Poll
Which is better hardware measuring method for chess ? (Closed)
| Best (highest) kN/s | 26 | 59% | |
| Fixed Depth | 18 | 41% |
![[tr]](/mwf/v12/flags/tr.png "Turkey")
> Expected time to a certain depth would be more useful...
For example,a such excellent benchmark list (with the highest nodes/s. at depth 22) already exist:
http://www.xtremesystems.org/forums/showthread.php?t=202139
And i don't see any reason to repeat (clone) of similar list
BTW,i prefer to work differently and i think the current R4 benchmark list is a good indicator for measuring hardware speeds
As Jean-Paul said...a lot of comments and less work
One thing more,if anybody has a different and better idea of creating perfect benchmark list,i wish him good luck and success !!!
Regards,
Sedat
The test at your link starts well, but then goes a little astray. The test should be sorted based on time to a fixed depth (expected time to depth by taking the median of N runs would be even better), rather than kn/s.
It would be fairly easy to show that this test gives more reproducible results than using your test methodology.
Note that the test I describe is only unbiased when the number of threads or processes is constant. There is an error when comparing tests with different number of threads or processes because at a constant depth, the processor using more threads or processes will generally have a wider tree. There is no easy way to normalize for this effect though...
People may be tired of this testing already, but I will start a new thread later tonight with this testing methodology and results from my computers in case anyone still cares...
It would be fairly easy to show that this test gives more reproducible results than using your test methodology.
Note that the test I describe is only unbiased when the number of threads or processes is constant. There is an error when comparing tests with different number of threads or processes because at a constant depth, the processor using more threads or processes will generally have a wider tree. There is no easy way to normalize for this effect though...
People may be tired of this testing already, but I will start a new thread later tonight with this testing methodology and results from my computers in case anyone still cares...
![[fr]](/mwf/v12/flags/fr.png "France")
> and if the engine is deterministic, even gives reproducible results. kn/s from the...
Problem is, with multi-core machines you get no reproducibility at all...
I tried Sedat's position with my dual-core (3 GHz), and obtained a range of 206 to 240 kN/s with several trials.
So much for data records!
Sure, with a multi-process or multi-threaded engine, you will get a distribution. But I claim that Sedat's method is flawed in three ways:
1) Use of kn/s is highly dependent on position and depth and even hash size. Time to depth for a fixed depth is a less noisy estimator.
2) The selected position should not include tactics near the root. The initial position is a good choice in this respect.
3) Taking several trials and picking the highest estimate is a terribly flawed approach. Looking for values on the tails of distribution curves is never a good idea because it is biased by how many trials are performed. Taking the median of multiple trials is a much better approach.
1) Use of kn/s is highly dependent on position and depth and even hash size. Time to depth for a fixed depth is a less noisy estimator.
2) The selected position should not include tactics near the root. The initial position is a good choice in this respect.
3) Taking several trials and picking the highest estimate is a terribly flawed approach. Looking for values on the tails of distribution curves is never a good idea because it is biased by how many trials are performed. Taking the median of multiple trials is a much better approach.
> But I claim that Sedat's method is flawed in three ways:
Sorry,but i don't agree with your statements...
And i strongly believe in that there is no such position where Rybka's kns each time will be same
I have tested it in other positions too (no include tactics near the root
) and the kns values are changing each time,difference is going approx 1%-15%Btw,i have no patience to see your superior method and i will be very glad if you inform about your position
And please no more comments...prove it with reality !!!
Best Regards,
Sedat Canbaz
I do not claim that kn/s will be the same. Obviously they won't be the same.
kn/ s from the same multi process or multi threaded engine will fall in a distribution. A near optimal way to estimate the "central" value of this distribution is to take the median of several trials. The worst way to generate this estimate is to take the maximum value from a number of trials.
kn/s also a poor metric for processor speed. It adds in additional random noise by not even attempting to run the trial under similar conditions. Time to depth is clearly a better metric, allowing fast a slow processors to be compared under more similar circumstances.
As I stated earlier, the opening position should be fine for this analysis since it is a quiet position and is not subject to problems caused by having engines in some trials discover tactics that are not discovered in other trials, resulting in distributions with increased sigma.
kn/ s from the same multi process or multi threaded engine will fall in a distribution. A near optimal way to estimate the "central" value of this distribution is to take the median of several trials. The worst way to generate this estimate is to take the maximum value from a number of trials.
kn/s also a poor metric for processor speed. It adds in additional random noise by not even attempting to run the trial under similar conditions. Time to depth is clearly a better metric, allowing fast a slow processors to be compared under more similar circumstances.
As I stated earlier, the opening position should be fine for this analysis since it is a quiet position and is not subject to problems caused by having engines in some trials discover tactics that are not discovered in other trials, resulting in distributions with increased sigma.
> The worst way to generate this estimate is to take the maximum value from a number of trials.
Completely agreed. But to take the maximum value is flattering!
If you want to take the median of several trials, you have to specify the "several" (5?, 10?) and then it's no longer fun...
The only undisputed result is from 1 core/thread, and a given position (start_position is OK), hashsize, and depth.
Of course, if you don't mind to have estimates +/-10%, any "specified method" will give that.
> Time to depth is clearly a better metric
Are you sure/have you tested it? (compared to kN/s when a given depth is attained)
If you want to take the median of several trials, you have to specify the "several" (5?, 10?) and then it's no longer fun...
More trials will of course result in less variance, but using the median of N trials will always be nearly optimal.
Are you sure/have you tested it? (compared to kN/s when a given depth is attained)
I've not tested this thesis, but I would be surprised if it was wrong for two reasons:
The first is that there are good reasons to believe that being at a certain depth is the best single number we have for evaluating the strength of a certain move (that is why we are running to a fixed depth after all, as opposed to running to a fixed number of evaluated nodes). So if we had a choice between an engine running on computer A that reaches depth D in an hour while calculating X kn/s and the same engine running on computer B that reaches depth D in a half hour while calculating X/2 kn/s, we should say computer B is stronger. No?
