Not logged inCSS-Forum
Forum CSS-Online Help Search Login
CSS-Shop Impressum Datenschutz
Up Topic Hauptforen / CSS-Forum / ICCR (Independent Computer Chess Rating)
- - By Sedat Canbaz Date 2011-12-29 11:37
Hello Dear Chess Friends,

I have a new idea/project for 2012:
-Let's create a new decent rating list,which will based on latest fast processors with slow time control

Here are the rules/conditions for ICCR rating

ICCR Conditions:
----------------
Processors          minimum Intel/AMD Six-Core or higher
Operating System    Windows 64-bit
Used Chess GUI      Fritz,Shredder,Arena,Aquarium,ChessGUI,ChessPatner,Winboard...
Tournament Type     Engine-Matches,Round-Robin,Swiss,Gauntlet
Time Control        45 Minutes + 15 Seconds
HashTable Size      1024 MB or higher
Ponder              OFF
Opening Book        Perfect 2012a or newer version
Book Learning       OFF
Tablebases          Nalimov/Gaviota/TripleBases...
TB Cache            128 MB or higher

More details:
-------------
Actually we are a little bit late for creating a such decent rating list,but anyway its better late than never
Yes...it's time to build a new decent rating on latest fast processors with slow time control
And there is no doubt in case of such new rating,this list will include a lot of high quality valuble games
ICCR Testers will receive oftenly new updates of my Perfect opening book series
ICCR will accept only Testers which are owners of latest fast processors:
-i7 970/i7 980X/i7 990X/i7 3930k/i7 960X/2x Intel Xeon/AMD Phenom II X6 1035T/1055T/1100T...
Note:Quads or slower processors will be allowed only for testing in case of match:single - against single processor engine
The time control (45m+15s) will be adapted on i7 980X 3.33 GHz,so e.g on AMD Phenom II X6 1035T will be played with 70m+15s
Or other example,the time control on i7 3930k @4.60 GHz will adapted to 30m+15s

All experienced Testers from all over the World are welcome to join in the new team:ICCR

ICCR rating project will be cancelled,in case of no interest between 3 months period of time

For any questions please feel free to contact:
http://www.sedatcanbaz.com/chess/contact/

Best Regards,
Sedat Canbaz
Parent - - By Michael Scheidl Date 2011-12-29 11:47
With six or more CPU cores, my first idea for (more or less) long time control ratings would be:

3 cores each, 20m+10s, ponder on.

But I do not have strong arguments for it and maybe it's just a matter of taste. Also, I could not contribute myself because my comps miss these standards, by far.

Good luck with the project anyway!
Parent - - By Sedat Canbaz Date 2011-12-29 12:12 Edited 2011-12-29 12:18
[quote="Michael Scheidl"]
With six or more CPU cores, my first idea for (more or less) long time control ratings would be:

3 cores each, 20m+10s, ponder on.

But I do not have strong arguments for it and maybe it's just a matter of taste. Also, I could not contribute myself because my comps miss these standards, by far.

Good luck with the project anyway!
[/quote]

Thank you

Actually your idea is not bad...but anyway i think my idea is better,if you are looking for better chess engine performance and for more quality games

So...i can give you another example for the current issue:
-Games played on i7 980X 3.33 GHz,6 cores, 45m+15s, Ponder OFF are approx. equal to i7 980X 3.33 GHz,6 cores, 35m+15s, Ponder ON
*I mean, the ELO performance will be approx. equal if i run Auto232 matches,e.g 45m+15s Ponder OFF against 35m+10s Ponder ON

About testing engines with 3 Cores+Ponder ON,20m+10s
-In case of such test....the games,which is planning to be played on i7 980X 3.33 GHz,6 cores, 45m+15s, Ponder OFF will be min 100 ELO stronger than i7 980X 3.33 GHz,3 Cores+Ponder ON,20m+10s

Best,
Sedat
Parent - - By Sedat Canbaz Date 2011-12-30 16:51
Hello Chess Friends,

POLL:Do you support ICCR ?
http://www.sedatcanbaz.com/chess/iccr/

Thanks in advance,
Sedat Canbaz
Parent - - By Clemens Keck Date 2011-12-30 17:28
Dear Sedat

what you mean with " support" ? Do you mean join in with hardware?
I voted for "yes" in sight of "good idea". The ICCR could be a nice new list with engines at highend level.
But I would not join in computers.
some reasons:
- I like games with ponder ON
- I like to use different books and testsuites
- I have own projetcs in work

