for Komodo 10.4 +16 -14 =70
A next 100 games is running
On my i5 laptop 2cores same engines, same TC, I have after 75 games
+12 -8 =55 for Houdini 5
Conclusions sofar Komodo has improved its results, it looks on par with Houdini 5. Off course more results needed to confirm
beram wrote:First match results 100games on my AMD x6 1090T
TC 3m2s,both 4cores, ponder off, privat book 25 lines
A small margin win for Komodo 10.4 +16 -14 =70
A next 100 games is running
On my i5 laptop 2cores same engines, same TC, I have after 75 games
+12 -8 =55 for Houdini 5
Conclusions sofar Komodo has improved its results, it looks on par with Houdini 5. Off course more results needed to confirm
At 2' + 1" vs. Houdini 5 I got +13 elo on four threads in 3000 games but just + 1 elo in 13,700 single thread games. I think Komodo results vs. Houdini get better with more time and/or more threads.
beram wrote:First match results 100games on my AMD x6 1090T
TC 3m2s,both 4cores, ponder off, privat book 25 lines
A small margin win for Komodo 10.4 +16 -14 =70
A next 100 games is running
On my i5 laptop 2cores same engines, same TC, I have after 75 games
+12 -8 =55 for Houdini 5
Conclusions sofar Komodo has improved its results, it looks on par with Houdini 5. Off course more results needed to confirm
At 2' + 1" vs. Houdini 5 I got +13 elo on four threads in 3000 games but just + 1 elo in 13,700 single thread games. I think Komodo results vs. Houdini get better with more time and/or more threads.
I don't think so look at CCRL there it is the other way around
Houdini 5 is just doing better with more cores and with more time(on 4cpu) against Komodo 10.3
CCRL 40/40
H5 vs K103 4cpu 57,5% (+27-9=84)
H5 vs K103 1cpu 54,8% (+9-4=39)
CCRL 4/40:
H5 vs K103 4cpu 55% (+31-17=92)
H5 vs K103 1cpu 55,4% (+169-82=562)
On fast.gm Komodo 10.3 on 1CPU is behind Houdini 5 by 40 ELO at 60 sec, 14 elo at 10 min and 7 elo à 1h.
Time controls seems to helps more Komodo than number of core.
But to be sure of that, you might want to test a match 1CPU vs 1CPU at reasonnable time control and then at the time control another match with 4CPU against 4CPU. Then you ll be sure.
beram wrote:First match results 100games on my AMD x6 1090T
TC 3m2s,both 4cores, ponder off, privat book 25 lines
A small margin win for Komodo 10.4 +16 -14 =70
A next 100 games is running
On my i5 laptop 2cores same engines, same TC, I have after 75 games
+12 -8 =55 for Houdini 5
Conclusions sofar Komodo has improved its results, it looks on par with Houdini 5. Off course more results needed to confirm
beram wrote:First match results 100games on my AMD x6 1090T
TC 3m2s,both 4cores, ponder off, privat book 25 lines
A small margin win for Komodo 10.4 +16 -14 =70
A next 100 games is running
On my i5 laptop 2cores same engines, same TC, I have after 75 games
+12 -8 =55 for Houdini 5
Conclusions sofar Komodo has improved its results, it looks on par with Houdini 5. Off course more results needed to confirm
At 2' + 1" vs. Houdini 5 I got +13 elo on four threads in 3000 games but just + 1 elo in 13,700 single thread games. I think Komodo results vs. Houdini get better with more time and/or more threads.
I don't think so look at CCRL there it is the other way around
Houdini 5 is just doing better with more cores and with more time(on 4cpu) against Komodo 10.3
CCRL 40/40
H5 vs K103 4cpu 57,5% (+27-9=84)
H5 vs K103 1cpu 54,8% (+9-4=39)
CCRL 4/40:
H5 vs K103 4cpu 55% (+31-17=92)
H5 vs K103 1cpu 55,4% (+169-82=562)
Larry quoted 3000 games and 13,700 games. The CCRL game count is not enough to draw many conclusions. 3 of the 4 matches cited were less than 100 games, with huge error margins. Plus Larry was talking about Komodo 10.4 and not 10.3. Basically, you need a lot of games to draw many conclusions unless one program is massively stronger than another.
beram wrote:First match results 100games on my AMD x6 1090T
TC 3m2s,both 4cores, ponder off, privat book 25 lines
A small margin win for Komodo 10.4 +16 -14 =70
A next 100 games is running
On my i5 laptop 2cores same engines, same TC, I have after 75 games
+12 -8 =55 for Houdini 5
Conclusions sofar Komodo has improved its results, it looks on par with Houdini 5. Off course more results needed to confirm
beram wrote:First match results 100games on my AMD x6 1090T
TC 3m2s,both 4cores, ponder off, privat book 25 lines
A small margin win for Komodo 10.4 +16 -14 =70
A next 100 games is running
On my i5 laptop 2cores same engines, same TC, I have after 75 games
+12 -8 =55 for Houdini 5
Conclusions sofar Komodo has improved its results, it looks on par with Houdini 5. Off course more results needed to confirm
At 2' + 1" vs. Houdini 5 I got +13 elo on four threads in 3000 games but just + 1 elo in 13,700 single thread games. I think Komodo results vs. Houdini get better with more time and/or more threads.
I don't think so look at CCRL there it is the other way around
Houdini 5 is just doing better with more cores and with more time(on 4cpu) against Komodo 10.3
CCRL 40/40
H5 vs K103 4cpu 57,5% (+27-9=84) total games 120
H5 vs K103 1cpu 54,8% (+9-4=39) total games 52
CCRL 4/40:
H5 vs K103 4cpu 55% (+31-17=92) total games 140
H5 vs K103 1cpu 55,4% (+169-82=562) total games 813
Larry quoted 3000 games and 13,700 games. The CCRL game count is not enough to draw many conclusions. 3 of the 4 matches cited were less than 100 games, with huge error margins. Plus Larry was talking about Komodo 10.4 and not 10.3. Basically, you need a lot of games to draw many conclusions unless one program is massively stronger than another.
Again I need to correct 3 of the 4 matches CCRL quoted where more than 100 games
There is no proof for better performance with more cores for earlier Komodo version 10.3. In fact it is the other way around.
There is no reason it will be the other way around for this new 10.4 version