The second advantage of time to depth is that it is free of Vas' artificial corrections when comparing the results of the engine running on computers using different numbers of cores. There is still some unknown added strength associated will having a wider tree, but at least we don't need to wonder if Vas' formula is always right.
More trials will of course result in less variance, but using the median of N trials will always be nearly optimal.
Are you sure/have you tested it? (compared to kN/s when a given depth is attained)
I've not tested this thesis, but I would be surprised if it was wrong for two reasons:
The first is that there are good reasons to believe that being at a certain depth is the best single number we have for evaluating the strength of a certain move (that is why we are running to a fixed depth after all, as opposed to running to a fixed number of evaluated nodes). So if we had a choice between an engine running on computer A that reaches depth D in an hour while calculating X kn/s and the same engine running on computer B that reaches depth D in a half hour while calculating X/2 kn/s, we should say computer B is stronger. No?
The second advantage of time to depth is that it is free of Vas' artificial corrections when comparing the results of the engine running on computers using different numbers of cores. There is still some unknown added strength associated will having a wider tree, but at least we don't need to wonder if Vas' formula is always right.
> Are you sure/have you tested it? (compared to kN/s when a given depth is attained)
>
> I've not tested this thesis, but ...
Well, doesn't take long to test it (say half an hour!...)
Maybe I'll do it myself
![[de]](/mwf/v12/flags/de.png "Germany")
Hi Sedat, to what time does this test correspond? Ie. after which time is Rybka moving in a 10 min game?
Btw., the FEN is
Here are my test runs on a Q6700@2.66 GHz, 4CPU, 32 MB Hash after one minute:
321 kN/s
[-0.66] d=14 1.Rae1 Qg7 2.Qh5 Ra5 3.Qe2 a2 4.Ra1 Ng5 5.Re3 Qh7 6.Qb2 Rda8 7.Qxb6 Qf5 8.Qb3 Kh8 9.c4 Ne4 10.Qb2 Ra4 11.Qe2 (0:00:30) 9476kN
[-0.66] d=13 1.Rae1 Qg7 2.Qh5 Ra5 3.Qe2 a2 4.Ra1 Ng5 5.Re3 Qh7 6.Qb2 Rda8 7.Qxb6 Qf5 8.Qb3 Kh8 9.c4 Ne4 10.Qb2 Ra4 11.Qe2 (0:00:29) 9026kN
[-0.64] d=12 1.Qd2 Kh8 2.Qc2 Rg8 3.Qd3 Ra5 4.d5 Nf8 5.Re3 Qg7 (0:00:09) 3027kN
[-0.80] d=12 1.h4 Ra7 2.Qc1 Rda8 3.Qc2 a2 4.c4 Kg7 5.d5 Qc5 6.Re4 b5 7.Rg4 (0:00:09) 3027kN
[-0.69] d=11 1.h4 Ra7 2.Qc1 Rda8 3.Qc2 a2 4.c4 Kg7 5.d5 Qc5 6.Qc3 (0:00:05) 1814kN
[-0.64] d=10 1.Qd2 Rc8 2.Re3 Qc4 3.Qd1 Ra7 4.Qb1 b5 5.Rg3 Kh8 (0:00:02) 832kN
[-0.64] d=9 1.Qd2 Kh8 2.Qe3 Ra7 3.Qc1 (0:00:02) 659kN
[-0.67] d=9 1.h4 a2 2.Qe3 Ra7 3.Qd2 Rda8 4.d5 Ra3 5.Rc6 Qe5 6.Rxb6 Qxc3 7.Rxa2 (0:00:01) 608kN
[-0.68] d=9 1.Qg6 Kh8 2.Qd3 Ra5 3.Ra2 f5 4.Qe3 Rg8 5.Re7 Qd6 (0:00:01) 461kN
[-0.13] d=8 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Kxg8 5.Qe8 Nf8 6.Re7 Qh6 7.Qf7 (0:00:00) 175kN
[-0.26] d=8 1.Qg6 Kh8 2.Rae1 Qg7 3.Qd3 Ra7 4.Rxb6 a2 5.Ra1 Rg8 6.Qf3 Ng5 7.Qxf6 Qxf6 (0:00:00) 158kN
[-0.41] d=8 1.Qg6 Kh8 2.Rae1 Qg7 3.Qd3 Ra7 4.Rxb6 a2 5.Ra1 Rg8 6.Qf3 Ng5 7.Qxf6 Qxf6 (0:00:00) 158kN
[-0.53] d=7 1.Qg6 Kh8 2.Rae1 Qg7 3.Qd3 Ra7 4.Rxb6 a2 5.Ra1 Rg8 6.Qf3 Ng5 7.Qxf6 (0:00:00) 102kN
[-0.24] d=7 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Kxg8 5.Qe8 Nf8 6.Re7 Qg6 7.Qb5 (0:00:00) 26kN
[-0.36] d=6 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Kxg8 5.Qe8 Nf8 6.Re7 Qg6 (0:00:00) 10kN
[-0.27] d=5 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Qxg8 (0:00:00) 4kN
[-0.29] d=4 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 (0:00:00) 3kN
[-0.09] d=3 1.Qg6 Qg7 2.Qd3 Ra5 3.Rxb6 a2 4.c4 (0:00:00) 0kN
[+0.00] d=2 1.Qg6 Qg7 2.Qd3 Qb7 3.Qg6 Qg7 4.Qd3 Qb7 5.Qg6 Qg7 6.Qd3 Qb7 7.Qg6 Qg7 8.Qd3 Qb7 9.Qg6 Qg7 10.Qd3 Qb7 11.Qg6 Qg7 12.Qd3 Qb7 13.Qg6 Qg7 14.Qd3 Qb7 15.Qg6 Qg7 16.Qd3 (0:00:00) 0kN
[-0.02] d=1 1.Qg6 Qg7 2.Qd3 Qb7 3.h4 (0:00:00) 0kN
321kN/s
[-0.76] d=15 1.Qc1 a2 2.Qc2 Rda8 3.d5 R6a7 4.Qg6 Kh8 5.Qd3 Qd8 6.c4 Ng5 7.Re2 Qd6 8.Qd4 Re7 (0:00:43) 14055kN
[-0.69] d=14 1.Qc1 a2 2.Qc2 Rda8 3.d5 R6a7 4.Qg6 Kh8 5.Qd3 Qd8 (0:00:17) 5549kN
[-0.58] d=13 1.Qc1 a2 2.Qc2 Rda8 3.d5 R6a7 4.Qg6 Kh8 5.Qd3 Qd8 (0:00:16) 5325kN
[-0.68] d=13 1.Rae1 Qg7 2.Qh5 Ra5 3.Qe2 a2 4.Ra1 Rda8 5.Rxb6 Ng5 6.Qc4 Kh8 7.Qd3 Qh7 (0:00:16) 5318kN
[-0.76] d=13 1.Qd2 Kh8 2.Qc2 Rg8 3.Qd3 Ra4 4.d5 Qg7 5.g4 Ng5 6.Rxb6 (0:00:10) 3127kN
[-0.69] d=12 1.Qd2 Kh8 2.Qc2 Rg8 3.Qd3 Ra4 4.d5 Qg7 5.g4 Ng5 6.Rxb6 f5 7.Kf1 Ne4 (0:00:03) 1059kN
[-0.65] d=11 1.Qd2 Kh8 2.Qc2 Rg8 3.Qd3 Ra4 4.d5 Ng5 5.Rxf6 Qh7 6.Qxh7 Kxh7 7.Rxb6 Nxh3 8.Kf1 (0:00:02) 674kN
[-0.60] d=10 1.Qd2 Kh8 2.Qd3 Ra5 3.Ra2 f5 4.Rh6 Kg7 5.Qd2 Rc8 (0:00:01) 375kN
[-0.60] d=9 1.Qd2 Kh8 2.Qd3 Ra5 3.Ra2 f5 4.Rh6 Kg7 5.Qd2 Rc8 (0:00:00) 267kN
[-0.63] d=9 1.Qg6 Kh8 2.Qc2 Rg8 3.Qd3 Ra4 4.Ra2 Nf8 5.Re1 Ng6 6.Qc2 (0:00:00) 231kN
[-0.59] d=8 1.Qg6 Kh8 2.Qc2 Rg8 3.h4 Nf8 4.Re3 Qg7 5.g3 Qd7 6.Qe2 (0:00:00) 137kN
[-0.13] d=7 1.Qg6 Qg7 2.Qd3 Rda8 3.Re3 Ra5 4.Rg3 Rg5 5.Qe4 Raa5 6.Qe6 (0:00:00) 17kN
[-0.13] d=6 1.Qg6 Qg7 2.Qd3 Rda8 3.Re3 Ra5 4.Rg3 Rg5 5.Qe4 Raa5 6.Qe6 (0:00:00) 15kN
[-0.05] d=5 1.Qg6 Qg7 2.Qd3 Ra5 3.Rxb6 Qc7 4.Rb3 (0:00:00) 1kN
[-0.04] d=4 1.Qg6 Qg7 2.Qd3 Ra5 3.Rxb6 Qc7 4.Re6 a2 5.Ree1 (0:00:00) 0kN
[+0.00] d=3 1.Qg6 Qg7 2.Qd3 Qb7 3.Qg6 Qg7 4.Qd3 Qb7 5.Qg6 Qg7 6.Qd3 Qb7 7.Qg6 Qg7 8.Qd3 Qb7 9.Qg6 Qg7 10.Qd3 Qb7 11.Qg6 Qg7 12.Qd3 Qb7 13.Qg6 Qg7 14.Qd3 Qb7 15.Qg6 Qg7 16.Qd3 (0:00:00) 0kN
[-0.12] d=3 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 (0:00:00) 0kN
[-0.24] d=2 1.Qg6 Kh8 2.Qd3 Ra5 3.Ra2 Kg8 4.Ra1 (0:00:00) 0kN
[-0.08] d=1 1.Qg6 Qg7 2.Qd3 Qb7 3.Qg6 (0:00:00) 0kN
[-0.08] d=1 1.Qg6 Qg7 2.Qd3 Qb7 3.Qg6 (0:00:00) 0kN
317 kN/s
[-0.75] d=16 1.Rae1 Qg7 2.Qh5 Ra7 3.Rxb6 a2 4.Ra1 Qc7 5.Rb2 Qxc3 6.Raxa2 Rxa2 7.Rxa2 Qxd4 8.Re2 Nf8 9.Kh2 Qd6 10.g3 Kg7 11.Re4 Ng6 (0:00:55) 17740kN
[-0.67] d=15 1.Rae1 Qg7 2.Qh5 Ra7 3.Rxb6 a2 4.Ra1 Qc7 5.Rb2 Qxc3 6.Raxa2 Rxa2 7.Rxa2 Qxd4 8.Re2 Nf8 9.Re1 Rd7 10.Qh6 Re7 (0:00:40) 12308kN
[-0.62] d=14 1.Rae1 Qg7 2.Qh5 a2 3.Re7 a1Q 4.Rxg7 Kxg7 5.Qg4 Ng5 6.Rxa1 Rxa1 7.Kh2 Rc1 8.f4 (0:00:34) 10600kN
[-0.76] d=14 1.Qd2 Kh8 2.Qd3 Ra5 3.Re3 a2 4.Re2 Rda8 5.d5 Qd6 6.c4 Ra3 7.Qd2 (0:00:21) 6167kN
[-0.68] d=13 1.Qd2 Kh8 2.Qd3 Ra5 3.Re3 a2 4.Re2 Rda8 5.d5 Qd6 6.c4 Ra3 7.Qd4 (0:00:13) 3860kN
[-0.72] d=12 1.Qd2 Kh8 2.Qd3 Ra5 3.Re3 b5 4.Qd2 Raa8 5.h4 Rg8 6.d5 Qc4 (0:00:05) 1510kN
[-0.58] d=11 1.Qd2 Kh8 2.Qc2 Rg8 3.Qd3 Ra5 4.c4 a2 5.Re2 Rga8 6.d5 Ra3 (0:00:03) 861kN
[-0.58] d=10 1.Qg6 Kh8 2.Qc2 Rg8 3.Qd3 Ra5 4.c4 a2 5.Re2 Rga8 6.d5 Ra3 (0:00:01) 282kN
[-0.64] d=9 1.Qg6 Kh8 2.Qc2 Rg8 3.Qd3 Ra5 4.c4 a2 5.Qb3 Qg7 6.g4 Ng5 7.Rxb6 f5 8.Rxa2 Rxa2 9.Qxa2 Nxh3 10.Kh2 (0:00:00) 152kN
[-0.63] d=8 1.Qg6 Kh8 2.Qc2 Rg8 3.h4 Nf8 4.Ree1 Ng6 5.Qf5 Nf4 6.g3 Qd6 7.h5 (0:00:00) 85kN
[-0.19] d=7 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Kxg8 5.Qe8 Nf8 6.Re7 Qh6 7.Qf7 Kh8 8.Re8 Qc1 9.Kh2 Qf4 10.Kg1 Qc1 11.Kh2 Qf4 12.Kg1 Qc1 13.Kh2 Qf4 14.Kg1 Qc1 15.Kh2 Qf4 16.Kg1 (0:00:00) 19kN
[-0.18] d=7 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Kxg8 5.Qe8 Nf8 6.Re7 Qh6 7.Qf7 (0:00:00) 19kN
[-0.30] d=6 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Qe4 Ra7 (0:00:00) 6kN
[-0.17] d=5 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Kxg8 5.Qe8 Nf8 6.Re7 (0:00:00) 3kN
[-0.04] d=4 1.Qg6 Qg7 2.Qd3 Ra5 3.Rxb6 Qc7 4.Re6 a2 5.Ree1 (0:00:00) 1kN
[-0.11] d=3 1.Qg6 Qg7 2.Qd3 Ra5 3.Rxb6 a2 4.Re6 (0:00:00) 0kN
[+0.03] d=3 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 (0:00:00) 0kN
[-0.12] d=3 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 (0:00:00) 0kN
[-0.24] d=2 1.Qg6 Kh8 2.Qd3 Ra5 3.Ra2 Kg8 (0:00:00) 0kN
[-0.08] d=1 1.Qg6 Qg7 2.Qd3 Qb7 3.Qg6 (0:00:00) 0kN
Btw., the FEN is
3r2k1/2q4n/rp2Rp1Q/8/3P4/p1P4P/5PP1/R5K1 w - -
Here are my test runs on a Q6700@2.66 GHz, 4CPU, 32 MB Hash after one minute:
321 kN/s
[-0.66] d=14 1.Rae1 Qg7 2.Qh5 Ra5 3.Qe2 a2 4.Ra1 Ng5 5.Re3 Qh7 6.Qb2 Rda8 7.Qxb6 Qf5 8.Qb3 Kh8 9.c4 Ne4 10.Qb2 Ra4 11.Qe2 (0:00:30) 9476kN
[-0.66] d=13 1.Rae1 Qg7 2.Qh5 Ra5 3.Qe2 a2 4.Ra1 Ng5 5.Re3 Qh7 6.Qb2 Rda8 7.Qxb6 Qf5 8.Qb3 Kh8 9.c4 Ne4 10.Qb2 Ra4 11.Qe2 (0:00:29) 9026kN