I wish you all the best for 2012

best regards, Clemens
Parent - - By Sedat Canbaz Date 2011-12-30 18:32
[quote="Clemens Keck"]
Dear Sedat

what you mean with " support" ? Do you mean join in with hardware?
I voted for "yes" in sight of "good idea". The ICCR could be a nice new list with engines at highend level.
But I would not join in computers.
some reasons:
- I like games with ponder ON
- I like to use different books and testsuites
- I have own projetcs in work

I wish you all the best for 2012

best regards, Clemens
[/quote]

Hello dear Clemens,

Thank you for your kind words...

With 'support' = i mean that do you like my new project/idea

About pondering ON/OFF
Sure pondering ON is better,but only in this case:
-if you play via Internet,e.g Playchess server... or in Auto232 mode

But if you play on same PC:then my recommendation is Ponder OFF+maximum cores (you will produce much more quality games than using Ponder ON+1 core or half of your max.cores of your machine)
And for me its much more important:Quality than Quantity
This is just my view of course...i know very well too that all of us think and see the life differently and respect that too

And i wish you all the best in 2012 too,
Sedat
Parent - - By Sedat Canbaz Date 2012-01-02 00:55
First of all many thanks to all Voters

Wow...its true that still there are not many Voters (probably due to many people were busy with the new year holiday)

But however i did not expect that there will be a such BIG interest/support over my new ICCR project

After a such support....probably i will not cancel my new project and i will continue to work over it (even after 3 months)

Best Wishes,
Sedat



Voting Statistics:
----------------------
Yes (good idea):39 Voters
No (bad idea):6 Voters
Total:45 Voters
Parent - - By Sedat Canbaz Date 2012-01-02 20:07
Hello Chess Friends,

Just i'd like to inform you that:
-Due to many interest,requests...i will allow Quads too,but only with Hyper Threading OFF

Fore More Information:
http://www.sedatcanbaz.com/chess/iccr/

Note:Fritz Chess Benchmark will be used for measuring the processor chess speed
For accurate testing:its strongly recommended to be used the bellow time controls:

Hardware-Processor        Speed    Cores   kN/s   Time Control
---------------------------------------------------------------
2x Intel Xeon X5667     @ 4.60 GHz   12    28827   14m+15s
2x Intel Xeon X5680     @ 4.18 GHz   12    26276   17m+15s
2x Intel Xeon X5680     @ 4.00 GHZ   12    24293   20m+15s
Intel Core i7 3930k     @ 4.80 GHz    6    20356   28m+15s
Intel Core i7 980X      @ 4.60 GHz    6    19791   29m+15s
Intel Core i7 980X      @ 4.33 GHz    6    18709   32m+15s
2x Intel Core QX9775    @ 4.00 GHz    8    18063   33m+15s
Intel Core i7 970       @ 4.16 GHz    6    17302   36m+15s
Intel Core i7 2600K     @ 5.20 GHz    4    15516   42m+15s
2x Intel Core QX9775      3.20 GHz    8    14784   45m+15s
Intel Core i7 980X        3.33 GHz    6    14744   45m+15s
Intel Core i5 2500k     @ 4.90 GHz    4    14316   47m+15s
AMD Phenom II X6 1055T  @ 4.06 GHz    6    14012   48m+15s
AMD Phenom II X6 1090T  @ 3.80 GHz    6    13599   50m+15s
2x Intel Xeon E5472       3.00 GHz    8    13573   50m+15s
AMD FX X8 8150          @ 4.21 GHz    8    13331   52m+15s
Intel Core i5 760       @ 4.44 GHz    4    12912   54m+15s
Intel Core i5 2500k     @ 4.49 GHz    4    12784   55m+15s
Intel Core i7 920       @ 4.00 GHz    4    12454   57m+15s
AMD Phenom II X6 1090T  @ 3.60 GHz    6    12286   58m+15s
Intel Core i7 960       @ 4.00 GHz    4    11988   59m+15s
AMD FX X8 8150            3.60 GHz    8    11867   60m+15s
AMD Phenom II X6 1100T    3.30 GHz    6    11476   63m+15s
AMD Phenom II X6 1090T    3.20 GHz    6    11259   65m+15s
Intel Core i5 750       @ 3.80 GHz    4    11075   66m+15s
Intel Core 2 Quad Q9650 @ 3.82 GHz    4    10730   69m+15s
AMD Phenom II X6 1075T    3.00 GHz    6    10601   70m+15s
Intel Core i7 2600K       3.40 GHz    4    10473   71m+15s
Intel Core i7 920       @ 3.30 GHz    4    10350   72m+15s
Intel Core 2 Quad Q9550 @ 3.60 GHz    4    10192   73m+15s
AMD FX X8 8120            3.20 GHz    8    10019   75m+15s
AMD Phenom II X6 1055T    2.80 GHz    6    9945    75m+15s
Intel Core 2 Quad Q9550 @ 3.40 GHz    4    9469    80m+15s
AMD Phenom II X6 1035T    2.60 GHz    6    9283    82m+15s
AMD Athlon II X4 955    @ 3.91 GHz    4    9192    83m+15s
Intel Core i7 860         2.80 GHz    4    8715    88m+15s
Intel Core 2 QX9650       3.00 GHz    4    8560    90m+15s
Intel Core i5 750         2.66 GHz    4    8335    93m+15s
Intel Core 2 Quad Q9450 @ 2.80 GHz    4    7747   102m+15s
Intel Core 2 QX6700       2.66 GHz    4    7231   110m+15s
AMD Phenom II X4 940      3.00 GHz    4    6898   116m+15s
Intel Core 2 Quad Q6600   2.40 GHz    4    6771   119m+15s
AMD Phenom X4 9700        2.40 GHz    4    5725   144m+15s
2x AMD Opteron 270        2.00 GHz    4    4581   185m+15s


