Benchmark Reviews Discussion Forum Right Header

Go Back Benchmark Reviews Discussion Forum > Hardware > Overclockers Lounge

Overclockers Lounge Post your projects and results in this section.

Reply
Thread Tools Display Modes
#1
Unread 4th January 2010, 11:21 PM
LiquidNitrogenOverclocking LiquidNitrogenOverclocking is offline
Kilobyte
Join Date: Jan 2010
Posts: 35
A Benchmark of mine

If any of you overclockers are bored, I did write a program earlier this summer that would be a good benchmark application.

It takes around 9 hours to run on a 2.4 GHz AMD, which means as systems inevitably get faster and faster, the performance comparison won't get "down to the decimal points" for quite a while. Even if computers were 100 times faster than the system mentioned above, it would still take about 5.5 minutes to run to completion.

What the program does is solve every checkers endgame with 6 pieces, provided one side has no more than 3 men. It's over 1 billion positions, so it will write about 1 GB of data to your hard drive.

It will also create a "report.txt" file, showing the longest winning position for each of the various configurations (3 kings vs. 2 kings and 1 checker, 1 king + 2 checkers vs. 2 kings + 1 checker, 2 kings vs. 3 checkers in the 5-piece subdatabase, etc.) and also the time it takes to solve the database, iteration by iteration.

The longest win is surprisingly 167 moves (counting play by both sides) and it is rather intricate.

Anyway, I would like to see how long it takes on your various systems, and I'd like to build a copy of the fastest one, so I can run a version of this program that will solve the play for up to 10 pieces, which is 8.5 trillion different positions.

For those who want to help out, download it from here

http://www.liquidnitrogenoverclockin...b6_builder.zip

You have to decompress it before you run it, of course.

Only do this if you have several hours (like 5 - 8 hours) that you don't need to use your system.

Last edited by LiquidNitrogenOverclocking : 5th January 2010 at 11:55 AM. Reason: estimated the total time it takes to run a little better
Reply With Quote
#2
Unread 5th January 2010, 05:07 AM
XJnine XJnine is offline
Gigabyte
Join Date: Mar 2007
Location: Jacksonville, FL
Posts: 166
Send a message via ICQ to XJnine
Running it for you...

I put my cpu back to stock speeds since OC'ing may give inaccurate results and if you were to use a system like this you may not get the same OC results. After the initial run I will run it again overclocked for comparison.

Is this program multithreaded? How will I know how much longer it will take to finish?
__________________
Asrock Z68 Fatal1ty mobo
Intel Core i7 2600k CPU @ 4.8ghz
Noctua NH-D14 HSF
2x PNY GeForce 680 GTX SLI
32gb G.Skill Sniper DDR3 1866 @ 2133
2x Samsung 840Pro 256gb SSD's, Raid 0
2x Seagate Barracuda 3tb HD's, RAID 1
Vertex 4 128gb SRT Cache Drive
Memorex MRX-800 Blu-Ray drive
Corsair 1000 watt power supply
Fractal Design Define XL R2 Case
Dell U3011 30" monitor

Last edited by XJnine : 5th January 2010 at 05:16 AM.
Reply With Quote
#3
Unread 5th January 2010, 07:23 AM
LiquidNitrogenOverclocking LiquidNitrogenOverclocking is offline
Kilobyte
Join Date: Jan 2010
Posts: 35
Quote:
Originally Posted by XJnine View Post
Running it for you...

I put my cpu back to stock speeds since OC'ing may give inaccurate results and if you were to use a system like this you may not get the same OC results. After the initial run I will run it again overclocked for comparison.

Is this program multithreaded? How will I know how much longer it will take to finish?
Hi,

And thanks for being first onboard!

Your idea about stock vs. OC is a great idea. Now why couldn't I think of that?

The program is a single-threaded 64-bit application. It might seem strange to create a 64-bit application for a game that has only 32 squares, but I found that using 64-bit "words" in my checkers move generator GREATLY REDUCED the number of lines of code needed for the Move Generator.

Typically, you need about 200 or so instructions to be able to make a legal move generator in the game of checkers, no matter if you use an Array Move Generator, or a so-called "Bitboard".

If you discount the curly braces in my C-code, I was able to get it down to about 19 lines of executable code. I'm working on publishing a paper on the move generator and the "overuse of bit aperture" to perform an almost RISC-like bit-manipulation to gain speed.

Having said that, here comes the paradox: I used a sub-optimal means to solve the checkers endgames. I implemented what is called a Forward Move Generator, which is easy to code, easy to expand (to other endgames I wish to solve) and was much easier to write. There is a much faster technique that uses both Forward and Reverse move generation in combination to solve endgames by only visiting the positions that could potentially impact the next iteration. By keeping a "list" of positions that were just solved on the first 2 Forward Move passes, you can then visit these positions, let the other side to move make a "backwards move", and it must make a forward move from there into one of the positions previously resolved. If this resolves the position, or changes the distance-to-win/loss currently known, you add it to the list to "visit" on the next pass, where you would unmove the other side to get a bunch of positions that would play into that, etc.

Unmoving and moving is a little more involved, but it has a tremendous speedup advantage. The forward mover must revisit every position of the database being solved on every iteration. After the 3rd pass, you can see, there are fewer and fewer positions resolved, so mostly the program is parsing the entire database and not doing much solving. The reverse move generator flies through in "no time", so it wouldn't be much of a benchmarking tool.

When I tackle the 10-piece databases, I will be using a Forward and Reverse move generator, on a "copy" of the system that produces the fastest results from here.

Thanks again for your help! I should really put some feedback into the program to let the end user know how much work is left for it to do.

The 6-piece database will contain 16 different "slices". So, if you can read your "slice number" on the screen, that will give you some indication of how much is left.

Please email your report.txt file to me at LiquidNitrogenOverclocking@hushmail.com and let me know what "name" or "handle" you would like on our benchmarking web page, or if you would rather just be anonymous.