[-0.64] d=12 1.Qd2 Kh8 2.Qc2 Rg8 3.Qd3 Ra5 4.d5 Nf8 5.Re3 Qg7 (0:00:09) 3027kN
[-0.80] d=12 1.h4 Ra7 2.Qc1 Rda8 3.Qc2 a2 4.c4 Kg7 5.d5 Qc5 6.Re4 b5 7.Rg4 (0:00:09) 3027kN
[-0.69] d=11 1.h4 Ra7 2.Qc1 Rda8 3.Qc2 a2 4.c4 Kg7 5.d5 Qc5 6.Qc3 (0:00:05) 1814kN
[-0.64] d=10 1.Qd2 Rc8 2.Re3 Qc4 3.Qd1 Ra7 4.Qb1 b5 5.Rg3 Kh8 (0:00:02) 832kN
[-0.64] d=9 1.Qd2 Kh8 2.Qe3 Ra7 3.Qc1 (0:00:02) 659kN
[-0.67] d=9 1.h4 a2 2.Qe3 Ra7 3.Qd2 Rda8 4.d5 Ra3 5.Rc6 Qe5 6.Rxb6 Qxc3 7.Rxa2 (0:00:01) 608kN
[-0.68] d=9 1.Qg6 Kh8 2.Qd3 Ra5 3.Ra2 f5 4.Qe3 Rg8 5.Re7 Qd6 (0:00:01) 461kN
[-0.13] d=8 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Kxg8 5.Qe8 Nf8 6.Re7 Qh6 7.Qf7 (0:00:00) 175kN
[-0.26] d=8 1.Qg6 Kh8 2.Rae1 Qg7 3.Qd3 Ra7 4.Rxb6 a2 5.Ra1 Rg8 6.Qf3 Ng5 7.Qxf6 Qxf6 (0:00:00) 158kN
[-0.41] d=8 1.Qg6 Kh8 2.Rae1 Qg7 3.Qd3 Ra7 4.Rxb6 a2 5.Ra1 Rg8 6.Qf3 Ng5 7.Qxf6 Qxf6 (0:00:00) 158kN
[-0.53] d=7 1.Qg6 Kh8 2.Rae1 Qg7 3.Qd3 Ra7 4.Rxb6 a2 5.Ra1 Rg8 6.Qf3 Ng5 7.Qxf6 (0:00:00) 102kN
[-0.24] d=7 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Kxg8 5.Qe8 Nf8 6.Re7 Qg6 7.Qb5 (0:00:00) 26kN
[-0.36] d=6 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Kxg8 5.Qe8 Nf8 6.Re7 Qg6 (0:00:00) 10kN
[-0.27] d=5 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Qxg8 (0:00:00) 4kN
[-0.29] d=4 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 (0:00:00) 3kN
[-0.09] d=3 1.Qg6 Qg7 2.Qd3 Ra5 3.Rxb6 a2 4.c4 (0:00:00) 0kN
[+0.00] d=2 1.Qg6 Qg7 2.Qd3 Qb7 3.Qg6 Qg7 4.Qd3 Qb7 5.Qg6 Qg7 6.Qd3 Qb7 7.Qg6 Qg7 8.Qd3 Qb7 9.Qg6 Qg7 10.Qd3 Qb7 11.Qg6 Qg7 12.Qd3 Qb7 13.Qg6 Qg7 14.Qd3 Qb7 15.Qg6 Qg7 16.Qd3 (0:00:00) 0kN
[-0.02] d=1 1.Qg6 Qg7 2.Qd3 Qb7 3.h4 (0:00:00) 0kN
321kN/s
[-0.76] d=15 1.Qc1 a2 2.Qc2 Rda8 3.d5 R6a7 4.Qg6 Kh8 5.Qd3 Qd8 6.c4 Ng5 7.Re2 Qd6 8.Qd4 Re7 (0:00:43) 14055kN
[-0.69] d=14 1.Qc1 a2 2.Qc2 Rda8 3.d5 R6a7 4.Qg6 Kh8 5.Qd3 Qd8 (0:00:17) 5549kN
[-0.58] d=13 1.Qc1 a2 2.Qc2 Rda8 3.d5 R6a7 4.Qg6 Kh8 5.Qd3 Qd8 (0:00:16) 5325kN
[-0.68] d=13 1.Rae1 Qg7 2.Qh5 Ra5 3.Qe2 a2 4.Ra1 Rda8 5.Rxb6 Ng5 6.Qc4 Kh8 7.Qd3 Qh7 (0:00:16) 5318kN
[-0.76] d=13 1.Qd2 Kh8 2.Qc2 Rg8 3.Qd3 Ra4 4.d5 Qg7 5.g4 Ng5 6.Rxb6 (0:00:10) 3127kN
[-0.69] d=12 1.Qd2 Kh8 2.Qc2 Rg8 3.Qd3 Ra4 4.d5 Qg7 5.g4 Ng5 6.Rxb6 f5 7.Kf1 Ne4 (0:00:03) 1059kN
[-0.65] d=11 1.Qd2 Kh8 2.Qc2 Rg8 3.Qd3 Ra4 4.d5 Ng5 5.Rxf6 Qh7 6.Qxh7 Kxh7 7.Rxb6 Nxh3 8.Kf1 (0:00:02) 674kN
[-0.60] d=10 1.Qd2 Kh8 2.Qd3 Ra5 3.Ra2 f5 4.Rh6 Kg7 5.Qd2 Rc8 (0:00:01) 375kN
[-0.60] d=9 1.Qd2 Kh8 2.Qd3 Ra5 3.Ra2 f5 4.Rh6 Kg7 5.Qd2 Rc8 (0:00:00) 267kN
[-0.63] d=9 1.Qg6 Kh8 2.Qc2 Rg8 3.Qd3 Ra4 4.Ra2 Nf8 5.Re1 Ng6 6.Qc2 (0:00:00) 231kN
[-0.59] d=8 1.Qg6 Kh8 2.Qc2 Rg8 3.h4 Nf8 4.Re3 Qg7 5.g3 Qd7 6.Qe2 (0:00:00) 137kN
[-0.13] d=7 1.Qg6 Qg7 2.Qd3 Rda8 3.Re3 Ra5 4.Rg3 Rg5 5.Qe4 Raa5 6.Qe6 (0:00:00) 17kN
[-0.13] d=6 1.Qg6 Qg7 2.Qd3 Rda8 3.Re3 Ra5 4.Rg3 Rg5 5.Qe4 Raa5 6.Qe6 (0:00:00) 15kN
[-0.05] d=5 1.Qg6 Qg7 2.Qd3 Ra5 3.Rxb6 Qc7 4.Rb3 (0:00:00) 1kN
[-0.04] d=4 1.Qg6 Qg7 2.Qd3 Ra5 3.Rxb6 Qc7 4.Re6 a2 5.Ree1 (0:00:00) 0kN
[+0.00] d=3 1.Qg6 Qg7 2.Qd3 Qb7 3.Qg6 Qg7 4.Qd3 Qb7 5.Qg6 Qg7 6.Qd3 Qb7 7.Qg6 Qg7 8.Qd3 Qb7 9.Qg6 Qg7 10.Qd3 Qb7 11.Qg6 Qg7 12.Qd3 Qb7 13.Qg6 Qg7 14.Qd3 Qb7 15.Qg6 Qg7 16.Qd3 (0:00:00) 0kN
[-0.12] d=3 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 (0:00:00) 0kN
[-0.24] d=2 1.Qg6 Kh8 2.Qd3 Ra5 3.Ra2 Kg8 4.Ra1 (0:00:00) 0kN
[-0.08] d=1 1.Qg6 Qg7 2.Qd3 Qb7 3.Qg6 (0:00:00) 0kN
[-0.08] d=1 1.Qg6 Qg7 2.Qd3 Qb7 3.Qg6 (0:00:00) 0kN
317 kN/s
[-0.75] d=16 1.Rae1 Qg7 2.Qh5 Ra7 3.Rxb6 a2 4.Ra1 Qc7 5.Rb2 Qxc3 6.Raxa2 Rxa2 7.Rxa2 Qxd4 8.Re2 Nf8 9.Kh2 Qd6 10.g3 Kg7 11.Re4 Ng6 (0:00:55) 17740kN
[-0.67] d=15 1.Rae1 Qg7 2.Qh5 Ra7 3.Rxb6 a2 4.Ra1 Qc7 5.Rb2 Qxc3 6.Raxa2 Rxa2 7.Rxa2 Qxd4 8.Re2 Nf8 9.Re1 Rd7 10.Qh6 Re7 (0:00:40) 12308kN
[-0.62] d=14 1.Rae1 Qg7 2.Qh5 a2 3.Re7 a1Q 4.Rxg7 Kxg7 5.Qg4 Ng5 6.Rxa1 Rxa1 7.Kh2 Rc1 8.f4 (0:00:34) 10600kN
[-0.76] d=14 1.Qd2 Kh8 2.Qd3 Ra5 3.Re3 a2 4.Re2 Rda8 5.d5 Qd6 6.c4 Ra3 7.Qd2 (0:00:21) 6167kN
[-0.68] d=13 1.Qd2 Kh8 2.Qd3 Ra5 3.Re3 a2 4.Re2 Rda8 5.d5 Qd6 6.c4 Ra3 7.Qd4 (0:00:13) 3860kN
[-0.72] d=12 1.Qd2 Kh8 2.Qd3 Ra5 3.Re3 b5 4.Qd2 Raa8 5.h4 Rg8 6.d5 Qc4 (0:00:05) 1510kN
[-0.58] d=11 1.Qd2 Kh8 2.Qc2 Rg8 3.Qd3 Ra5 4.c4 a2 5.Re2 Rga8 6.d5 Ra3 (0:00:03) 861kN
[-0.58] d=10 1.Qg6 Kh8 2.Qc2 Rg8 3.Qd3 Ra5 4.c4 a2 5.Re2 Rga8 6.d5 Ra3 (0:00:01) 282kN
[-0.64] d=9 1.Qg6 Kh8 2.Qc2 Rg8 3.Qd3 Ra5 4.c4 a2 5.Qb3 Qg7 6.g4 Ng5 7.Rxb6 f5 8.Rxa2 Rxa2 9.Qxa2 Nxh3 10.Kh2 (0:00:00) 152kN
[-0.63] d=8 1.Qg6 Kh8 2.Qc2 Rg8 3.h4 Nf8 4.Ree1 Ng6 5.Qf5 Nf4 6.g3 Qd6 7.h5 (0:00:00) 85kN
[-0.19] d=7 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Kxg8 5.Qe8 Nf8 6.Re7 Qh6 7.Qf7 Kh8 8.Re8 Qc1 9.Kh2 Qf4 10.Kg1 Qc1 11.Kh2 Qf4 12.Kg1 Qc1 13.Kh2 Qf4 14.Kg1 Qc1 15.Kh2 Qf4 16.Kg1 (0:00:00) 19kN
[-0.18] d=7 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Kxg8 5.Qe8 Nf8 6.Re7 Qh6 7.Qf7 (0:00:00) 19kN
[-0.30] d=6 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Qe4 Ra7 (0:00:00) 6kN
[-0.17] d=5 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Kxg8 5.Qe8 Nf8 6.Re7 (0:00:00) 3kN
[-0.04] d=4 1.Qg6 Qg7 2.Qd3 Ra5 3.Rxb6 Qc7 4.Re6 a2 5.Ree1 (0:00:00) 1kN
[-0.11] d=3 1.Qg6 Qg7 2.Qd3 Ra5 3.Rxb6 a2 4.Re6 (0:00:00) 0kN
[+0.03] d=3 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 (0:00:00) 0kN
[-0.12] d=3 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 (0:00:00) 0kN
[-0.24] d=2 1.Qg6 Kh8 2.Qd3 Ra5 3.Ra2 Kg8 (0:00:00) 0kN
[-0.08] d=1 1.Qg6 Qg7 2.Qd3 Qb7 3.Qg6 (0:00:00) 0kN
> Hi Sedat, to what time does this test correspond? Ie. after which time is Rybka moving in a 10 min game?
Hello Felix,
Many thanks for your bench...in my testings almost each time i see different moving in time
For example,here are available all Rybka 4 benchaMarks's screenshots and you can check R4's moving in time:
http://sedatchess.110mb.com/games/r4bench.zip
BTW,we have almost same results,comparing your hardware with mine
And (in the next web update) your results will be added too
Best,
Sedat
for a reliable benchmark you should maybe use one position and request kN/s at e.g. depth=15. the kN/s get higher with higher depth, so if you compare different depths you don't get good results.
>depth=15
Yes...your idea is good,but in my testing kns values are almost same,e.g in depth 14,depth 15,depth 16....
Thats why i recommend:
1.Run the benchmark minimum 3-4 times (send the highest Kn/s of your results)
And it does not mean that e.g in depth 15,your R4 bench kns results will be highest
One thing more:if we set in depth 15,for slower machines-yes,but what about for faster machines,its will be hard to catch 15 depth with screenshot
Lets say the bench condition is set to depth 18,this time the slower machines will take many time to reach that depth (minutes or maybe hour)
Remember also:in slower time controls,Rybka 4 is starting to use the tablebeases and that hurts the Kn/s performance
*Solution for that,during benchmarking the tablebases should be off (with a such conditions our R4 bench will be become much more complicated)
Best,
Sedat
Hi,
Folks may not so interested in benchmarks, I posted some when building Pauls machine as we knew it would be of interest. I am content knowing I would be number 1 with a 6 core machine at least for 24 hours Benchmark speed figures are also not always the same as what people run there machines normally 24/7 ... To get the very best figures you also need to turn off as much stuff you have running in the background as poss
http://valid.canardpc.com/show_oc.php?id=1378774
Regards
Folks may not so interested in benchmarks, I posted some when building Pauls machine as we knew it would be of interest. I am content knowing I would be number 1 with a 6 core machine at least for 24 hours
Benchmark speed figures are also not always the same as what people run there machines normally 24/7 ... To get the very best figures you also need to turn off as much stuff you have running in the background as posshttp://valid.canardpc.com/show_oc.php?id=1378774
Regards
Hello Chess Friends !