Calculation Method - Time Control:
Intel Core i7 980X 3.33GHz=45m+15s (official time control)
Average Game Lenght:72 moves x15 sec = 18 Minutes
TOTAL= 45 Minutes + 18 Minutes = 63 Minutes per side (45m+15s)

Intel Core i7 980X @4.33GHz=32m+15s
Average Game Lenght:72 moves x15 sec = 18 Minutes
TOTAL= 32 Minutes + 18 Minutes = 50 Minutes per side (32m+15s)

18709 kns / 14744 kns = 1,268
63 Minutes / 1.268 = 49,68 Minutes - 18 Minutes = 32m+15s
63 Minutes / 50 Minutes = 1.26

Greetings,
Sedat
Parent - - By Sedat Canbaz Date 2012-01-04 11:09
New Update:

The new official time control will be 40m+10s and it will be adapted to i7 980X 3.33 GHz
I slightly change the time control,due to i plan to add the Gladiator 2011 games:
http://www.sedatcanbaz.com/chess/scct-60m10s-gladiators/
For more accurate ELO calculation:its not allowed testing 6 core against 4 core on same machine
The reason,we plan to test the engines with adapted time controls under different speed hardwares
*Otherwise,in case of such wrong test,the Engines ELO Performance will be effected
*For right testing (6c against 4c):
-Engines should be tested with adapted time controls,on two separate machines via Auto232 mode

For more information:
http://www.sedatcanbaz.com/chess/iccr/

Best Regards,
Sedat
Parent - - By Sedat Canbaz Date 2012-01-04 23:25
I have started a new hardware speed Elo test between i7 980X @4.33GHz against i7 920 @3.3GHz

Auto232 Conditions:
------------------------
i7 980X@4.33GHz (4min+2sec)
i7 920 @3.3GHz (9min+2sec)
Windows XP x64 Prof
Hyper-Threading OFF
Large Pages OFF
Ponder ON
Perfect 2012a
512 MB Hashtable size
Gaviota TB
64 MB Cache
Auto232 mode (both PCs are connected via null-modem cable)
Engine:Houdini 2.0c x64

More details:
--------------
For right now,i7 920 @3.3GHz (9min+2sec) is leading:+4=12-3
After approx. 100-200 games,i will calculate the ELO performance between 4 cores and 6 cores ...
And in case of almost same ELO performance/strength:
-i plan to combine ICCR played game results by Quads/Six-Cores... in one chess engine version
The current used adapted time controls are calculated with the same method system as ICCR