Last edited by LiquidNitrogenOverclocking : 5th January 2010 at 07:25 AM. Reason: typo corrections
Reply With Quote
#4
Unread 5th January 2010, 07:40 AM
LiquidNitrogenOverclocking LiquidNitrogenOverclocking is offline
Kilobyte
Join Date: Jan 2010
Posts: 35
8-piece Database Results for 4 Kings vs. 4 Kings

Below are the results from my "All Kings" generator, which I ran on a very slow Macintosh G4 @ 500 MHz, just to look for ways to improve my buffering technique that will be used to solve endgames that have more positions than my RAM can hold. Since you have to solve slices with all kings first, it's a good idea to just get them out of the way. So I did 4K4K, 5K3K, 5K4K, 4K5K, and 5K5K is just way to slow on that hardware.

Here is what those results looked like

Code:
Thursday, March 19, 2009 @ 1:43:36 PM
***********************************************************************************
* 8-piece database, 4 white kings 0 white checkers vs. 4 red kings 0 red checkers *
***********************************************************************************

slice 1 of 8-piece database jump pass complete (global slice # 232).

jump pass wins = 339573858, jump pass losses = 18221225, jump pass draws = 100684741.
unknown positions remaining to be resolved = 277801176.


BUFFER EFFICIENCY = 99.99644470% with 7611270 disk reads and 214244320113 RAM probes.

Thursday, March 19, 2009 @ 5:40:10 PM
DB 8, slice 1, (global slice 232): iteration 1 completed.
*********************************************************

wins resolved this pass = 12428917, losses resolved this pass = 14493238, draws resolved this pass = 10102501.
win lengths improved this pass = 0, loss lengths that changed this pass = 0.

total move pass wins = 12428917, total move pass losses = 14493238, total move pass draws = 10102501.
cumulative win lengths improved = 0, cumulative loss lengths that changed = 0.

unknown positions remaining to be resolved = 240776520.


BUFFER EFFICIENCY = 99.99649811% with 7611306 disk reads and 217308349177 RAM probes.

Thursday, March 19, 2009 @ 9:36:30 PM
DB 8, slice 1, (global slice 232): iteration 2 completed.
*********************************************************

wins resolved this pass = 2892627, losses resolved this pass = 697528, draws resolved this pass = 832414.
win lengths improved this pass = 25021, loss lengths that changed this pass = 639.

total move pass wins = 15321544, total move pass losses = 15190766, total move pass draws = 10934915.
cumulative win lengths improved = 25021, cumulative loss lengths that changed = 639.

unknown positions remaining to be resolved = 236353951.


BUFFER EFFICIENCY = 99.99654388% with 7611306 disk reads and 220271050887 RAM probes.

Friday, March 20, 2009 @ 1:31:58 AM
DB 8, slice 1, (global slice 232): iteration 3 completed.
*********************************************************

wins resolved this pass = 264730, losses resolved this pass = 88738, draws resolved this pass = 82721.
win lengths improved this pass = 13808, loss lengths that changed this pass = 3105.

total move pass wins = 15586274, total move pass losses = 15279504, total move pass draws = 11017636.
cumulative win lengths improved = 38829, cumulative loss lengths that changed = 3744.

unknown positions remaining to be resolved = 235917762.


BUFFER EFFICIENCY = 99.99658966% with 7611306 disk reads and 223225014620 RAM probes.

Friday, March 20, 2009 @ 5:25:57 AM
DB 8, slice 1, (global slice 232): iteration 4 completed.
*********************************************************

wins resolved this pass = 102812, losses resolved this pass = 30801, draws resolved this pass = 13124.
win lengths improved this pass = 9925, loss lengths that changed this pass = 1944.

total move pass wins = 15689086, total move pass losses = 15310305, total move pass draws = 11030760.
cumulative win lengths improved = 48754, cumulative loss lengths that changed = 5688.

unknown positions remaining to be resolved = 235771025.


BUFFER EFFICIENCY = 99.99663544% with 7611306 disk reads and 226178145286 RAM probes.

Friday, March 20, 2009 @ 9:18:50 AM
DB 8, slice 1, (global slice 232): iteration 5 completed.
*********************************************************

wins resolved this pass = 78888, losses resolved this pass = 22346, draws resolved this pass = 7078.
win lengths improved this pass = 12407, loss lengths that changed this pass = 2776.

total move pass wins = 15767974, total move pass losses = 15332651, total move pass draws = 11037838.
cumulative win lengths improved = 61161, cumulative loss lengths that changed = 8464.

unknown positions remaining to be resolved = 235662713.


BUFFER EFFICIENCY = 99.99668121% with 7611306 disk reads and 229131161933 RAM probes.

Friday, March 20, 2009 @ 1:14:13 PM
DB 8, slice 1, (global slice 232): iteration 6 completed.
*********************************************************

wins resolved this pass = 68045, losses resolved this pass = 17896, draws resolved this pass = 6774.
win lengths improved this pass = 15806, loss lengths that changed this pass = 3534.

total move pass wins = 15836019, total move pass losses = 15350547, total move pass draws = 11044612.
cumulative win lengths improved = 76967, cumulative loss lengths that changed = 11998.

unknown positions remaining to be resolved = 235569998.


BUFFER EFFICIENCY = 99.99671936% with 7611306 disk reads and 232084119414 RAM probes.

Friday, March 20, 2009 @ 5:08:30 PM
DB 8, slice 1, (global slice 232): iteration 7 completed.
*********************************************************

wins resolved this pass = 51986, losses resolved this pass = 15866, draws resolved this pass = 4997.
win lengths improved this pass = 18983, loss lengths that changed this pass = 4798.

total move pass wins = 15888005, total move pass losses = 15366413, total move pass draws = 11049609.
cumulative win lengths improved = 95950, cumulative loss lengths that changed = 16796.

unknown positions remaining to be resolved = 235497149.