First of all I would like to thank to:
---------------------------------
Tsvetan Milchev
Pavel Hase
Servet Kultur
Felix King
Jesse Gersenson
Last Updated Results:
---------------------
Intel Core i7 920
Intel Core i7 860
Intel Core Q6600
Intel Quad Q6700
Intel Core i3 M370
Intel Core 2 Duo P8400
For Deep Rypka 4 BenchMark List:
http://sedatchess.110mb.com/index.php?p=1_72
Kind Regards,
Sedat Canbaz
First of all I would like to thank to:
---------------------------------
Tsvetan Milchev
Pavel Hase
Servet Kultur
Felix King
Jesse Gersenson
Last Updated Results:
---------------------
Intel Core i7 920
Intel Core i7 860
Intel Core Q6600
Intel Quad Q6700
Intel Core i3 M370
Intel Core 2 Duo P8400
For Deep Rypka 4 BenchMark List:
http://sedatchess.110mb.com/index.php?p=1_72
Kind Regards,
Sedat Canbaz
I now have results for large pages enabled (353 kN/s):
336 kN/s
[-0.64] d=16 1.Rae1 Qg7 2.Qh5 Ra7 3.Rxb6 a2 4.Ra1 Qc7 5.Rb3 Qc6 6.h4 Nf8 7.Qf5 Qd5 8.Qc2 Rda8 9.Qb2 Rg7 10.g3 (0:00:46) 15624kN
[-0.64] d=15 1.Rae1 Qg7 2.Qh5 Ra7 3.Rxb6 a2 4.Ra1 Qc7 5.Rb3 Qc6 6.h4 Nf8 7.Qf5 Qd5 8.Qc2 Rda8 9.Qb2 Rg7 10.g3 (0:00:41) 13997kN
[-0.50] d=14 1.Rae1 Qg7 2.Qh5 Ra7 3.Rxb6 a2 4.Ra1 Qc7 5.Rb3 Qc6 6.f4 Qd5 7.Qg6 Kh8 8.Qc2 Rda8 9.Qb2 Kg7 10.Rb5 Qe4 (0:00:09) 2778kN
[-0.50] d=13 1.Rae1 Qg7 2.Qh5 Ra7 3.Rxb6 a2 4.Ra1 Qc7 5.Rb3 Qc6 6.f4 Qd5 7.Qg6 Kh8 8.Qc2 Rda8 9.Qb2 Kg7 10.Rb5 Qe4 (0:00:07) 1922kN
[-0.49] d=13 1.Rae1 Qg7 2.Qh5 Ra7 3.Rxb6 a2 4.Ra1 Qc7 5.Rb3 Qc6 6.f4 Qd5 7.Qg6 Rg7 8.Qc2 (0:00:07) 1801kN
[-0.61] d=12 1.Rae1 Qg7 2.Qh5 Ra7 3.Rxb6 a2 4.Ra1 Qc7 5.Rb2 Qxc3 6.Raxa2 Rxa2 7.Rxa2 Qxd4 8.Qg6 Kh8 9.Ra3 Qe5 10.Qc2 Ng5 11.Re3 Qa1 (0:00:03) 832kN
[-0.57] d=11 1.Rae1 Qg7 2.Qh5 Ra7 3.Rxb6 a2 4.Ra1 Qc7 5.Rb3 Qc6 6.Qf5 Ng5 7.Rb2 Rda8 8.Qc2 (0:00:03) 766kN
[-0.52] d=11 1.Rae1 Qg7 2.Qh5 a2 3.Re7 Qxe7 4.Rxe7 (0:00:02) 731kN
[-0.64] d=11 1.Qc1 a2 2.Qc2 Kh8 3.Qd2 Ra3 4.Rxa2 Rxa2 5.Qxa2 Qxc3 6.d5 Ng5 7.Re7 Qc1 8.Kh2 Qf4 9.Kg1 Rc8 10.Re1 Rc1 (0:00:01) 509kN
[-0.64] d=10 1.Qc1 a2 2.Qc2 Rda8 3.d5 R6a7 4.Qg6 Kh8 5.Qd3 Rd8 6.Re2 Ra5 7.c4 (0:00:01) 374kN
[-0.54] d=9 1.Qc1 Qf7 2.Qe3 a2 3.Qe2 Rda8 4.h4 Qg6 5.Qc4 Kh8 6.Ree1 Ra3 7.d5 Rg8 8.g3 (0:00:01) 338kN
[-0.66] d=9 1.Re2 Qf7 2.Rea2 Rda8 3.Qe3 b5 4.Qd3 Qc4 5.Qc2 Ng5 6.Qg6 (0:00:00) 229kN
[-0.52] d=8 1.Qg6 Kh8 2.Rae1 Qg7 3.Re8 Rxe8 4.Rxe8 Nf8 5.Qh5 Kg8 6.Qb5 Qa7 7.Qd5 Kg7 8.Qe4 (0:00:00) 99kN
[-0.30] d=7 1.Qg6 Kh8 2.Rae1 Qg7 3.Re8 Rxe8 4.Rxe8 Nf8 5.Qh5 Kg8 6.Qb5 Ra7 (0:00:00) 10kN
[-0.31] d=6 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Kxg8 5.Qe8 Nf8 6.Re7 Qg6 (0:00:00) 3kN
[-0.22] d=5 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Qxg8 5.Re8 (0:00:00) 0kN
[-0.22] d=4 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Qxg8 5.Re8 (0:00:00) 0kN
[-0.22] d=3 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Qxg8 (0:00:00) 0kN
[-0.14] d=2 1.Qg6 Kh8 2.Qd3 Qb7 3.Rae1 Rg8 (0:00:00) 0kN
[+0.25] d=1 1.Qg6 Kh8 2.Qd3 (0:00:00) 0kN
330 kN/s
[-0.55] d=15 1.Rae1 Qg7 2.Qh5 Ra5 3.Qe2 a2 4.Ra1 Ng5 5.Re3 Qg6 6.Qb2 Rda8 7.Qxb6 Qc2 8.Qxf6 Rf8 9.Qb6 Qxf2 10.Kh1 Qf5 11.c4 Rf6 12.Qd8 Kg7 13.Ree1 Ra4 14.Qe8 Qc2 (0:00:28) 9137kN
[-0.55] d=14 1.Rae1 Qg7 2.Qh5 Ra5 3.Qe2 a2 4.Ra1 Ng5 5.Re3 Qg6 6.Qb2 Rda8 7.Qxb6 Qc2 8.Qxf6 Rf8 9.Qb6 Qxf2 10.Kh1 Qf5 11.c4 Rf6 12.Qd8 Kg7 13.Ree1 Ra4 14.Qe8 Qc2 (0:00:24) 7744kN
[-0.52] d=13 1.Rae1 Qg7 2.Qh5 Ra5 3.Qe2 a2 4.Ra1 Ng5 5.Re3 Qg6 6.Qb2 Rda8 7.Qxb6 Qc2 8.Qxf6 Rf8 9.Qb6 Qxf2 10.Kh1 Qf5 11.c4 Rf6 12.Qd8 Kg7 13.Ree1 Nf7 14.Qc7 Ra8 (0:00:22) 6974kN
[-0.58] d=13 1.Rae1 Qg7 2.Qh5 Ra5 3.Qe2 a2 4.Ra1 Ng5 5.Re3 Qg6 6.Qb2 Rda8 7.Qxb6 Kh8 8.c4 Ne4 9.d5 (0:00:20) 6226kN
[-0.69] d=13 1.Qd2 Kh8 2.Qd3 Ra7 3.d5 Qd7 4.Re4 Ng5 5.Re3 b5 6.Qd4 Nh7 7.c4 bxc4 (0:00:17) 5102kN
[-0.70] d=12 1.Qd2 Kh8 2.Qd3 Ra7 3.d5 Qc5 4.c4 b5 5.cxb5 Rxd5 6.Qg3 Rg5 7.Qd3 Rf5 8.b6 Qxf2 (0:00:05) 1235kN
[-0.63] d=11 1.Qd2 Kh8 2.Qd3 Ra7 3.c4 Ra4 4.Rxa3 Rxc4 5.Ra1 Rcxd4 6.Qb3 Rd2 (0:00:02) 389kN
[-0.60] d=10 1.Qd2 Kh8 2.Qd3 Ra7 3.Re3 a2 4.Re2 Rda8 5.d5 b5 6.d6 Qc4 7.Qd2 Ng5 8.Re7 Rxe7 9.dxe7 Nf7 10.Rxa2 (0:00:01) 311kN
[-0.65] d=9 1.Qd2 Rc8 2.Qa2 Kh8 3.Re3 f5 4.Qe6 f4 5.Rf3 Rg8 6.Kh2 (0:00:00) 95kN
[-0.65] d=8 1.Qd2 Rc8 2.Qa2 Kh8 3.Re3 f5 4.Qe6 f4 5.Rf3 Rg8 6.Kh2 (0:00:00) 88kN
[-0.67] d=8 1.Qg6 Kh8 2.Re2 Ra7 3.Qd3 Rc8 4.Re3 a2 5.Qd2 Qc4 6.Qb2 b5 (0:00:00) 88kN
[-0.06] d=7 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Kxg8 5.Qe8 Nf8 6.Re7 Qg6 7.Qb5 a2 (0:00:00) 35kN
[-0.28] d=7 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Kxg8 5.Qe8 Nf8 6.Re7 (0:00:00) 25kN
[-0.43] d=7 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Kxg8 5.Qe8 Nf8 6.Re7 (0:00:00) 25kN
[-0.55] d=6 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Kxg8 5.Qe8 Nf8 6.Qb5 Qa7 7.h4 (0:00:00) 15kN
[-0.32] d=5 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Kxg8 5.Re8 (0:00:00) 4kN
[+0.00] d=4 1.Qg6 Qg7 2.Qd3 Qb7 3.Qg6 Qg7 4.Qd3 Qb7 5.Qg6 Qg7 6.Qd3 Qb7 7.Qg6 Qg7 8.Qd3 Qb7 9.Qg6 Qg7 10.Qd3 Qb7 11.Qg6 Qg7 12.Qd3 Qb7 13.Qg6 Qg7 14.Qd3 Qb7 15.Qg6 Qg7 16.Qd3 (0:00:00) 1kN
[+0.00] d=3 1.Qg6 Qg7 2.Qd3 Qb7 3.Qg6 Qg7 4.Qd3 Qb7 5.Qg6 Qg7 6.Qd3 Qb7 7.Qg6 Qg7 8.Qd3 Qb7 9.Qg6 Qg7 10.Qd3 Qb7 11.Qg6 Qg7 12.Qd3 Qb7 13.Qg6 Qg7 14.Qd3 Qb7 15.Qg6 Qg7 16.Qd3 (0:00:00) 1kN
[+0.06] d=2 1.Qg6 Qg7 2.Qd3 Ra5 (0:00:00) 0kN
[+0.12] d=1 1.Qg6 Kh8 2.Qd3 (0:00:00) 0kN
312 kN/s
[-0.73] d=15 1.Re4 a2 2.Qg6 Qg7 3.Qf5 Qd7 4.Qg6 Kh8 5.Rh4 Rf8 6.c4 Ra4 7.g3 Qc7 8.Qe4 Rfa8 9.Kg2 R8a7 10.Qf5 (0:01:10) 21913kN
[-0.66] d=14 1.Rae1 Qg7 2.Qh5 Ra5 3.Qe2 a2 4.Ra1 Ng5 5.Re3 Qg6 6.Qb2 Rda8 7.Qxb6 Kh8 8.Qb2 Ne4 (0:00:14) 4429kN
[-0.66] d=13 1.Rae1 Qg7 2.Qh5 Ra5 3.Qe2 a2 4.Ra1 Ng5 5.Re3 Qg6 6.Qb2 Rda8 7.Qxb6 Kh8 8.Qb2 Ne4 (0:00:14) 4315kN
[-0.68] d=13 1.Re4 a2 2.Qg6 Qg7 3.Qf5 Qd7 4.Qg6 Kh8 5.Rh4 Rf8 6.c4 Ra3 7.Qc2 Rfa8 8.Qg6 Qe7 9.Qf5 Rc3 10.d5 Ra7 (0:00:11) 3169kN
[-0.68] d=12 1.Re4 a2 2.Qg6 Qg7 3.Qf5 Qd7 4.Qg6 Kh8 5.Rh4 Rf8 6.c4 Ra3 7.Qc2 Rfa8 8.Qg6 Qe7 9.Qf5 Rc3 10.d5 Ra7 11.Re4 (0:00:07) 1763kN
[-0.64] d=11 1.Re4 a2 2.Qg6 Qg7 3.Qf5 Qf7 4.Rg4 Kh8 5.Rh4 Re8 6.c4 Ra3 7.d5 Re2 8.d6 Rd2 9.c5 (0:00:04) 1144kN
[-0.50] d=10 1.Qc1 a2 2.Qc2 Kh8 3.Rxa2 Rxa2 4.Qxa2 Qxc3 5.Re7 Qc1 6.Kh2 Qf4 7.Kg1 Rc8 8.g3 Qf3 9.Qb1 Ng5 10.h4 Qa3 (0:00:01) 447kN
[-0.53] d=9 1.Qg6 Kh8 2.Qc2 a2 3.Rxa2 Rxa2 4.Qxa2 Qxc3 5.Re7 Qc1 6.Kh2 Qf4 7.Kg1 Ng5 8.Qe2 Qc1 9.Kh2 Qf4 10.Kg1 Qc1 11.Kh2 Qf4 12.Kg1 Qc1 13.Kh2 Qf4 14.Kg1 Qc1 15.Kh2 Qf4 16.Kg1 (0:00:01) 391kN