Greetings
Sedat
Parent - - By Clemens Keck Date 2012-01-05 00:19
Dear Sedat

there is one important thing I recommend you to think well about:
ELO performance on multiple cores is NOT measurable between 4 core and 6 core computers!
The main reason is: the engines are different in their work with 2, 4, 6, 8 etc. cores.
Some benefit more from +cores, some less.
So if you do this test with houdini, you only can say you meassured Houdini 2.0c x64, and even this result will be not really correct.

regards, Clemens
Parent - By Sedat Canbaz Date 2012-01-05 02:54 Edited 2012-01-05 03:00
Thank you for your useful comments dear Clemens

A good point...

Actually this issue is too complicated...

Yes,sure...there can be X mp engine which will have worse or stronger ELO performance on 4core or 6core machines
But not for Rybka or Houdini,see the bellow SCCT Auto232 results

About your question...
Hmm...then is appearing another BIG problem-how we can calculate the all results played by 4core and 6core in ICCR rating list

And its completely wrong...if we start testing the engines on same machine-6c against 4c

And will explain you why...

Let's say we used a engine,which all cores work properly,e.g Houdini 2.0c

1)For Example on my i7 980X 3.33 Six core machine:
-Intel Core i7 980X 3.33 GHz  6 core   14744 kns   40m+10s
-Intel Core i7 980X 3.33 GHz  4 core   10490 kns   40m+10s
*Note that Time control:40m+10s (for both engines)

*Elo difference is expected to be around 25 ELO (on i7 980X @4.33GHz between 6c and 4c)

-----------------------

2) For Example,let's say other Tester played many games only by 4core engines on Quad-i7 2600k 3.40GHz:
Intel Core i7 2600K       3.40 GHz    4    10473   61m+10s

So...i am asking you which one is stronger,if there will be mach between them?
-Intel Core i7 980X 3.33 GHz  6 core   14744 kns  40m+10s
or
Intel Core i7 2600K  3.40 GHz 4 core   10473 kns  61m+10s

In my opinion,in case of similar conditions...it will be ended at equal ELO performance for both engines (i mean i7 980X 6c against i7 2600K 4c)

---------------------
Another VERY IMPORTANT question:which one is stronger,if there will be mach between them?
-Intel Core i7 980X  3.33 GHz  4 core   10490 kns   40m+10s
or
-Intel Core i7 2600K 3.40 GHz  4 core   10473 kns   61m+10s

In my opinion,in case of such match...it will be ended approx.25 ELO in favor for i7 2600K 4c

----------------------------------

Conclusion:
In case of testing the engines,e.g 6c against 4c on same PC- there will be big misunderstanding results-wrong Elo calculations
All played standings will be effected....you can not see and compare the real ELO strength of the engines
Actually the right testing (6c against 4c) on same machine will be ok,if:
- all Rating games are played with same time control and on same hardware
The problem is appearing due to adapting the time controls under many different speed processors
And after all...really i am disappointed...it seems no any right solution for the current complicated issue
But anyway, i think the current Auto232 hardware ELO method calculation is better,of course i mean for NON-buggy MP engines

*************************************

More details about Houdini and Rybka with 4 cores or 6 cores,which i believe that both engines are playing at full performance

SCCT Auto232-ELO difference:71
Rank  Name                      Elo    +    -  games  score oppo. draws
   1 Houdini 2.0b Pro x64 6c    3426   17   17  1009   69%  3307   46%
   5 Houdini 2.0b Pro x64 4c    3355   17   17  1009   51%  3350   47%

 

SCCT Auto232-ELO difference:67
Rank  Name                      Elo    +    -  games  score oppo. draws
   4 Deep Rybka 4.1 x64 6c      3367   18   18   857   62%  3301   57%
   7 Deep Rybka 4.1 x64 4c      3300   13   13  1603   47%  3321   56%


That means, i7 980X @4.33GHz is approx.70 Elo stronger than against i7 920 @3.0GHz  

    
Best,
Sedat
Up Topic Hauptforen / CSS-Forum / ICCR (Independent Computer Chess Rating)

Powered by mwForum 2.29.3 © 1999-2014 Markus Wichitill