BUFFER EFFICIENCY = 99.99676514% with 7611306 disk reads and 235037021529 RAM probes.

Friday, March 20, 2009 @ 9:03:49 PM
DB 8, slice 1, (global slice 232): iteration 8 completed.
*********************************************************

wins resolved this pass = 43123, losses resolved this pass = 13031, draws resolved this pass = 4469.
win lengths improved this pass = 19227, loss lengths that changed this pass = 5032.

total move pass wins = 15931128, total move pass losses = 15379444, total move pass draws = 11054078.
cumulative win lengths improved = 115177, cumulative loss lengths that changed = 21828.

unknown positions remaining to be resolved = 235436526.


BUFFER EFFICIENCY = 99.99680328% with 7611306 disk reads and 237989882724 RAM probes.

Saturday, March 21, 2009 @ 12:59:48 AM
DB 8, slice 1, (global slice 232): iteration 9 completed.
*********************************************************

wins resolved this pass = 36210, losses resolved this pass = 12051, draws resolved this pass = 3386.
win lengths improved this pass = 19481, loss lengths that changed this pass = 5499.

total move pass wins = 15967338, total move pass losses = 15391495, total move pass draws = 11057464.
cumulative win lengths improved = 134658, cumulative loss lengths that changed = 27327.

unknown positions remaining to be resolved = 235384879.


BUFFER EFFICIENCY = 99.99684143% with 7611306 disk reads and 240942707702 RAM probes.

Saturday, March 21, 2009 @ 4:57:26 AM
DB 8, slice 1, (global slice 232): iteration 10 completed.
*********************************************************

wins resolved this pass = 31518, losses resolved this pass = 11136, draws resolved this pass = 2578.
win lengths improved this pass = 19597, loss lengths that changed this pass = 5720.

total move pass wins = 15998856, total move pass losses = 15402631, total move pass draws = 11060042.
cumulative win lengths improved = 154255, cumulative loss lengths that changed = 33047.

unknown positions remaining to be resolved = 235339647.


BUFFER EFFICIENCY = 99.99687958% with 7611306 disk reads and 243895505285 RAM probes.

Saturday, March 21, 2009 @ 8:50:07 AM
DB 8, slice 1, (global slice 232): iteration 11 completed.
*********************************************************

wins resolved this pass = 28702, losses resolved this pass = 10459, draws resolved this pass = 2415.
win lengths improved this pass = 20354, loss lengths that changed this pass = 6199.

total move pass wins = 16027558, total move pass losses = 15413090, total move pass draws = 11062457.
cumulative win lengths improved = 174609, cumulative loss lengths that changed = 39246.

unknown positions remaining to be resolved = 235298071.


BUFFER EFFICIENCY = 99.99691772% with 7611306 disk reads and 246848281922 RAM probes.

Saturday, March 21, 2009 @ 12:45:18 PM
DB 8, slice 1, (global slice 232): iteration 12 completed.
*********************************************************

wins resolved this pass = 25760, losses resolved this pass = 8258, draws resolved this pass = 2486.
win lengths improved this pass = 20746, loss lengths that changed this pass = 6418.

total move pass wins = 16053318, total move pass losses = 15421348, total move pass draws = 11064943.
cumulative win lengths improved = 195355, cumulative loss lengths that changed = 45664.

unknown positions remaining to be resolved = 235261567.


BUFFER EFFICIENCY = 99.99695587% with 7611306 disk reads and 249801038536 RAM probes.

Saturday, March 21, 2009 @ 4:41:05 PM
DB 8, slice 1, (global slice 232): iteration 13 completed.
*********************************************************

wins resolved this pass = 22993, losses resolved this pass = 6680, draws resolved this pass = 2257.
win lengths improved this pass = 22326, loss lengths that changed this pass = 6938.

total move pass wins = 16076311, total move pass losses = 15428028, total move pass draws = 11067200.
cumulative win lengths improved = 217681, cumulative loss lengths that changed = 52602.

unknown positions remaining to be resolved = 235229637.


BUFFER EFFICIENCY = 99.99698639% with 7611306 disk reads and 252753774681 RAM probes.

Saturday, March 21, 2009 @ 8:35:27 PM
DB 8, slice 1, (global slice 232): iteration 14 completed.
*********************************************************

wins resolved this pass = 20461, losses resolved this pass = 4875, draws resolved this pass = 2200.
win lengths improved this pass = 22433, loss lengths that changed this pass = 6888.

total move pass wins = 16096772, total move pass losses = 15432903, total move pass draws = 11069400.
cumulative win lengths improved = 240114, cumulative loss lengths that changed = 59490.

unknown positions remaining to be resolved = 235202101.


BUFFER EFFICIENCY = 99.99702454% with 7611306 disk reads and 255706492324 RAM probes.

Sunday, March 22, 2009 @ 12:30:40 AM
DB 8, slice 1, (global slice 232): iteration 15 completed.
*********************************************************

wins resolved this pass = 17388, losses resolved this pass = 4063, draws resolved this pass = 1476.
win lengths improved this pass = 22236, loss lengths that changed this pass = 7154.

total move pass wins = 16114160, total move pass losses = 15436966, total move pass draws = 11070876.
cumulative win lengths improved = 262350, cumulative loss lengths that changed = 66644.

unknown positions remaining to be resolved = 235179174.


BUFFER EFFICIENCY = 99.99705505% with 7611306 disk reads and 258659191641 RAM probes.

Sunday, March 22, 2009 @ 4:26:37 AM
DB 8, slice 1, (global slice 232): iteration 16 completed.
*********************************************************

wins resolved this pass = 12807, losses resolved this pass = 2933, draws resolved this pass = 1387.
win lengths improved this pass = 20608, loss lengths that changed this pass = 6638.

total move pass wins = 16126967, total move pass losses = 15439899, total move pass draws = 11072263.
cumulative win lengths improved = 282958, cumulative loss lengths that changed = 73282.

unknown positions remaining to be resolved = 235162047.


BUFFER EFFICIENCY = 99.99709320% with 7611306 disk reads and 261611878067 RAM probes.

Sunday, March 22, 2009 @ 8:21:54 AM
DB 8, slice 1, (global slice 232): iteration 17 completed.
*********************************************************

wins resolved this pass = 9545, losses resolved this pass = 2233, draws resolved this pass = 953.
win lengths improved this pass = 16696, loss lengths that changed this pass = 5878.

total move pass wins = 16136512, total move pass losses = 15442132, total move pass draws = 11073216.
cumulative win lengths improved = 299654, cumulative loss lengths that changed = 79160.

unknown positions remaining to be resolved = 235149316.


BUFFER EFFICIENCY = 99.99712372% with 7611306 disk reads and 264564552787 RAM probes.

Sunday, March 22, 2009 @ 12:15:02 PM
DB 8, slice 1, (global slice 232): iteration 18 completed.
*********************************************************

wins resolved this pass = 6876, losses resolved this pass = 1710, draws resolved this pass = 895.
win lengths improved this pass = 12393, loss lengths that changed this pass = 4596.

total move pass wins = 16143388, total move pass losses = 15443842, total move pass draws = 11074111.
cumulative win lengths improved = 312047, cumulative loss lengths that changed = 83756.

unknown positions remaining to be resolved = 235139835.


BUFFER EFFICIENCY = 99.99715424% with 7611306 disk reads and 267517219068 RAM probes.

Sunday, March 22, 2009 @ 4:10:23 PM
DB 8, slice 1, (global slice 232): iteration 19 completed.
*********************************************************

wins resolved this pass = 5265, losses resolved this pass = 1363, draws resolved this pass = 670.
win lengths improved this pass = 9209, loss lengths that changed this pass = 3968.

total move pass wins = 16148653, total move pass losses = 15445205, total move pass draws = 11074781.
cumulative win lengths improved = 321256, cumulative loss lengths that changed = 87724.

unknown positions remaining to be resolved = 235132537.


BUFFER EFFICIENCY = 99.99718475% with 7611306 disk reads and 270469877948 RAM probes.

Sunday, March 22, 2009 @ 8:06:28 PM
DB 8, slice 1, (global slice 232): iteration 20 completed.
*********************************************************

wins resolved this pass = 4040, losses resolved this pass = 972, draws resolved this pass = 427.
win lengths improved this pass = 6863, loss lengths that changed this pass = 2813.

total move pass wins = 16152693, total move pass losses = 15446177, total move pass draws = 11075208.
cumulative win lengths improved = 328119, cumulative loss lengths that changed = 90537.

unknown positions remaining to be resolved = 235127098.


BUFFER EFFICIENCY = 99.99721527% with 7611306 disk reads and 273422530841 RAM probes.

Monday, March 23, 2009 @ 12:02:03 AM
DB 8, slice 1, (global slice 232): iteration 21 completed.
*********************************************************

wins resolved this pass = 3244, losses resolved this pass = 759, draws resolved this pass = 322.
win lengths improved this pass = 5199, loss lengths that changed this pass = 1877.

total move pass wins = 16155937, total move pass losses = 15446936, total move pass draws = 11075530.
cumulative win lengths improved = 333318, cumulative loss lengths that changed = 92414.

unknown positions remaining to be resolved = 235122773.


BUFFER EFFICIENCY = 99.99724579% with 7611306 disk reads and 276375180026 RAM probes.

Monday, March 23, 2009 @ 3:56:10 AM
DB 8, slice 1, (global slice 232): iteration 22 completed.
*********************************************************

wins resolved this pass = 2724, losses resolved this pass = 631, draws resolved this pass = 249.
win lengths improved this pass = 3778, loss lengths that changed this pass = 1153.

total move pass wins = 16158661, total move pass losses = 15447567, total move pass draws = 11075779.
cumulative win lengths improved = 337096, cumulative loss lengths that changed = 93567.

unknown positions remaining to be resolved = 235119169.


BUFFER EFFICIENCY = 99.99727631% with 7611306 disk reads and 279327826411 RAM probes.

Monday, March 23, 2009 @ 7:52:27 AM
DB 8, slice 1, (global slice 232): iteration 23 completed.
*********************************************************

wins resolved this pass = 1781, losses resolved this pass = 407, draws resolved this pass = 243.
win lengths improved this pass = 2701, loss lengths that changed this pass = 743.

total move pass wins = 16160442, total move pass losses = 15447974, total move pass draws = 11076022.
cumulative win lengths improved = 339797, cumulative loss lengths that changed = 94310.

unknown positions remaining to be resolved = 235116738.


BUFFER EFFICIENCY = 99.99730682% with 7611306 disk reads and 282280470605 RAM probes.

Monday, March 23, 2009 @ 11:45:50 AM
DB 8, slice 1, (global slice 232): iteration 24 completed.
*********************************************************

wins resolved this pass = 1003, losses resolved this pass = 224, draws resolved this pass = 171.
win lengths improved this pass = 2436, loss lengths that changed this pass = 555.

total move pass wins = 16161445, total move pass losses = 15448198, total move pass draws = 11076193.
cumulative win lengths improved = 342233, cumulative loss lengths that changed = 94865.

unknown positions remaining to be resolved = 235115340.


BUFFER EFFICIENCY = 99.99732971% with 7611306 disk reads and 285233112576 RAM probes.

Monday, March 23, 2009 @ 3:39:51 PM
DB 8, slice 1, (global slice 232): iteration 25 completed.
*********************************************************

wins resolved this pass = 543, losses resolved this pass = 108, draws resolved this pass = 43.
win lengths improved this pass = 1417, loss lengths that changed this pass = 325.

total move pass wins = 16161988, total move pass losses = 15448306, total move pass draws = 11076236.
cumulative win lengths improved = 343650, cumulative loss lengths that changed = 95190.

unknown positions remaining to be resolved = 235114646.


BUFFER EFFICIENCY = 99.99736023% with 7611306 disk reads and 288185753107 RAM probes.

Monday, March 23, 2009 @ 7:34:23 PM
DB 8, slice 1, (global slice 232): iteration 26 completed.
*********************************************************

wins resolved this pass = 289, losses resolved this pass = 72, draws resolved this pass = 46.
win lengths improved this pass = 751, loss lengths that changed this pass = 153.

total move pass wins = 16162277, total move pass losses = 15448378, total move pass draws = 11076282.
cumulative win lengths improved = 344401, cumulative loss lengths that changed = 95343.

unknown positions remaining to be resolved = 235114239.


BUFFER EFFICIENCY = 99.99738312% with 7611306 disk reads and 291138393261 RAM probes.

Monday, March 23, 2009 @ 11:30:22 PM
DB 8, slice 1, (global slice 232): iteration 27 completed.
*********************************************************

wins resolved this pass = 130, losses resolved this pass = 28, draws resolved this pass = 7.
win lengths improved this pass = 429, loss lengths that changed this pass = 80.

total move pass wins = 16162407, total move pass losses = 15448406, total move pass draws = 11076289.
cumulative win lengths improved = 344830, cumulative loss lengths that changed = 95423.

unknown positions remaining to be resolved = 235114074.


BUFFER EFFICIENCY = 99.99741364% with 7611306 disk reads and 294091033078 RAM probes.

Tuesday, March 24, 2009 @ 3:26:17 AM
DB 8, slice 1, (global slice 232): iteration 28 completed.
*********************************************************

wins resolved this pass = 27, losses resolved this pass = 4, draws resolved this pass = 8.
win lengths improved this pass = 192, loss lengths that changed this pass = 51.

total move pass wins = 16162434, total move pass losses = 15448410, total move pass draws = 11076297.
cumulative win lengths improved = 345022, cumulative loss lengths that changed = 95474.

unknown positions remaining to be resolved = 235114035.


BUFFER EFFICIENCY = 99.99743652% with 7611306 disk reads and 297043672846 RAM probes.

Tuesday, March 24, 2009 @ 7:21:19 AM
DB 8, slice 1, (global slice 232): iteration 29 completed.
*********************************************************

wins resolved this pass = 1, losses resolved this pass = 0, draws resolved this pass = 0.
win lengths improved this pass = 71, loss lengths that changed this pass = 8.

total move pass wins = 16162435, total move pass losses = 15448410, total move pass draws = 11076297.
cumulative win lengths improved = 345093, cumulative loss lengths that changed = 95482.

unknown positions remaining to be resolved = 235114034.


BUFFER EFFICIENCY = 99.99745941% with 7611306 disk reads and 299996312560 RAM probes.

Tuesday, March 24, 2009 @ 11:16:48 AM
DB 8, slice 1, (global slice 232): iteration 30 completed.
*********************************************************

wins resolved this pass = 0, losses resolved this pass = 0, draws resolved this pass = 0.
win lengths improved this pass = 4, loss lengths that changed this pass = 1.

total move pass wins = 16162435, total move pass losses = 15448410, total move pass draws = 11076297.
cumulative win lengths improved = 345097, cumulative loss lengths that changed = 95483.

unknown positions remaining to be resolved = 235114034.


BUFFER EFFICIENCY = 99.99749756% with 7611306 disk reads and 303685233274 RAM probes.

Tuesday, March 24, 2009 @ 12:24:00 PM
DB 8, slice 1, (global slice 232): iteration 31 completed.
*********************************************************

wins resolved this pass = 0, losses resolved this pass = 0, draws resolved this pass = 0.
win lengths improved this pass = 1, loss lengths that changed this pass = 0.

total move pass wins = 16162435, total move pass losses = 15448410, total move pass draws = 11076297.
cumulative win lengths improved = 345098, cumulative loss lengths that changed = 95483.

unknown positions remaining to be resolved = 0.


BUFFER EFFICIENCY = 99.99749756% with 7611306 disk reads and 304282482716 RAM probes.

Tuesday, March 24, 2009 @ 1:15:10 PM
DB 8, slice 1, (global slice 232): iteration 32 completed.
*********************************************************

wins resolved this pass = 0, losses resolved this pass = 0, draws resolved this pass = 0.
win lengths improved this pass = 0, loss lengths that changed this pass = 0.

total move pass wins = 16162435, total move pass losses = 15448410, total move pass draws = 11076297.
cumulative win lengths improved = 345098, cumulative loss lengths that changed = 95483.

unknown positions remaining to be resolved = 0.

Tuesday, March 24, 2009 @ 1:22:33 PM
*************************************
* 8-piece database slice # 1 solved *
*************************************

***********************************************************************************
* 8-piece database, 4 white kings 0 white checkers vs. 4 red kings 0 red checkers *
***********************************************************************************

TOTAL POSITIONS: 736281000
TOTAL BYTES:     0 (buffered)


TOTAL WINS   (white to move)........... 355736293
TOTAL DRAWS  (white to move)........... 346875072
TOTAL LOSSES (white to move)........... 33669635

Wins resolved as jumps................. 339573858
Draws resolved as jumps................ 100684741
Losses resolved as jumps............... 18221225

Wins resolved from moving.............. 16162435
Draws resolved from moving............. 11076297
Losses resolved from moving............ 15448410

Wins with NO JUMPS for EITHER SIDE..... 7705209
Draws with NO JUMPS for EITHER SIDE.... 153104694
Losses with NO JUMPS for EITHER SIDE... 197297

Draws resolved from unknowns........... 235114034

Cumulative win length improvements..... 345098
Cumulative loss length changes......... 95483


Database resolved after 34 iterations (includes the JUMP pass).
The longest win requires 115 ply to complete.

There are 40 wins of the same length in this database slice.
One of the longest white to move and win positions: (white pieces start at top of board)

***************************************************************************************************************
*----------*----------*----------*----------*----------*----------*----------*----------*----------*----------*
*----------*----------*----------*----------*----------*----------*----------*----------*----------*----------*
*----------*----------*----------*----------*----------*----------*----------*----------*----------*----------*
*----------*----------*----------*----------*----------*----------*----------*----------*----------*----------*
*----------*----------*----------*----------*----------*----------*----------*----------*----------*----------*
***************************************************************************************************************
*----------*##########*          *##########*          *##########*          *##########* -------- *----------*
*----------*##########*          *##########*          *##########*          *##########* |WWWWWW| *----------*
*----------*### 32 ###*          *### 31 ###*          *### 30 ###*          *### 29 ###* -------- *----------*
*----------*##########*          *##########*          *##########*          *##########* |WWWWWW| *----------*
*----------*##########*          *##########*          *##########*          *##########* -------- *----------*
***************************************************************************************************************
*----------*          *##########*          *##########* -------- *##########*          *##########*----------*
*----------*          *##########*          *##########* |RRRRRR| *##########*          *##########*----------*
*----------*          *### 28 ###*          *### 27 ###* -------- *### 26 ###*          *### 25 ###*----------*
*----------*          *##########*          *##########* |RRRRRR| *##########*          *##########*----------*
*----------*          *##########*          *##########* -------- *##########*          *##########*----------*
***************************************************************************************************************
*----------*##########*          *##########*          *##########*          *##########* -------- *----------*
*----------*##########*          *##########*          *##########*          *##########* |WWWWWW| *----------*
*----------*### 24 ###*          *### 23 ###*          *### 22 ###*          *### 21 ###* -------- *----------*
*----------*##########*          *##########*          *##########*          *##########* |WWWWWW| *----------*
*----------*##########*          *##########*          *##########*          *##########* -------- *----------*
***************************************************************************************************************
*----------*          *##########*          *##########* -------- *##########*          *##########*----------*
*----------*          *##########*          *##########* |RRRRRR| *##########*          *##########*----------*
*----------*          *### 20 ###*          *### 19 ###* -------- *### 18 ###*          *### 17 ###*----------*
*----------*          *##########*          *##########* |RRRRRR| *##########*          *##########*----------*
*----------*          *##########*          *##########* -------- *##########*          *##########*----------*
***************************************************************************************************************
*----------*##########*          *##########*          *##########*          *##########*          *----------*
*----------*##########*          *##########*          *##########*          *##########*          *----------*
*----------*### 16 ###*          *### 15 ###*          *### 14 ###*          *### 13 ###*          *----------*
*----------*##########*          *##########*          *##########*          *##########*          *----------*
*----------*##########*          *##########*          *##########*          *##########*          *----------*
***************************************************************************************************************
*----------*          *##########* -------- *##########*          *##########*          *##########*----------*
*----------*          *##########* |RRRRRR| *##########*          *##########*          *##########*----------*
*----------*          *### 12 ###* -------- *### 11 ###*          *### 10 ###*          *### 09 ###*----------*
*----------*          *##########* |RRRRRR| *##########*          *##########*          *##########*----------*
*----------*          *##########* -------- *##########*          *##########*          *##########*----------*
***************************************************************************************************************
*----------*##########*          *##########* -------- *##########*          *##########* -------- *----------*
*----------*##########*          *##########* |WWWWWW| *##########*          *##########* |RRRRRR| *----------*
*----------*### 08 ###*          *### 07 ###* -------- *### 06 ###*          *### 05 ###* -------- *----------*
*----------*##########*          *##########* |WWWWWW| *##########*          *##########* |RRRRRR| *----------*
*----------*##########*          *##########* -------- *##########*          *##########* -------- *----------*
***************************************************************************************************************
*----------*          *##########*          *##########*          *##########* -------- *##########*----------*
*----------*          *##########*          *##########*          *##########* |WWWWWW| *##########*----------*
*----------*          *### 04 ###*          *### 03 ###*          *### 02 ###* -------- *### 01 ###*----------*
*----------*          *##########*          *##########*          *##########* |WWWWWW| *##########*----------*
*----------*          *##########*          *##########*          *##########* -------- *##########*----------*
***************************************************************************************************************
*----------*----------*----------*----------*----------*----------*----------*----------*----------*----------*
*----------*----------*----------*----------*----------*----------*----------*----------*----------*----------*
*----------*----------*----------*----------*----------*----------*----------*----------*----------*----------*
*----------*----------*----------*----------*----------*----------*----------*----------*----------*----------*
*----------*----------*----------*----------*----------*----------*----------*----------*----------*----------*

Where there are NO JUMPS for either side, the longest win requires 115 ply to complete.

There are 4 JUMPLESS wins of the same length in this database slice.
One of those longest white to move and win positions: (white pieces start at top of board)

***************************************************************************************************************
*----------*----------*----------*----------*----------*----------*----------*----------*----------*----------*
*----------*----------*----------*----------*----------*----------*----------*----------*----------*----------*
*----------*----------*----------*----------*----------*----------*----------*----------*----------*----------*
*----------*----------*----------*----------*----------*----------*----------*----------*----------*----------*
*----------*----------*----------*----------*----------*----------*----------*----------*----------*----------*
***************************************************************************************************************
*----------*##########*          *##########*          *##########*          *##########* -------- *----------*
*----------*##########*          *##########*          *##########*          *##########* |WWWWWW| *----------*
*----------*### 32 ###*          *### 31 ###*          *### 30 ###*          *### 29 ###* -------- *----------*
*----------*##########*          *##########*          *##########*          *##########* |WWWWWW| *----------*
*----------*##########*          *##########*          *##########*          *##########* -------- *----------*
***************************************************************************************************************
*----------*          *##########*          *##########*          *##########*          *##########*----------*
*----------*          *##########*          *##########*          *##########*          *##########*----------*
*----------*          *### 28 ###*          *### 27 ###*          *### 26 ###*          *### 25 ###*----------*
*----------*          *##########*          *##########*          *##########*          *##########*----------*
*----------*          *##########*          *##########*          *##########*          *##########*----------*
***************************************************************************************************************
*----------*##########*          *##########* -------- *##########* -------- *##########* -------- *----------*
*----------*##########*          *##########* |WWWWWW| *##########* |RRRRRR| *##########* |WWWWWW| *----------*
*----------*### 24 ###*          *### 23 ###* -------- *### 22 ###* -------- *### 21 ###* -------- *----------*
*----------*##########*          *##########* |WWWWWW| *##########* |RRRRRR| *##########* |WWWWWW| *----------*
*----------*##########*          *##########* -------- *##########* -------- *##########* -------- *----------*
***************************************************************************************************************
*----------*          *##########*          *##########*          *##########*          *##########*----------*
*----------*          *##########*          *##########*          *##########*          *##########*----------*
*----------*          *### 20 ###*          *### 19 ###*          *### 18 ###*          *### 17 ###*----------*
*----------*          *##########*          *##########*          *##########*          *##########*----------*
*----------*          *##########*          *##########*          *##########*          *##########*----------*
***************************************************************************************************************
*----------*##########*          *##########* -------- *##########*          *##########*          *----------*
*----------*##########*          *##########* |RRRRRR| *##########*          *##########*          *----------*
*----------*### 16 ###*          *### 15 ###* -------- *### 14 ###*          *### 13 ###*          *----------*
*----------*##########*          *##########* |RRRRRR| *##########*          *##########*          *----------*
*----------*##########*          *##########* -------- *##########*          *##########*          *----------*
***************************************************************************************************************
*----------*          *##########*          *##########* -------- *##########*          *##########*----------*
*----------*          *##########*          *##########* |RRRRRR| *##########*          *##########*----------*
*----------*          *### 12 ###*          *### 11 ###* -------- *### 10 ###*          *### 09 ###*----------*
*----------*          *##########*          *##########* |RRRRRR| *##########*          *##########*----------*
*----------*          *##########*          *##########* -------- *##########*          *##########*----------*
***************************************************************************************************************
*----------*##########*          *##########*          *##########*          *##########*          *----------*
*----------*##########*          *##########*          *##########*          *##########*          *----------*
*----------*### 08 ###*          *### 07 ###*          *### 06 ###*          *### 05 ###*          *----------*
*----------*##########*          *##########*          *##########*          *##########*          *----------*
*----------*##########*          *##########*          *##########*          *##########*          *----------*
***************************************************************************************************************
*----------*          *##########*          *##########* -------- *##########* -------- *##########*----------*
*----------*          *##########*          *##########* |WWWWWW| *##########* |RRRRRR| *##########*----------*
*----------*          *### 04 ###*          *### 03 ###* -------- *### 02 ###* -------- *### 01 ###*----------*
*----------*          *##########*          *##########* |WWWWWW| *##########* |RRRRRR| *##########*----------*
*----------*          *##########*          *##########* -------- *##########* -------- *##########*----------*
***************************************************************************************************************
*----------*----------*----------*----------*----------*----------*----------*----------*----------*----------*
*----------*----------*----------*----------*----------*----------*----------*----------*----------*----------*
*----------*----------*----------*----------*----------*----------*----------*----------*----------*----------*
*----------*----------*----------*----------*----------*----------*----------*----------*----------*----------*
*----------*----------*----------*----------*----------*----------*----------*----------*----------*----------*

Last edited by LiquidNitrogenOverclocking : 5th January 2010 at 07:44 AM. Reason: Changed "All King" to "All Kings"
Reply With Quote
#5
Unread 5th January 2010, 07:56 AM
Olin Coles's Avatar
Olin Coles Olin Coles is offline
Executive Editor
Join Date: Feb 2007
Posts: 2,951
My primary concern is consistency. Does this tool run the exact same moves in the same order with every test? I'm happy to try it if each run is identical to the previous, but if it's random then there's room for variation that I can't allow.
__________________
You can follow Benchmark Reviews on Facebook and Twitter!
Reply With Quote
#6
Unread 5th January 2010, 08:20 AM
LiquidNitrogenOverclocking LiquidNitrogenOverclocking is offline
Kilobyte
Join Date: Jan 2010
Posts: 35
Quote:
Originally Posted by Olin Coles View Post
My primary concern is consistency. Does this tool run the exact same moves in the same order with every test? I'm happy to try it if each run is identical to the previous, but if it's random then there's room for variation that I can't allow.
The test executes the same way every time. It is completely deterministic.
Reply With Quote
#7
Unread 5th January 2010, 08:44 AM
XJnine XJnine is offline
Gigabyte
Join Date: Mar 2007
Location: Jacksonville, FL
Posts: 166
Send a message via ICQ to XJnine
Is this something you've considered rewriting in OpenCL/CUDA?

I'm not a programming guy but it seems like you could take advantage of the parallelism of the GPU and run all of your possible combinations (1K2C vs 1K1C, etc) at once. Since it appears that one endgame result doesn't influence the other I would think you could greatly benefit buy running the various permutations all at the same time.

I could very easily be wrong, just ask my wife...

Do you think that could be something that would work for you?
__________________
Asrock Z68 Fatal1ty mobo
Intel Core i7 2600k CPU @ 4.8ghz
Noctua NH-D14 HSF
2x PNY GeForce 680 GTX SLI
32gb G.Skill Sniper DDR3 1866 @ 2133
2x Samsung 840Pro 256gb SSD's, Raid 0
2x Seagate Barracuda 3tb HD's, RAID 1
Vertex 4 128gb SRT Cache Drive
Memorex MRX-800 Blu-Ray drive
Corsair 1000 watt power supply
Fractal Design Define XL R2 Case
Dell U3011 30" monitor
Reply With Quote
#8
Unread 5th January 2010, 09:12 AM
LiquidNitrogenOverclocking LiquidNitrogenOverclocking is offline
Kilobyte
Join Date: Jan 2010
Posts: 35
Quote:
Originally Posted by XJnine View Post
Is this something you've considered rewriting in OpenCL/CUDA?

I'm not a programming guy but it seems like you could take advantage of the parallelism of the GPU and run all of your possible combinations (1K2C vs 1K1C, etc) at once. Since it appears that one endgame result doesn't influence the other I would think you could greatly benefit buy running the various permutations all at the same time.

I could very easily be wrong, just ask my wife...

Do you think that could be something that would work for you?
The problem is the dependencies.

A checker can promote into a king.

So, you need to solve any combination that has kings before you can hope to solve the solve configuration with the king replaced by a checker.

Having said that, there is a parallelization that is exploitable. This can be done not with a coding platform of choice, but rather a separate instance of the same code, which "knows" when it can "kick in" and solve a slice whose dependent portion is complete.

Take 3 kings vs. 3 kings as the example.

That is the first 6-piece slice to solve. It requires the 3 kings vs. 2 kings slice from the 5-piece database to be solved, since a king can jump a king and land in that. Likewise, a double-jump can land in the 3 kings against 1 king slice from the 4-piece database, and so on.

But, once the 3 kings vs. 3 kings slice is solved, you can solve 2 kings + 1 checker vs. 3 kings, and 3 kings vs. 2 kings + 1 checker.

If you look at the output, you see I do that form of parallelization, but I don't do any 4-way or 8-way parallelizations, since 6-pieces doesn't feature that many parallelizations.

For the 10-piece databases, you are absolutely correct.

After solving 5K0C vs. 5K0C (5 kings + 0 checkers vs. the same), you can do:

4K1C vs. 5K0C and 5K0C vs. 4K1C

This spawns another pair that can be solved simultaneously (once these are completed of course):

4K1C vs. 4K1C and 3K2C vs. 5K0C

as well as this configuration:

5K0C vs. 3K2C

All of these will promote into a slice that has one less checker and one more king.

So let's say I wanted to exploit this 3-way parallelism (which can be generalized to n-ways).

I just need 3 copies of the same program running, 1 of which is the "main solver" trying to get as many "all kings" slices solved as possible, the other two are the "de-promotion" solvers, trying to solve any slices that contain checkers.

The first depromoter can call "fopen()" to see if I saved a particular database file to the disk yet, say "5K0C vs. 4K1C.d10" for example. If the call to "fopen()" fails, it will look for some other file, and so on. If it finds it, it knows it can solve a slice with one less checker from either side, as mentioned above. As calls to find files fail, the program can just wait a few minutes (I set my default to check through the entire list of unsolved slices every 10 minutes) and keep trying, until it solved everything on its list.

As long as all of my parallel solvers look for different files, and, collectively, the cover all the bases, the database can be solved rather expediently on muti-core systems.

So, your observation was astute!

Last edited by LiquidNitrogenOverclocking : 5th January 2010 at 09:20 AM. Reason: adding something for a little more clarity
Reply With Quote
#9
Unread 5th January 2010, 10:55 AM
XJnine XJnine is offline
Gigabyte
Join Date: Mar 2007
Location: Jacksonville, FL
Posts: 166
Send a message via ICQ to XJnine
All done, the report is on its way. I'll get my system oc'd and do another run tonight. FYI, I was using my system for web browsing while this ran but my CPU usage was never very high so I don't think it impacted the calculations very much.
__________________
Asrock Z68 Fatal1ty mobo
Intel Core i7 2600k CPU @ 4.8ghz
Noctua NH-D14 HSF
2x PNY GeForce 680 GTX SLI
32gb G.Skill Sniper DDR3 1866 @ 2133
2x Samsung 840Pro 256gb SSD's, Raid 0
2x Seagate Barracuda 3tb HD's, RAID 1
Vertex 4 128gb SRT Cache Drive
Memorex MRX-800 Blu-Ray drive
Corsair 1000 watt power supply
Fractal Design Define XL R2 Case
Dell U3011 30" monitor
Reply With Quote
#10
Unread 5th January 2010, 11:13 AM
LiquidNitrogenOverclocking LiquidNitrogenOverclocking is offline
Kilobyte
Join Date: Jan 2010
Posts: 35
Quote:
Originally Posted by XJnine View Post
All done, the report is on its way. I'll get my system oc'd and do another run tonight. FYI, I was using my system for web browsing while this ran but my CPU usage was never very high so I don't think it impacted the calculations very much.
Thanks, I got it!

Let me know if & how you want your timetrial displayed at:

http://www.liquidnitrogenoverclockin..._results.shtml

Currently you are #1 with a solving time of 5:30:37 which is much faster then the time of just under 9 hours for the 2.4 GHz AMD I used to do the same run.

Last edited by LiquidNitrogenOverclocking : 5th January 2010 at 11:38 AM. Reason: said "19 hours" instead of "9 hours" in my original post... oops!
Reply With Quote
Sponsored links
Reply


Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

vB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump


All times are GMT -7. The time now is 06:58 AM.

Benchmark Reviews Discussion Forum
Powered by vBulletin® Version 3.6.8
Copyright ©2000 - 2017, Jelsoft Enterprises Ltd.
Remove advertisements

Follow Benchmark Reviews on FacebookReceive Tweets from Benchmark Reviews on Twitter