[-0.57] d=8 1.Qc1 a2 2.Qc2 Rda8 3.d5 Qf7 4.Qf5 Ra5 5.h4 Re8 6.Qg4 Kh8 7.c4 Rxe6 (0:00:00) 227kN
[-0.61] d=8 1.Qg6 Kh8 2.Qd3 Ra7 3.Re3 Qd6 4.Qc4 Nf8 5.Qb3 Kg7 6.Ra2 (0:00:00) 179kN
[-0.05] d=7 1.Qg6 Qg7 2.Qd3 Rda8 3.d5 Ng5 4.Ree1 a2 5.d6 (0:00:00) 11kN
[-0.05] d=6 1.Qg6 Qg7 2.Qd3 Rda8 3.d5 Ng5 4.Ree1 a2 5.d6 (0:00:00) 10kN
[-0.13] d=6 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 (0:00:00) 3kN
[-0.25] d=5 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 (0:00:00) 2kN
[-0.05] d=4 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 (0:00:00) 1kN
[-0.09] d=4 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 (0:00:00) 1kN
[-0.21] d=3 1.Qg6 Kh8 2.Qd3 Ra7 3.d5 (0:00:00) 0kN
[-0.14] d=2 1.Qg6 Kh8 2.Qd3 Qb7 3.Rae1 Rg8 (0:00:00) 0kN
[+0.24] d=1 1.Qg6 Kh8 2.Qd3 (0:00:00) 0kN
[+0.07] d=1 1.Qg6 Kh8 2.Qc2 Rxd4 (0:00:00) 0kN
339 kN/s
[-0.68] d=16 1.Rae1 Qg7 2.Qh5 Ra5 3.Qe2 a2 4.Ra1 Ng5 5.Re3 Qg6 6.Qb2 Rda8 7.Qxb6 Qc2 8.Qc6 Rb8 9.h4 Qh7 10.Re8 Rxe8 11.Qxe8 Kg7 12.Qb8 Nf7 13.Qb4 Rh5 (0:00:49) 16720kN
[-0.68] d=15 1.Rae1 Qg7 2.Qh5 Ra5 3.Qe2 a2 4.Ra1 Ng5 5.Re3 Qg6 6.Qb2 Rda8 7.Qxb6 Qc2 8.Qc6 Rb8 9.h4 Qh7 10.Re8 Rxe8 11.Qxe8 Kg7 12.Qb8 Nf7 13.Qb4 Rh5 (0:00:40) 13055kN
[-0.63] d=14 1.Rae1 Qg7 2.Qh5 Ra5 3.Qe2 a2 4.Ra1 Ng5 5.Re3 Qg6 6.Qb2 Rda8 7.Qxb6 Kh8 8.Qb7 R5a7 (0:00:30) 9892kN
[-0.68] d=14 1.Qc1 Rda8 2.Qe3 R6a7 3.d5 Qc5 4.Qd3 Rc8 5.Re3 a2 6.d6 Rd8 7.Qg6 Rg7 (0:00:21) 6798kN
[-0.57] d=13 1.Qc1 a2 2.Qc2 Rda8 3.d5 R6a7 4.d6 Qf7 5.Qf5 Kf8 (0:00:11) 3182kN
[-0.56] d=12 1.Qc1 a2 2.Qc2 Rda8 3.d5 R6a7 4.d6 Qf7 5.Ree1 Ng5 6.h4 Ne6 7.Qb2 Ra3 (0:00:09) 2802kN
[-0.65] d=12 1.Qd2 Kh8 2.Qc2 Rg8 3.Qd3 Ra5 4.c4 a2 5.Re2 Rga8 6.d5 Ra3 7.d6 Qc5 8.Qd2 (0:00:05) 1406kN
[-0.64] d=11 1.Qd2 Kh8 2.Qd3 Ra7 3.Re3 Nf8 4.d5 a2 5.c4 Qc5 6.Re4 b5 7.Rh4 Kg7 (0:00:04) 1176kN
[-0.65] d=11 1.Qg6 Kh8 2.Qd3 Ra7 3.Re3 Nf8 4.d5 a2 5.c4 Qc5 6.Re4 b5 7.Rh4 Kg7 (0:00:04) 1176kN
[-0.64] d=10 1.Qg6 Kh8 2.Qd3 Ra7 3.Re3 Qd6 4.Qc4 Nf8 5.Qb3 Rda8 6.Re2 Kg7 7.g3 (0:00:02) 367kN
[-0.38] d=9 1.Qg6 Kh8 2.Rae1 Qg7 3.Re8 Rxe8 4.Qxe8 Nf8 5.Qe2 Ra7 6.d5 (0:00:00) 68kN
[-0.50] d=8 1.Qg6 Kh8 2.Rae1 Qg7 3.Qd3 Rg8 4.g4 Ra7 5.Rxb6 a2 6.Ra1 Ng5 7.f4 (0:00:00) 55kN
[-0.21] d=7 1.Qg6 Qg7 2.Qd3 Rda8 3.Ree1 Ng5 4.d5 Qh7 5.Qd4 Qf5 6.d6 Ra4 7.Qxb6 Ne4 (0:00:00) 16kN
[-0.11] d=7 1.Qg6 Qg7 2.Qd3 Rda8 3.Ree1 Ng5 4.d5 Qh7 5.Qd4 Qf5 6.d6 (0:00:00) 14kN
[-0.23] d=6 1.Qg6 Qg7 2.Qc2 Ng5 3.Re3 Qf7 4.h4 Nh7 5.Rg3 (0:00:00) 5kN
[-0.03] d=6 1.Qg6 Qg7 2.Qd3 Rda8 3.Ree1 a2 4.d5 Ng5 5.d6 Qh7 6.Qb5 Qc2 (0:00:00) 4kN
[-0.15] d=5 1.Qg6 Qg7 2.Qd3 Rda8 3.Qc4 Qf7 4.Qe2 (0:00:00) 2kN
[+0.00] d=4 1.Qg6 Qg7 2.Qd3 Qb7 3.Re7 (0:00:00) 0kN
[+0.00] d=3 1.Qg6 Qg7 2.Qd3 Qb7 3.Re7 (0:00:00) 0kN
[+0.00] d=2 1.Qg6 Qg7 2.Qd3 Qb7 3.Qg6 Qg7 4.Qd3 Qb7 5.Qg6 Qg7 6.Qd3 Qb7 7.Qg6 Qg7 8.Qd3 Qb7 9.Qg6 Qg7 10.Qd3 Qb7 11.Qg6 Qg7 12.Qd3 Qb7 13.Qg6 Qg7 14.Qd3 Qb7 15.Qg6 Qg7 16.Qd3 (0:00:00) 0kN
353 kN/s
[-0.65] d=15 1.Rae1 Qg7 2.Qh5 Ra5 3.Qe2 a2 4.Ra1 Ng5 5.Re3 Qg6 6.Qb2 Rda8 7.Qxb6 Kh8 8.c4 Ra4 9.h4 Ne4 10.Qc6 Nd2 11.c5 Rg8 12.h5 Qb1 (0:00:33) 11206kN
[-0.60] d=14 1.Rae1 Qg7 2.Qh5 Ra5 3.Qe2 a2 4.Ra1 Ng5 5.Re3 Qg6 6.Qb2 Rda8 7.Qxb6 Kh8 8.c4 Qf5 9.Qb2 Ra4 10.h4 Ne4 11.c5 Qd5 (0:00:24) 8031kN
[-0.55] d=13 1.Rae1 Qg7 2.Qh5 Ra5 3.Qe2 a2 4.Ra1 Ng5 5.Re3 Qh7 6.Qb2 Qa7 7.h4 Nf7 8.Re2 (0:00:19) 6624kN
[-0.63] d=12 1.Qd2 Kh8 2.Qd3 Ra7 3.d5 Qf7 4.Qf5 Qd7 5.d6 a2 6.Qd5 (0:00:06) 2248kN
[-0.54] d=11 1.Qc1 a2 2.h4 Kh8 3.Qd2 Rda8 4.d5 Qc4 5.Re7 R6a7 6.Rxa7 Rxa7 7.d6 Nf8 8.Qh6 Kg8 9.Qxf6 Qe4 10.Kh2 (0:00:02) 834kN
[-0.58] d=10 1.Qc1 a2 2.Qc2 Ra3 3.Rxa2 Rxc3 4.Qg6 Kh8 5.Re1 Rg8 6.Qf5 Qg7 7.g3 Ng5 8.Re3 Rc1 9.Kh2 (0:00:01) 454kN
[-0.61] d=10 1.Qg6 Kh8 2.Qd3 Ra7 3.d5 Qc5 4.c4 Ra4 5.Re4 Ng5 6.Rh4 Kg7 (0:00:00) 210kN
[-0.55] d=9 1.Qg6 Kh8 2.Qd3 Ra7 3.d5 Qc5 4.c4 b5 5.cxb5 Rxd5 6.Qg6 Ra8 7.Rae1 Rg8 8.Re8 (0:00:00) 84kN
[-0.55] d=8 1.Qg6 Kh8 2.Qd3 Ra7 3.d5 Qc5 4.c4 b5 5.cxb5 Rxd5 6.Qg6 Ra8 7.Rae1 Rg8 8.Re8 (0:00:00) 49kN
[-0.44] d=7 1.Qg6 Kh8 2.Rae1 Qg7 3.Qe4 Ra7 4.Rxb6 a2 5.Ra1 Rg8 6.d5 (0:00:00) 19kN
[-0.24] d=7 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Kxg8 5.Qe8 Nf8 6.Re7 (0:00:00) 11kN
[-0.36] d=6 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Kxg8 5.Qf5 (0:00:00) 10kN
[-0.15] d=5 1.Qg6 Qg7 2.Qd3 Rda8 3.Qc4 Qf7 4.Qe2 (0:00:00) 3kN
[+0.12] d=5 1.Qg6 Qg7 2.Qd3 Qb7 3.Qg6 Qg7 4.Qd3 Qb7 5.Qg6 Qg7 6.Qd3 Qb7 7.Qg6 Qg7 8.Qd3 Qb7 9.Qg6 Qg7 10.Qd3 Qb7 11.Qg6 Qg7 12.Qd3 Qb7 13.Qg6 Qg7 14.Qd3 Qb7 15.Qg6 Qg7 16.Qd3 (0:00:00) 1kN
[+0.00] d=4 1.Qg6 Qg7 2.Qd3 Qb7 3.Qg6 Qg7 4.Qd3 Qb7 5.Qg6 Qg7 6.Qd3 Qb7 7.Qg6 Qg7 8.Qd3 Qb7 9.Qg6 Qg7 10.Qd3 Qb7 11.Qg6 Qg7 12.Qd3 Qb7 13.Qg6 Qg7 14.Qd3 Qb7 15.Qg6 Qg7 16.Qd3 (0:00:00) 0kN
[+0.00] d=3 1.Qg6 Qg7 2.Qd3 Qb7 3.Qg6 Qg7 4.Qd3 Qb7 5.Qg6 Qg7 6.Qd3 Qb7 7.Qg6 Qg7 8.Qd3 Qb7 9.Qg6 Qg7 10.Qd3 Qb7 11.Qg6 Qg7 12.Qd3 Qb7 13.Qg6 Qg7 14.Qd3 Qb7 15.Qg6 Qg7 16.Qd3 (0:00:00) 0kN
[+0.00] d=2 1.Qg6 Qg7 2.Qd3 Qb7 3.Qg6 Qg7 4.Qd3 Qb7 5.Qg6 Qg7 6.Qd3 Qb7 7.Qg6 Qg7 8.Qd3 Qb7 9.Qg6 Qg7 10.Qd3 Qb7 11.Qg6 Qg7 12.Qd3 Qb7 13.Qg6 Qg7 14.Qd3 Qb7 15.Qg6 Qg7 16.Qd3 (0:00:00) 0kN
[-0.04] d=1 1.Qg6 Qg7 2.Qd3 Qb7 3.Qg6 Qg7 4.Qd3 Qb7 5.Qg6 Qg7 6.Qd3 Qb7 7.Qg6 Qg7 8.Qd3 Qb7 9.Qg6 Qg7 10.Qd3 Qb7 11.Qg6 Qg7 12.Qd3 Qb7 13.Qg6 Qg7 14.Qd3 Qb7 15.Qg6 Qg7 16.Qd3 (0:00:00) 0kN
[-0.04] d=1 1.Qg6 Qg7 2.Qd3 Qb7 3.Qg6 (0:00:00) 0kN
[-0.47] d=1 1.Qe3 a2 2.Qd3 (0:00:00) 0kN
336 kN/s
[-0.64] d=16 1.Rae1 Qg7 2.Qh5 Ra7 3.Rxb6 a2 4.Ra1 Qc7 5.Rb3 Qc6 6.h4 Nf8 7.Qf5 Qd5 8.Qc2 Rda8 9.Qb2 Rg7 10.g3 (0:00:46) 15624kN
[-0.64] d=15 1.Rae1 Qg7 2.Qh5 Ra7 3.Rxb6 a2 4.Ra1 Qc7 5.Rb3 Qc6 6.h4 Nf8 7.Qf5 Qd5 8.Qc2 Rda8 9.Qb2 Rg7 10.g3 (0:00:41) 13997kN
[-0.50] d=14 1.Rae1 Qg7 2.Qh5 Ra7 3.Rxb6 a2 4.Ra1 Qc7 5.Rb3 Qc6 6.f4 Qd5 7.Qg6 Kh8 8.Qc2 Rda8 9.Qb2 Kg7 10.Rb5 Qe4 (0:00:09) 2778kN
[-0.50] d=13 1.Rae1 Qg7 2.Qh5 Ra7 3.Rxb6 a2 4.Ra1 Qc7 5.Rb3 Qc6 6.f4 Qd5 7.Qg6 Kh8 8.Qc2 Rda8 9.Qb2 Kg7 10.Rb5 Qe4 (0:00:07) 1922kN
[-0.49] d=13 1.Rae1 Qg7 2.Qh5 Ra7 3.Rxb6 a2 4.Ra1 Qc7 5.Rb3 Qc6 6.f4 Qd5 7.Qg6 Rg7 8.Qc2 (0:00:07) 1801kN
[-0.61] d=12 1.Rae1 Qg7 2.Qh5 Ra7 3.Rxb6 a2 4.Ra1 Qc7 5.Rb2 Qxc3 6.Raxa2 Rxa2 7.Rxa2 Qxd4 8.Qg6 Kh8 9.Ra3 Qe5 10.Qc2 Ng5 11.Re3 Qa1 (0:00:03) 832kN
[-0.57] d=11 1.Rae1 Qg7 2.Qh5 Ra7 3.Rxb6 a2 4.Ra1 Qc7 5.Rb3 Qc6 6.Qf5 Ng5 7.Rb2 Rda8 8.Qc2 (0:00:03) 766kN
[-0.52] d=11 1.Rae1 Qg7 2.Qh5 a2 3.Re7 Qxe7 4.Rxe7 (0:00:02) 731kN
[-0.64] d=11 1.Qc1 a2 2.Qc2 Kh8 3.Qd2 Ra3 4.Rxa2 Rxa2 5.Qxa2 Qxc3 6.d5 Ng5 7.Re7 Qc1 8.Kh2 Qf4 9.Kg1 Rc8 10.Re1 Rc1 (0:00:01) 509kN
[-0.64] d=10 1.Qc1 a2 2.Qc2 Rda8 3.d5 R6a7 4.Qg6 Kh8 5.Qd3 Rd8 6.Re2 Ra5 7.c4 (0:00:01) 374kN
[-0.54] d=9 1.Qc1 Qf7 2.Qe3 a2 3.Qe2 Rda8 4.h4 Qg6 5.Qc4 Kh8 6.Ree1 Ra3 7.d5 Rg8 8.g3 (0:00:01) 338kN
[-0.66] d=9 1.Re2 Qf7 2.Rea2 Rda8 3.Qe3 b5 4.Qd3 Qc4 5.Qc2 Ng5 6.Qg6 (0:00:00) 229kN
[-0.52] d=8 1.Qg6 Kh8 2.Rae1 Qg7 3.Re8 Rxe8 4.Rxe8 Nf8 5.Qh5 Kg8 6.Qb5 Qa7 7.Qd5 Kg7 8.Qe4 (0:00:00) 99kN
[-0.30] d=7 1.Qg6 Kh8 2.Rae1 Qg7 3.Re8 Rxe8 4.Rxe8 Nf8 5.Qh5 Kg8 6.Qb5 Ra7 (0:00:00) 10kN
[-0.31] d=6 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Kxg8 5.Qe8 Nf8 6.Re7 Qg6 (0:00:00) 3kN
[-0.22] d=5 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Qxg8 5.Re8 (0:00:00) 0kN
[-0.22] d=4 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Qxg8 5.Re8 (0:00:00) 0kN
[-0.22] d=3 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Qxg8 (0:00:00) 0kN
[-0.14] d=2 1.Qg6 Kh8 2.Qd3 Qb7 3.Rae1 Rg8 (0:00:00) 0kN
[+0.25] d=1 1.Qg6 Kh8 2.Qd3 (0:00:00) 0kN
330 kN/s
[-0.55] d=15 1.Rae1 Qg7 2.Qh5 Ra5 3.Qe2 a2 4.Ra1 Ng5 5.Re3 Qg6 6.Qb2 Rda8 7.Qxb6 Qc2 8.Qxf6 Rf8 9.Qb6 Qxf2 10.Kh1 Qf5 11.c4 Rf6 12.Qd8 Kg7 13.Ree1 Ra4 14.Qe8 Qc2 (0:00:28) 9137kN
[-0.55] d=14 1.Rae1 Qg7 2.Qh5 Ra5 3.Qe2 a2 4.Ra1 Ng5 5.Re3 Qg6 6.Qb2 Rda8 7.Qxb6 Qc2 8.Qxf6 Rf8 9.Qb6 Qxf2 10.Kh1 Qf5 11.c4 Rf6 12.Qd8 Kg7 13.Ree1 Ra4 14.Qe8 Qc2 (0:00:24) 7744kN
[-0.52] d=13 1.Rae1 Qg7 2.Qh5 Ra5 3.Qe2 a2 4.Ra1 Ng5 5.Re3 Qg6 6.Qb2 Rda8 7.Qxb6 Qc2 8.Qxf6 Rf8 9.Qb6 Qxf2 10.Kh1 Qf5 11.c4 Rf6 12.Qd8 Kg7 13.Ree1 Nf7 14.Qc7 Ra8 (0:00:22) 6974kN
[-0.58] d=13 1.Rae1 Qg7 2.Qh5 Ra5 3.Qe2 a2 4.Ra1 Ng5 5.Re3 Qg6 6.Qb2 Rda8 7.Qxb6 Kh8 8.c4 Ne4 9.d5 (0:00:20) 6226kN
[-0.69] d=13 1.Qd2 Kh8 2.Qd3 Ra7 3.d5 Qd7 4.Re4 Ng5 5.Re3 b5 6.Qd4 Nh7 7.c4 bxc4 (0:00:17) 5102kN
[-0.70] d=12 1.Qd2 Kh8 2.Qd3 Ra7 3.d5 Qc5 4.c4 b5 5.cxb5 Rxd5 6.Qg3 Rg5 7.Qd3 Rf5 8.b6 Qxf2 (0:00:05) 1235kN
[-0.63] d=11 1.Qd2 Kh8 2.Qd3 Ra7 3.c4 Ra4 4.Rxa3 Rxc4 5.Ra1 Rcxd4 6.Qb3 Rd2 (0:00:02) 389kN
[-0.60] d=10 1.Qd2 Kh8 2.Qd3 Ra7 3.Re3 a2 4.Re2 Rda8 5.d5 b5 6.d6 Qc4 7.Qd2 Ng5 8.Re7 Rxe7 9.dxe7 Nf7 10.Rxa2 (0:00:01) 311kN
[-0.65] d=9 1.Qd2 Rc8 2.Qa2 Kh8 3.Re3 f5 4.Qe6 f4 5.Rf3 Rg8 6.Kh2 (0:00:00) 95kN
[-0.65] d=8 1.Qd2 Rc8 2.Qa2 Kh8 3.Re3 f5 4.Qe6 f4 5.Rf3 Rg8 6.Kh2 (0:00:00) 88kN
[-0.67] d=8 1.Qg6 Kh8 2.Re2 Ra7 3.Qd3 Rc8 4.Re3 a2 5.Qd2 Qc4 6.Qb2 b5 (0:00:00) 88kN
[-0.06] d=7 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Kxg8 5.Qe8 Nf8 6.Re7 Qg6 7.Qb5 a2 (0:00:00) 35kN
[-0.28] d=7 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Kxg8 5.Qe8 Nf8 6.Re7 (0:00:00) 25kN
[-0.43] d=7 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Kxg8 5.Qe8 Nf8 6.Re7 (0:00:00) 25kN
[-0.55] d=6 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Kxg8 5.Qe8 Nf8 6.Qb5 Qa7 7.h4 (0:00:00) 15kN
[-0.32] d=5 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Kxg8 5.Re8 (0:00:00) 4kN
[+0.00] d=4 1.Qg6 Qg7 2.Qd3 Qb7 3.Qg6 Qg7 4.Qd3 Qb7 5.Qg6 Qg7 6.Qd3 Qb7 7.Qg6 Qg7 8.Qd3 Qb7 9.Qg6 Qg7 10.Qd3 Qb7 11.Qg6 Qg7 12.Qd3 Qb7 13.Qg6 Qg7 14.Qd3 Qb7 15.Qg6 Qg7 16.Qd3 (0:00:00) 1kN
[+0.00] d=3 1.Qg6 Qg7 2.Qd3 Qb7 3.Qg6 Qg7 4.Qd3 Qb7 5.Qg6 Qg7 6.Qd3 Qb7 7.Qg6 Qg7 8.Qd3 Qb7 9.Qg6 Qg7 10.Qd3 Qb7 11.Qg6 Qg7 12.Qd3 Qb7 13.Qg6 Qg7 14.Qd3 Qb7 15.Qg6 Qg7 16.Qd3 (0:00:00) 1kN
[+0.06] d=2 1.Qg6 Qg7 2.Qd3 Ra5 (0:00:00) 0kN
[+0.12] d=1 1.Qg6 Kh8 2.Qd3 (0:00:00) 0kN
312 kN/s
[-0.73] d=15 1.Re4 a2 2.Qg6 Qg7 3.Qf5 Qd7 4.Qg6 Kh8 5.Rh4 Rf8 6.c4 Ra4 7.g3 Qc7 8.Qe4 Rfa8 9.Kg2 R8a7 10.Qf5 (0:01:10) 21913kN
[-0.66] d=14 1.Rae1 Qg7 2.Qh5 Ra5 3.Qe2 a2 4.Ra1 Ng5 5.Re3 Qg6 6.Qb2 Rda8 7.Qxb6 Kh8 8.Qb2 Ne4 (0:00:14) 4429kN
[-0.66] d=13 1.Rae1 Qg7 2.Qh5 Ra5 3.Qe2 a2 4.Ra1 Ng5 5.Re3 Qg6 6.Qb2 Rda8 7.Qxb6 Kh8 8.Qb2 Ne4 (0:00:14) 4315kN
[-0.68] d=13 1.Re4 a2 2.Qg6 Qg7 3.Qf5 Qd7 4.Qg6 Kh8 5.Rh4 Rf8 6.c4 Ra3 7.Qc2 Rfa8 8.Qg6 Qe7 9.Qf5 Rc3 10.d5 Ra7 (0:00:11) 3169kN
[-0.68] d=12 1.Re4 a2 2.Qg6 Qg7 3.Qf5 Qd7 4.Qg6 Kh8 5.Rh4 Rf8 6.c4 Ra3 7.Qc2 Rfa8 8.Qg6 Qe7 9.Qf5 Rc3 10.d5 Ra7 11.Re4 (0:00:07) 1763kN
[-0.64] d=11 1.Re4 a2 2.Qg6 Qg7 3.Qf5 Qf7 4.Rg4 Kh8 5.Rh4 Re8 6.c4 Ra3 7.d5 Re2 8.d6 Rd2 9.c5 (0:00:04) 1144kN
[-0.50] d=10 1.Qc1 a2 2.Qc2 Kh8 3.Rxa2 Rxa2 4.Qxa2 Qxc3 5.Re7 Qc1 6.Kh2 Qf4 7.Kg1 Rc8 8.g3 Qf3 9.Qb1 Ng5 10.h4 Qa3 (0:00:01) 447kN
[-0.53] d=9 1.Qg6 Kh8 2.Qc2 a2 3.Rxa2 Rxa2 4.Qxa2 Qxc3 5.Re7 Qc1 6.Kh2 Qf4 7.Kg1 Ng5 8.Qe2 Qc1 9.Kh2 Qf4 10.Kg1 Qc1 11.Kh2 Qf4 12.Kg1 Qc1 13.Kh2 Qf4 14.Kg1 Qc1 15.Kh2 Qf4 16.Kg1 (0:00:01) 391kN
[-0.57] d=8 1.Qc1 a2 2.Qc2 Rda8 3.d5 Qf7 4.Qf5 Ra5 5.h4 Re8 6.Qg4 Kh8 7.c4 Rxe6 (0:00:00) 227kN
[-0.61] d=8 1.Qg6 Kh8 2.Qd3 Ra7 3.Re3 Qd6 4.Qc4 Nf8 5.Qb3 Kg7 6.Ra2 (0:00:00) 179kN
[-0.05] d=7 1.Qg6 Qg7 2.Qd3 Rda8 3.d5 Ng5 4.Ree1 a2 5.d6 (0:00:00) 11kN
[-0.05] d=6 1.Qg6 Qg7 2.Qd3 Rda8 3.d5 Ng5 4.Ree1 a2 5.d6 (0:00:00) 10kN
[-0.13] d=6 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 (0:00:00) 3kN
[-0.25] d=5 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 (0:00:00) 2kN
[-0.05] d=4 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 (0:00:00) 1kN
[-0.09] d=4 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 (0:00:00) 1kN
[-0.21] d=3 1.Qg6 Kh8 2.Qd3 Ra7 3.d5 (0:00:00) 0kN
[-0.14] d=2 1.Qg6 Kh8 2.Qd3 Qb7 3.Rae1 Rg8 (0:00:00) 0kN
[+0.24] d=1 1.Qg6 Kh8 2.Qd3 (0:00:00) 0kN
[+0.07] d=1 1.Qg6 Kh8 2.Qc2 Rxd4 (0:00:00) 0kN
339 kN/s
[-0.68] d=16 1.Rae1 Qg7 2.Qh5 Ra5 3.Qe2 a2 4.Ra1 Ng5 5.Re3 Qg6 6.Qb2 Rda8 7.Qxb6 Qc2 8.Qc6 Rb8 9.h4 Qh7 10.Re8 Rxe8 11.Qxe8 Kg7 12.Qb8 Nf7 13.Qb4 Rh5 (0:00:49) 16720kN
[-0.68] d=15 1.Rae1 Qg7 2.Qh5 Ra5 3.Qe2 a2 4.Ra1 Ng5 5.Re3 Qg6 6.Qb2 Rda8 7.Qxb6 Qc2 8.Qc6 Rb8 9.h4 Qh7 10.Re8 Rxe8 11.Qxe8 Kg7 12.Qb8 Nf7 13.Qb4 Rh5 (0:00:40) 13055kN
[-0.63] d=14 1.Rae1 Qg7 2.Qh5 Ra5 3.Qe2 a2 4.Ra1 Ng5 5.Re3 Qg6 6.Qb2 Rda8 7.Qxb6 Kh8 8.Qb7 R5a7 (0:00:30) 9892kN
[-0.68] d=14 1.Qc1 Rda8 2.Qe3 R6a7 3.d5 Qc5 4.Qd3 Rc8 5.Re3 a2 6.d6 Rd8 7.Qg6 Rg7 (0:00:21) 6798kN
[-0.57] d=13 1.Qc1 a2 2.Qc2 Rda8 3.d5 R6a7 4.d6 Qf7 5.Qf5 Kf8 (0:00:11) 3182kN
[-0.56] d=12 1.Qc1 a2 2.Qc2 Rda8 3.d5 R6a7 4.d6 Qf7 5.Ree1 Ng5 6.h4 Ne6 7.Qb2 Ra3 (0:00:09) 2802kN
[-0.65] d=12 1.Qd2 Kh8 2.Qc2 Rg8 3.Qd3 Ra5 4.c4 a2 5.Re2 Rga8 6.d5 Ra3 7.d6 Qc5 8.Qd2 (0:00:05) 1406kN
[-0.64] d=11 1.Qd2 Kh8 2.Qd3 Ra7 3.Re3 Nf8 4.d5 a2 5.c4 Qc5 6.Re4 b5 7.Rh4 Kg7 (0:00:04) 1176kN
[-0.65] d=11 1.Qg6 Kh8 2.Qd3 Ra7 3.Re3 Nf8 4.d5 a2 5.c4 Qc5 6.Re4 b5 7.Rh4 Kg7 (0:00:04) 1176kN
[-0.64] d=10 1.Qg6 Kh8 2.Qd3 Ra7 3.Re3 Qd6 4.Qc4 Nf8 5.Qb3 Rda8 6.Re2 Kg7 7.g3 (0:00:02) 367kN
[-0.38] d=9 1.Qg6 Kh8 2.Rae1 Qg7 3.Re8 Rxe8 4.Qxe8 Nf8 5.Qe2 Ra7 6.d5 (0:00:00) 68kN
[-0.50] d=8 1.Qg6 Kh8 2.Rae1 Qg7 3.Qd3 Rg8 4.g4 Ra7 5.Rxb6 a2 6.Ra1 Ng5 7.f4 (0:00:00) 55kN
[-0.21] d=7 1.Qg6 Qg7 2.Qd3 Rda8 3.Ree1 Ng5 4.d5 Qh7 5.Qd4 Qf5 6.d6 Ra4 7.Qxb6 Ne4 (0:00:00) 16kN
[-0.11] d=7 1.Qg6 Qg7 2.Qd3 Rda8 3.Ree1 Ng5 4.d5 Qh7 5.Qd4 Qf5 6.d6 (0:00:00) 14kN
[-0.23] d=6 1.Qg6 Qg7 2.Qc2 Ng5 3.Re3 Qf7 4.h4 Nh7 5.Rg3 (0:00:00) 5kN
[-0.03] d=6 1.Qg6 Qg7 2.Qd3 Rda8 3.Ree1 a2 4.d5 Ng5 5.d6 Qh7 6.Qb5 Qc2 (0:00:00) 4kN
[-0.15] d=5 1.Qg6 Qg7 2.Qd3 Rda8 3.Qc4 Qf7 4.Qe2 (0:00:00) 2kN
[+0.00] d=4 1.Qg6 Qg7 2.Qd3 Qb7 3.Re7 (0:00:00) 0kN
[+0.00] d=3 1.Qg6 Qg7 2.Qd3 Qb7 3.Re7 (0:00:00) 0kN
[+0.00] d=2 1.Qg6 Qg7 2.Qd3 Qb7 3.Qg6 Qg7 4.Qd3 Qb7 5.Qg6 Qg7 6.Qd3 Qb7 7.Qg6 Qg7 8.Qd3 Qb7 9.Qg6 Qg7 10.Qd3 Qb7 11.Qg6 Qg7 12.Qd3 Qb7 13.Qg6 Qg7 14.Qd3 Qb7 15.Qg6 Qg7 16.Qd3 (0:00:00) 0kN
353 kN/s
[-0.65] d=15 1.Rae1 Qg7 2.Qh5 Ra5 3.Qe2 a2 4.Ra1 Ng5 5.Re3 Qg6 6.Qb2 Rda8 7.Qxb6 Kh8 8.c4 Ra4 9.h4 Ne4 10.Qc6 Nd2 11.c5 Rg8 12.h5 Qb1 (0:00:33) 11206kN
[-0.60] d=14 1.Rae1 Qg7 2.Qh5 Ra5 3.Qe2 a2 4.Ra1 Ng5 5.Re3 Qg6 6.Qb2 Rda8 7.Qxb6 Kh8 8.c4 Qf5 9.Qb2 Ra4 10.h4 Ne4 11.c5 Qd5 (0:00:24) 8031kN
[-0.55] d=13 1.Rae1 Qg7 2.Qh5 Ra5 3.Qe2 a2 4.Ra1 Ng5 5.Re3 Qh7 6.Qb2 Qa7 7.h4 Nf7 8.Re2 (0:00:19) 6624kN
[-0.63] d=12 1.Qd2 Kh8 2.Qd3 Ra7 3.d5 Qf7 4.Qf5 Qd7 5.d6 a2 6.Qd5 (0:00:06) 2248kN
[-0.54] d=11 1.Qc1 a2 2.h4 Kh8 3.Qd2 Rda8 4.d5 Qc4 5.Re7 R6a7 6.Rxa7 Rxa7 7.d6 Nf8 8.Qh6 Kg8 9.Qxf6 Qe4 10.Kh2 (0:00:02) 834kN
[-0.58] d=10 1.Qc1 a2 2.Qc2 Ra3 3.Rxa2 Rxc3 4.Qg6 Kh8 5.Re1 Rg8 6.Qf5 Qg7 7.g3 Ng5 8.Re3 Rc1 9.Kh2 (0:00:01) 454kN
[-0.61] d=10 1.Qg6 Kh8 2.Qd3 Ra7 3.d5 Qc5 4.c4 Ra4 5.Re4 Ng5 6.Rh4 Kg7 (0:00:00) 210kN
[-0.55] d=9 1.Qg6 Kh8 2.Qd3 Ra7 3.d5 Qc5 4.c4 b5 5.cxb5 Rxd5 6.Qg6 Ra8 7.Rae1 Rg8 8.Re8 (0:00:00) 84kN
[-0.55] d=8 1.Qg6 Kh8 2.Qd3 Ra7 3.d5 Qc5 4.c4 b5 5.cxb5 Rxd5 6.Qg6 Ra8 7.Rae1 Rg8 8.Re8 (0:00:00) 49kN
[-0.44] d=7 1.Qg6 Kh8 2.Rae1 Qg7 3.Qe4 Ra7 4.Rxb6 a2 5.Ra1 Rg8 6.d5 (0:00:00) 19kN
[-0.24] d=7 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Kxg8 5.Qe8 Nf8 6.Re7 (0:00:00) 11kN
[-0.36] d=6 1.Qg6 Kh8 2.Rae1 Rg8 3.Re8 Qg7 4.Rxg8 Kxg8 5.Qf5 (0:00:00) 10kN
[-0.15] d=5 1.Qg6 Qg7 2.Qd3 Rda8 3.Qc4 Qf7 4.Qe2 (0:00:00) 3kN
[+0.12] d=5 1.Qg6 Qg7 2.Qd3 Qb7 3.Qg6 Qg7 4.Qd3 Qb7 5.Qg6 Qg7 6.Qd3 Qb7 7.Qg6 Qg7 8.Qd3 Qb7 9.Qg6 Qg7 10.Qd3 Qb7 11.Qg6 Qg7 12.Qd3 Qb7 13.Qg6 Qg7 14.Qd3 Qb7 15.Qg6 Qg7 16.Qd3 (0:00:00) 1kN
[+0.00] d=4 1.Qg6 Qg7 2.Qd3 Qb7 3.Qg6 Qg7 4.Qd3 Qb7 5.Qg6 Qg7 6.Qd3 Qb7 7.Qg6 Qg7 8.Qd3 Qb7 9.Qg6 Qg7 10.Qd3 Qb7 11.Qg6 Qg7 12.Qd3 Qb7 13.Qg6 Qg7 14.Qd3 Qb7 15.Qg6 Qg7 16.Qd3 (0:00:00) 0kN
[+0.00] d=3 1.Qg6 Qg7 2.Qd3 Qb7 3.Qg6 Qg7 4.Qd3 Qb7 5.Qg6 Qg7 6.Qd3 Qb7 7.Qg6 Qg7 8.Qd3 Qb7 9.Qg6 Qg7 10.Qd3 Qb7 11.Qg6 Qg7 12.Qd3 Qb7 13.Qg6 Qg7 14.Qd3 Qb7 15.Qg6 Qg7 16.Qd3 (0:00:00) 0kN
[+0.00] d=2 1.Qg6 Qg7 2.Qd3 Qb7 3.Qg6 Qg7 4.Qd3 Qb7 5.Qg6 Qg7 6.Qd3 Qb7 7.Qg6 Qg7 8.Qd3 Qb7 9.Qg6 Qg7 10.Qd3 Qb7 11.Qg6 Qg7 12.Qd3 Qb7 13.Qg6 Qg7 14.Qd3 Qb7 15.Qg6 Qg7 16.Qd3 (0:00:00) 0kN
[-0.04] d=1 1.Qg6 Qg7 2.Qd3 Qb7 3.Qg6 Qg7 4.Qd3 Qb7 5.Qg6 Qg7 6.Qd3 Qb7 7.Qg6 Qg7 8.Qd3 Qb7 9.Qg6 Qg7 10.Qd3 Qb7 11.Qg6 Qg7 12.Qd3 Qb7 13.Qg6 Qg7 14.Qd3 Qb7 15.Qg6 Qg7 16.Qd3 (0:00:00) 0kN
[-0.04] d=1 1.Qg6 Qg7 2.Qd3 Qb7 3.Qg6 (0:00:00) 0kN
[-0.47] d=1 1.Qe3 a2 2.Qd3 (0:00:00) 0kN
>I now have results for large pages enabled (353 kN/s):
Thanks Felix,i hope to publish your results soon
Greetings,
Sedat
>> I now have results for large pages enabled (353 kN/s):
> Thanks Felix,i hope to publish your results soon
Why don't you publish 312 kN/s instead of 353 kN/s?
always the highest value :)
>Why don't you publish 312 kN/s instead of 353 kN/s?
Ok,but what about your hardware ??? i have no patience to see Ernest's results ?!
OK Sedat, I already gave them here for my Intel Core 2 Duo (E6600) @ 3GHz
http://rybkaforum.net/cgi-bin/rybkaforum/topic_show.pl?pid=287972#pid287972
http://rybkaforum.net/cgi-bin/rybkaforum/topic_show.pl?pid=287972#pid287972
![[us]](/mwf/v12/flags/us.png "United States")
Felix,
I'm joining this discussion very late, so I have a basic question: Is this about kN/s as shown on the UCI analysis output or the kN/s from the "Chess Benchmark" feature of, say, Fritz 12?
I'm joining this discussion very late, so I have a basic question: Is this about kN/s as shown on the UCI analysis output or the kN/s from the "Chess Benchmark" feature of, say, Fritz 12?
its Aquarium and the conditions as described by sedat
![[cz]](/mwf/v12/flags/cz.png "Czech Republic")
felix try
sudo apt-get install shutter
sudo apt-get install shutter
Hi,
Although the benchmark gives a rough idea of KNS as you can see in the pic two 6 core setups and the slower machine is getting the higher KNS
Although the benchmark gives a rough idea of KNS as you can see in the pic two 6 core setups and the slower machine is getting the higher KNS
> Although the benchmark gives a rough idea of KNS as you can see in the pic two 6 core setups and the slower machine is getting the higher KNS
As i mentioned before,the kns values depend mainly on:
1.Processor
2.Ram
3.Operating System
4.Virus,Spyware,Malware...
5.Hardware Stability
6.Use Large Pages ON/OFF
7.Chess GUI
8.HashTable
9.HT ON/OFF
10...
Note:I strongly believe in that,not only with Rybka bench,that happens even with the rest benchmarking utilities too
For example (SEE BELOW) even on identical same hardware conditions we get each time different results !!!
I hope my information will be helpful...
Best,
Sedat
I think that the 874 kN/s for 4 CPUs is bullshit (cheated)!
Hmm...now i understand why you wish a lot to visit SedatChess site links,to find holes in my lists ?
It seems you are really a big hardware expert,maybe you have a lot of data about chess benchmark lists,can you inform please more about your benchmark activity ?
About 874 kns results,i dont think that there is cheating...
If there is a cheating... its in your results,you confirmed that your R4 engine performs between 200-240 kns,thats mean 20 % speed difference
I have never noticed a such kns speed difference with the current R4 bench position even testing it with many many times...
Please run your R4 benchmark again,i think you are doing something wrong ...if you give more information about your hardware benchmark details,then i can try to help you
BTW,i know very well Tsvetan Milchev since many years,he is a trusted person and his opening books were many times as participant in SCCT Book Tourneys
Maybe the bellow data can help you to make your own estimations:
Fritz Chess BenchMark:
Source of Fritz Benchmarks:
http://www.jens-hartmann.at/Fritzmarks/
?It seems you are really a big hardware expert,maybe you have a lot of data about chess benchmark lists,can you inform please more about your benchmark activity ?
About 874 kns results,i dont think that there is cheating...
>I tried Sedat's position with my dual-core (3 GHz), and obtained a range of 206 to 240 kN/s with several trials.
If there is a cheating... its in your results,you confirmed that your R4 engine performs between 200-240 kns,thats mean 20 % speed difference
I have never noticed a such kns speed difference with the current R4 bench position even testing it with many many times...
Please run your R4 benchmark again,i think you are doing something wrong ...if you give more information about your hardware benchmark details,then i can try to help you
BTW,i know very well Tsvetan Milchev since many years,he is a trusted person and his opening books were many times as participant in SCCT Book Tourneys
Maybe the bellow data can help you to make your own estimations:
Fritz Chess BenchMark:
--------------------------------------------------------------
kN/s Processor Speed Threads Hardware User
--------------------------------------------------------------
18898 Intel i7 980X @4.41 GHz 6 CPUs Martin Thoresen
15893 Intel i7 975 @4.31 GHz 8 CPUs Chiller
13473 Intel i7 920 @3.80 GHz 4 CPUs Thorsten E.
Source of Fritz Benchmarks:
http://www.jens-hartmann.at/Fritzmarks/
Hi Sedat,
I am afraid your benchmark lists play a big role as ego-boosters, that's why giving the best number of several test trials can be criticized (also the fact that it is at the tail end of the probability distribution: one could imagine somebody doing the test 1000 times, the "best" value will then be at more than 3 standard deviations from the average).
Also note that your Fritz Chess BenchMark data example concerns Fritz, not Rybka and the "Intel i7 975" has 8 CPUs only because HyperThreading was on, and this is "good" for that BenchMark
Now my results: I did the test 10 times, lowest value was 202 kn/s, best was 240 kn/s : well that's fact, that's the way naughty multiprocessors behave, like it or not!
Finally, concerning this guy Milchev (Mischief?): his result is about 20% higher than what others post (same processor and correction for GHz), 60% higher than your own post (yes I know, you run at 3.33 GHz...), so I have my doubts (unless he can explain his data...).
Accepting in your list "surprising" and unexplained data casts a shadow of doubt on all the data and on the work itself.
But don't forget, Sedat, only those who do nothing will never be criticized...
I am afraid your benchmark lists play a big role as ego-boosters, that's why giving the best number of several test trials can be criticized (also the fact that it is at the tail end of the probability distribution: one could imagine somebody doing the test 1000 times, the "best" value will then be at more than 3 standard deviations from the average).
Also note that your Fritz Chess BenchMark data example concerns Fritz, not Rybka and the "Intel i7 975" has 8 CPUs only because HyperThreading was on, and this is "good" for that BenchMark
Now my results: I did the test 10 times, lowest value was 202 kn/s, best was 240 kn/s : well that's fact, that's the way naughty multiprocessors behave, like it or not!
Finally, concerning this guy Milchev (Mischief?
): his result is about 20% higher than what others post (same processor and correction for GHz), 60% higher than your own post (yes I know, you run at 3.33 GHz...), so I have my doubts (unless he can explain his data...).Accepting in your list "surprising" and unexplained data casts a shadow of doubt on all the data and on the work itself.
But don't forget, Sedat, only those who do nothing will never be criticized...
![[ar]](/mwf/v12/flags/ar.png "Argentina")
Sure, you will never be criticized!
If you really want to be a good competitive noticeable hardware expert,then
you should find/write about cheatings or suspicious results in all other lists ?!
For example,who is cheating here,in this Fritz Chess BenchMark list ?
Personally i don't think about above results that there is cheating
As i mentioned before,i strongly believe that the bench results depend on many factors (hardware stability,processor,ram mhz...)
Do you know what is hurting me ?
Why those clever people (including you) are looking for holes,...) only in my computer chess work ?
Maybe is that can be the reasons:i from Antalya/Turkey ??? or maybe the reason is jealousy ? or perhaps you think that i am from the 3rd world
If so... that's mean those people still have not been in Antalya,but trust me,it is one of the most beautiful places of the world
Actually that's true that almost all we know here each other from online ...
And i hope one day those people will changed their minds and will show a little bit respect when we met together in live,in the real life !
you should find/write about cheatings or suspicious results in all other lists ?!
For example,who is cheating here,in this Fritz Chess BenchMark list ?
--------------------------------------------------------------
kN/s Processor Speed Threads Hardware User
--------------------------------------------------------------
13473 Intel i7 920 @3.80 GHz 4 CPUs Thorsten E.
12228 Intel i5 750 @4,20 GHz 4 CPUs Team-Oh
10350 Intel i7 920 @3.30 GHz 4 CPUs Sedat CanbazPersonally i don't think about above results that there is cheating
As i mentioned before,i strongly believe that the bench results depend on many factors (hardware stability,processor,ram mhz...)
Do you know what is hurting me ?
Why those clever people (including you) are looking for holes,...) only in my computer chess work ?
Maybe is that can be the reasons:i from Antalya/Turkey ??? or maybe the reason is jealousy ? or perhaps you think that i am from the 3rd world
If so... that's mean those people still have not been in Antalya,but trust me,it is one of the most beautiful places of the world
Actually that's true that almost all we know here each other from online ...
And i hope one day those people will changed their minds and will show a little bit respect when we met together in live,in the real life !
> show a little bit respect
Come on, Sedat, I do think Turkey is a beautiful country (and I am not the only one with this opinion).
Don't be paranoid...
Just look at what some poor guy (named Coit
) deducted from your list:http://www.talkchess.com/forum/viewtopic.php?p=376421#376421
This is why, if some Milchev sends you surprisingly high data, you could ask him for some details: did he use LP, how many tries, special configuration?
After all, maybe we can profit from somebody else's experience!...
>After all, maybe we can profit from somebody else's experience!...
Sorry..i have no free time,plus i don't want to disturb Tsvetan
If you really need a profit from somebody else's experience...then i suggest you to ask for help Thorsten E.,it seems he has a very successful fast hardware
And i recommend you if you ask him (Thorsten E.) for any help,please don't use any similar words like 'Don't be paranoid' or 'obviously nonsense' or 'cheating',otherwise nobody will help you too
> plus i don't want to disturb Tsvetan
I bet Tsvetan would be really disturbed!!!
And sorry Sedat, you are a hard worker, but really paranoid...
>And sorry Sedat, you are a hard worker, but really paranoid.
No,no...i can start to count several negative words about you,but sorry...i will not go in your way and will not fall in your level
Have a good night,
Sedat
I think you're being rude and disrespectful.
>I think you're being rude and disrespectful.
Thanks a lot for your understanding...
I hope a such people (like Enest) will not damage the quality of this good chess forum
Best Regards,
Sedat
Sorry Sedat, but it guy yes damage the forum.
Follow your serious work.
Thank you very much!
Rubén Cómes
Follow your serious work.
Thank you very much!
Rubén Cómes
I let my machine run the position for 15 minuets and got 249kn/s. On the list the machine is listed at 224 kn/s. So, i think it is possible to get a 20% difference in results, if one waits long enough (and breaks the rules).
>(and breaks the rules).
Hello Jesse,
A few notes about the current bench position...
Using Time Control:10 Minutes/Game (after the engine played move) :
--------------------------------------------------------------------
1.Rybka 4 w32/x64 1Cpu 128 MB HashTable - almost each time the kN/s values are almost identical
2.Rybka 4 w32/x64 2Cpus 128 MB HashTable - the kN/s values are changing approx.1-5%
3.Rybka 4 w32/x64 4Cpus 128 MB HashTable - the kN/s values are changing approx.1-10%
Note that i have never noticed 20 % speed difference (on my computers),especially with one or two core
Probably our dear Enest during the bench (with 2 cores) used another engine or under different conditions
Best,
Sedat
![[in]](/mwf/v12/flags/in.png "India")
Hi Sedat,
Do you expect different results if one does a 10 minute game vis-a-vis infinite analysis?
Regards,
Do you expect different results if one does a 10 minute game vis-a-vis infinite analysis?
Regards,
>Do you expect different results if one does a 10 minute game vis-a-vis infinite analysis?
Actually i did not make any serious comparing between both (10min and infinite) so far
Maybe later i will test the bench position in both variations,but not for now...
I have to do a lot of work and i should concentrate on SCCT VII Book Tourney
Greetings,
Sedat
> 1.Rybka 4 w32/x64 1Cpu 128 MB HashTable - almost each time the kN/s values are almost identical
...no disagreement about that, even the "almost" is too much!
> ...especially with ...two core
2 cores, 4 cores, 6...: that's another story
Just ask Bob Hyatt about multicore kN/s
Crawly, you have a wonderful machine: why don't you do some extended kN/s tests?
![[be]](/mwf/v12/flags/be.png "Belgium")
What is very normal..all depends on the configuration from your system and the parts you have used for that system!
Higher clocked system (even with same type cpu) don't means you have a faster system!
JP.
Higher clocked system (even with same type cpu) don't means you have a faster system!
JP.
>What is very normal..all depends on the configuration from your system and the parts you have used for that system!
> Higher clocked system (even with same type cpu) don't means you have a faster system!
Totally i agree !
Normally this is known issue by most experienced hardware users,but sometimes is appearing a such people who should be informed well
Best,
Sedat
it's not that easy. Of course you are right that there are other factors influencing the results, but I also claim that your test method leads to a larger error bar and is therefore making such results more likely :)
>your test method leads to a larger error bar
Ok...if you are criticizing,you should suggest which benchmark list is better,which benchmark utility is the right tool for measuring the hardwares speed
And i wonder,is it available a such one or still does not exist ?!
Look,this list is not a commercial work,this is just my hobby and i don't sell anything or advertising something...
Note also that my work is completely free and if someone does not like Rybka bench list,i suggest him to use the button 'Ignore'
BTW,since many years all my work in computer chess activity is criticized (or attacked maybe you remember ?)
So sometime i wonder what i should do... to be not criticized ?
But anyway nowadays i don't care,i take it easy...i am already accustomed
Last,i have no free time to look for holes in others work (but believe me if i start i will find many holes,but due to respect of their work i don't go in this way...)
And another interesting note,i can not find much data to criticize of these people,exception only comments,no more no less
I wish you happy weekend,
Sedat
I think fixed depth testing is better. however, the best is still playing a lot of games. kN/s are something the engine reports, it's calculated assuming different factors for higher numbers of processors which may be different for different setups and so on.
I'm happy about what you are doing, you can see that in our "links" section on our website for example :) however, you wondered why there are not many people here willing to contribute to that test, I think it's because of the method :) So fixed depth is better, easier to do and more reliable.
It's a quite complicated subject, after all I would simply ask Lukas what is best
I'm happy about what you are doing, you can see that in our "links" section on our website for example :) however, you wondered why there are not many people here willing to contribute to that test, I think it's because of the method :) So fixed depth is better, easier to do and more reliable.
It's a quite complicated subject, after all I would simply ask Lukas what is best
>I think fixed depth testing is better
Is it available a such Rybka bench list or maybe i missed that one,sorry for the question if already exist?
If not,i have no patience to see and compare the difference between 'fixed depth' bench results with 'best kns' results
Only after appearing a such Rybka bench list with 'fixed depth' i can make any conclusion...currently i think its too early to make any comments about this issue
>It's a quite complicated subject, after all I would simply ask Lukas what is best
That's true...for a better measuring and conclusion we need more data
Note also these complications are available in all areas,not only with benchmarking
What about rest issues,e.g which engines are clones or even there are lot of people who can not decide which is the world strongest engine?,opening book performance,which is the best way of testing and measuring the elo strength of chess engines...
As we see,there are a lot of questions which are waiting for answers
>I'm happy about what you are doing, you can see that in our "links" section on our website for example :)
Thank you...if you check SedatChess's links you will notice RybkaChess site is included too
And i think you know very well what i think for Rybka:
-I am in love with small fish,i have never noticed a such amazing performance by other engine to be hold for a such long time !!
>however, you wondered why there are not many people here willing to contribute to that test, I think it's because of the method :) So fixed depth is better, easier to do and more reliable.
Yes...but i don't claim that my list is perfect and about the method with 'fixed depth' nowadays i am too busy,that's why its will be great if other person will create a such list ...
And i will be very glad if the 'hardware expert' will inform us about in which conditions Rybka should be tested,which way is the ideal one of measuring hardwares speed ?
Best,
Sedat
Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill