i7 Hyper-threading

Discussion of anything and everything relating to chess playing software and machines.

Moderators: hgm, Dann Corbit, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
kasinp
Posts: 217
Joined: Sat Dec 02, 2006 9:47 pm
Location: Toronto
Full name: Peter Kasinski

Re: i7 Hyper-threading

Post by kasinp » Fri Dec 24, 2010 11:53 am

FWIW, HT tests using Rybka 4 x64 on i7 970 (6 physical cores):

On i7 970 (6 physical cores) I ran 50 tactical positions three separate times (total of 6 tests) with HT on and off. For HT tests I used ProcessLasso to assign CPU afinities for Rybka to 0,2,4,6,8,10 as prescribed by Lucas C. on Rybka's site.

I was promarily interested in this question: to what degree is NPS indicative or useful in predicting time to solution?
My results (averaged from 3 tests):

HT off: R4 solved 50 of 50, average time 48.31 sec., average depth 14.86, CPU temperature during tests 54C.

HT on: R4 solved 50 of 50, average time 52.02 sec., average depth 14.68, CPU temeparure during tests 60C.

Meanwhile, CPU benchmark position (from Sedat C. site) yielded the following (again, averaged over three tests):

HT off: 661
HT on: 1102

Regards,
PK

jpqy
Posts: 533
Joined: Thu Apr 24, 2008 7:31 am
Location: Belgium

Re: i7 Hyper-threading

Post by jpqy » Fri Dec 24, 2010 11:53 am

Hi Robert,

This information is so usefull that it should be on your website!!
So that everbody can find it back and try it to find the best SD..or even put it in a Readme file!

Thanks for this super strong engine..the best i have test yet!

JP.

mwyoung
Posts: 2727
Joined: Wed May 12, 2010 8:00 pm

Re: i7 Hyper-threading

Post by mwyoung » Fri Dec 24, 2010 1:03 pm

yanquis1972 wrote:fwiw, with 8 cores my best result came w/ a split depth of 12 (by about ~3-4% over the default of 10). my guess is the best result for four cores is 10 but i'll let someone else confirm this.
I also have split depth set to 12. Was a about 250nps faster with sd12 over sd10 with 8 cores. 4780 vs 4519

brinan
Posts: 87
Joined: Fri Apr 23, 2010 3:34 pm
Location: London
Full name: bri nan
Contact:

Re: i7 Hyper-threading

Post by brinan » Fri Dec 24, 2010 1:45 pm

Processor
Intel® Core™ i3-330M Processor (3M Cache, 2.13 GHz, 2 cores, 4 hardware threads)
Memory
2,936MB
Operating System
Windows 7, Home Premium Edition (Build 7600).

Console Input
Houdini_15_x64.exe
info string POPCNT available
info string 128 MB Hash

setoption name threads value 2 [entered 4 on second run of autotune]
autotune

autotune Output Summary (data rearranged)

Code: Select all

                2 threads    4 threads

Average CPU time per thread:
                597.219 s    596.064 s
CPU-corrected average speed:
                2676 kN/s    3023 kN/s
Average Split_Depth speeds:
10             2694 kN/s    3083 kN/s     
12             2689 kN/s    3082 kN/s
14             2684 kN/s    3014 kN/s
16             2624 kN/s    2979 kN/s
18             2626 kN/s    2858 kN/s

Breakdowns:
Split_Depth 10
Position 1: 2694 kN/s    3114 kN/s
Position 2: 2736 kN/s    3100 kN/s
Position 3: 2676 kN/s    3081 kN/s
Position 4: 2670 kN/s    3036 kN/s

Split_Depth 12
Position 1: 2730 kN/s    3130 kN/s
Position 2: 2693 kN/s    3110 kN/s
Position 3: 2687 kN/s    3030 kN/s
Position 4: 2644 kN/s    3058 kN/s

Split_Depth 14
Position 1: 2673 kN/s    3043 kN/s
Position 2: 2705 kN/s    3054 kN/s
Position 3: 2634 kN/s    2951 kN/s
Position 4: 2725 kN/s    3007 kN/s

Split_Depth 16
Position 1: 2582 kN/s    2984 kN/s
Position 2: 2585 kN/s    3012 kN/s
Position 3: 2661 kN/s    2997 kN/s
Position 4: 2668 kN/s    2922 kN/s

Split_Depth 18
Position 1: 2623 kN/s    2819 kN/s
Position 2: 2642 kN/s    3018 kN/s
Position 3: 2635 kN/s    2842 kN/s
Position 4: 2606 kN/s    2753 kN/s
Click on www button for all event details and games.
News can be received automatically by subscribing to my RSS feed (click on icon near top).

Robert Flesher
Posts: 1246
Joined: Tue Aug 18, 2009 1:06 am

Re: i7 Hyper-threading

Post by Robert Flesher » Fri Dec 24, 2010 1:45 pm

Houdini wrote:
mwyoung wrote:After loading my new i7 laptop with a few chess engines. I am finding Rybka 4, Fritz 12, and Houdini are running faster using 8 logical cores. This is true with NPS and timed test positions. As this is my first i7 cpu, does Hyper-threading with the i7 help instead of hurt chess engines performance. This was not true in the past. In my testing I am clearly faster with Hyper-threading turned on running my chess engines.

Example Fritz 12

Fritz benchmark 4 cores. 5380 Nps 11.21

Fritz benchmark with HT 8 logical cores. 6995 Nps 14.57
Mark,

I've never experienced any useful improvement from hyper-threading for Houdini. Even if the nps is slightly higher, the overhead of the additional threads could very well reduce the actual playing strength.

Could you try the "autotune" feature of Houdini? This feature was intended for picking the best Split_Depth parameter, but it also functions as an accurate benchmark for multi-thread node speeds.

Double-click on the Houdini executable to open it in a console window.
Then enter the following commands to "autotune" for 4 threads.

Code: Select all

setoption name threads value 4
autotune
Houdini will run for about 10 minutes analyzing a number of positions for different values of split_depth. Make sure you're doing nothing else on your computer, just leave it running for 10 minutes. At the end you'll get a summary of the node speeds.

Then repeat the procedure for 8 threads.

Code: Select all

setoption name threads value 8
autotune
What results do you get?

I also find with my I7 920 that HT enabled produces faster NPS, but also solves best move, or tactical problems faster. Until I see proof that 4 threads is faster ( I cannot see it on my machine) versus 4 actual + 4 virtual threads, HT will remain on. Please explain further, as these are not the results I am finding.

mwyoung
Posts: 2727
Joined: Wed May 12, 2010 8:00 pm

Re: i7 Hyper-threading (Test done Houdini faster with HT 6%)

Post by mwyoung » Fri Dec 24, 2010 2:01 pm

Houdini wrote:
mwyoung wrote:After loading my new i7 laptop with a few chess engines. I am finding Rybka 4, Fritz 12, and Houdini are running faster using 8 logical cores. This is true with NPS and timed test positions. As this is my first i7 cpu, does Hyper-threading with the i7 help instead of hurt chess engines performance. This was not true in the past. In my testing I am clearly faster with Hyper-threading turned on running my chess engines.

Example Fritz 12

Fritz benchmark 4 cores. 5380 Nps 11.21

Fritz benchmark with HT 8 logical cores. 6995 Nps 14.57
Mark,

I've never experienced any useful improvement from hyper-threading for Houdini. Even if the nps is slightly higher, the overhead of the additional threads could very well reduce the actual playing strength.

Could you try the "autotune" feature of Houdini? This feature was intended for picking the best Split_Depth parameter, but it also functions as an accurate benchmark for multi-thread node speeds.

Double-click on the Houdini executable to open it in a console window.
Then enter the following commands to "autotune" for 4 threads.

Code: Select all

setoption name threads value 4
autotune
Houdini will run for about 10 minutes analyzing a number of positions for different values of split_depth. Make sure you're doing nothing else on your computer, just leave it running for 10 minutes. At the end you'll get a summary of the node speeds.

Then repeat the procedure for 8 threads.

Code: Select all

setoption name threads value 8
autotune
What results do you get?
4 Threads Houndini 1.5 64bit

info time 8043 nodes 39420830 nps 4901000 tbhits 0 cpuload 495
info time 9045 nodes 44990468 nps 4974000 tbhits 0 cpuload 504
info time 10048 nodes 50344588 nps 5010000 tbhits 0 cpuload 491
info time 11049 nodes 55748719 nps 5045000 tbhits 0 cpuload 494
info time 12059 nodes 61187975 nps 5074000 tbhits 0 cpuload 500
info time 13062 nodes 66632068 nps 5101000 tbhits 0 cpuload 493
info time 14066 nodes 71807386 nps 5105000 tbhits 0 cpuload 503
info time 15067 nodes 76482510 nps 5076000 tbhits 0 cpuload 496
info time 16070 nodes 82296358 nps 5121000 tbhits 0 cpuload 493
info time 17073 nodes 88200054 nps 5166000 tbhits 0 cpuload 489
info time 18075 nodes 94439086 nps 5224000 tbhits 0 cpuload 505
info time 19079 nodes 100372752 nps 5260000 tbhits 0 cpuload 495
info time 20087 nodes 105500714 nps 5252000 tbhits 0 cpuload 501
info time 21093 nodes 110750334 nps 5250000 tbhits 0 cpuload 494
info time 22095 nodes 116442285 nps 5270000 tbhits 0 cpuload 498
info time 23101 nodes 121723313 nps 5269000 tbhits 0 cpuload 502
info time 24105 nodes 126987228 nps 5268000 tbhits 0 cpuload 495
info time 25106 nodes 132859377 nps 5291000 tbhits 0 cpuload 498
info time 26110 nodes 138832132 nps 5317000 tbhits 0 cpuload 501
info time 27112 nodes 145039253 nps 5349000 tbhits 0 cpuload 498
info time 28120 nodes 151261745 nps 5379000 tbhits 0 cpuload 491
info time 29122 nodes 157408513 nps 5405000 tbhits 0 cpuload 505
Split_Depth 14, Position 4: 5405 kN/s
info time 1004 nodes 3494925 nps 3481000 tbhits 0 cpuload 495
info time 2007 nodes 7695850 nps 3834000 tbhits 0 cpuload 497
info time 3016 nodes 12264208 nps 4066000 tbhits 0 cpuload 500
info time 4021 nodes 17132121 nps 4260000 tbhits 0 cpuload 494
info time 5030 nodes 22274279 nps 4428000 tbhits 0 cpuload 494
info time 6038 nodes 27318074 nps 4524000 tbhits 0 cpuload 499
info time 7042 nodes 32478704 nps 4612000 tbhits 0 cpuload 493
info time 8048 nodes 37648848 nps 4678000 tbhits 0 cpuload 496
info time 9050 nodes 42568048 nps 4703000 tbhits 0 cpuload 496
info time 10054 nodes 47914105 nps 4765000 tbhits 0 cpuload 495
info time 11055 nodes 52701892 nps 4767000 tbhits 0 cpuload 498
info time 12057 nodes 57826217 nps 4796000 tbhits 0 cpuload 498
info time 13061 nodes 63551986 nps 4865000 tbhits 0 cpuload 493
info time 14063 nodes 69179557 nps 4919000 tbhits 0 cpuload 496
info time 15068 nodes 74563107 nps 4948000 tbhits 0 cpuload 502
info time 16070 nodes 79737506 nps 4961000 tbhits 0 cpuload 498
info time 17072 nodes 84889423 nps 4972000 tbhits 0 cpuload 494
info time 18079 nodes 90175969 nps 4987000 tbhits 0 cpuload 495
info time 19084 nodes 95504269 nps 5004000 tbhits 0 cpuload 490
info time 20086 nodes 100835678 nps 5020000 tbhits 0 cpuload 498
info time 21094 nodes 105877889 nps 5019000 tbhits 0 cpuload 502
info time 22099 nodes 111503765 nps 5045000 tbhits 0 cpuload 496
info time 23101 nodes 116892836 nps 5060000 tbhits 0 cpuload 504
info time 24104 nodes 122463259 nps 5080000 tbhits 0 cpuload 493
info time 25107 nodes 127005670 nps 5058000 tbhits 0 cpuload 495
info time 26113 nodes 132400471 nps 5070000 tbhits 0 cpuload 502
info time 27119 nodes 137990436 nps 5088000 tbhits 0 cpuload 492
info time 28124 nodes 143564206 nps 5104000 tbhits 0 cpuload 492
info time 29125 nodes 149226414 nps 5123000 tbhits 0 cpuload 504
Split_Depth 16, Position 1: 5120 kN/s
info time 1005 nodes 4118215 nps 4097000 tbhits 0 cpuload 496
info time 2009 nodes 9594428 nps 4775000 tbhits 0 cpuload 497
info time 3011 nodes 14969721 nps 4971000 tbhits 0 cpuload 505
info time 4014 nodes 19564528 nps 4874000 tbhits 0 cpuload 495
info time 5015 nodes 24620845 nps 4909000 tbhits 0 cpuload 494
info time 6016 nodes 29821191 nps 4956000 tbhits 0 cpuload 496
info time 7018 nodes 35372073 nps 5040000 tbhits 0 cpuload 492
info time 8024 nodes 40790227 nps 5083000 tbhits 0 cpuload 500
info time 9029 nodes 45964571 nps 5090000 tbhits 0 cpuload 492
info time 10030 nodes 50813391 nps 5066000 tbhits 0 cpuload 500
info time 11037 nodes 56181674 nps 5090000 tbhits 0 cpuload 493
info time 12040 nodes 61927118 nps 5143000 tbhits 0 cpuload 495
info time 13044 nodes 67456922 nps 5171000 tbhits 0 cpuload 501
info time 14046 nodes 72955963 nps 5194000 tbhits 0 cpuload 496
info time 15053 nodes 77824318 nps 5170000 tbhits 0 cpuload 493
info time 16059 nodes 83862570 nps 5222000 tbhits 0 cpuload 502
info time 17063 nodes 89468256 nps 5243000 tbhits 0 cpuload 495
info time 18065 nodes 95363207 nps 5278000 tbhits 0 cpuload 504
info time 19067 nodes 100100808 nps 5249000 tbhits 0 cpuload 498
info time 20071 nodes 105415326 nps 5252000 tbhits 0 cpuload 493
info time 21072 nodes 110935125 nps 5264000 tbhits 0 cpuload 498
info time 22079 nodes 116752576 nps 5287000 tbhits 0 cpuload 501
info time 23085 nodes 122552195 nps 5308000 tbhits 0 cpuload 492
info time 24087 nodes 128236329 nps 5323000 tbhits 0 cpuload 504
info time 25090 nodes 133988874 nps 5340000 tbhits 0 cpuload 491
info time 26091 nodes 139478215 nps 5345000 tbhits 0 cpuload 488
info time 27096 nodes 145350645 nps 5364000 tbhits 0 cpuload 494
info time 28099 nodes 151053949 nps 5375000 tbhits 0 cpuload 501
info time 29105 nodes 156959397 nps 5392000 tbhits 0 cpuload 492
Split_Depth 16, Position 2: 5377 kN/s
info time 1003 nodes 3153717 nps 3144000 tbhits 0 cpuload 495
info time 2006 nodes 7392709 nps 3685000 tbhits 0 cpuload 503
info time 3009 nodes 12515784 nps 4159000 tbhits 0 cpuload 495
info time 4014 nodes 17626954 nps 4391000 tbhits 0 cpuload 500
info time 5015 nodes 22238074 nps 4434000 tbhits 0 cpuload 496
info time 6017 nodes 27576984 nps 4583000 tbhits 0 cpuload 496
info time 7019 nodes 32767994 nps 4668000 tbhits 0 cpuload 498
info time 8027 nodes 38268805 nps 4767000 tbhits 0 cpuload 502
info time 9029 nodes 42924109 nps 4754000 tbhits 0 cpuload 498
info time 10035 nodes 48411182 nps 4824000 tbhits 0 cpuload 492
info time 11040 nodes 53646068 nps 4859000 tbhits 0 cpuload 504
info time 12046 nodes 58611913 nps 4865000 tbhits 0 cpuload 496
info time 13047 nodes 63707268 nps 4882000 tbhits 0 cpuload 506
info time 14053 nodes 69020746 nps 4911000 tbhits 0 cpuload 492
info time 15056 nodes 73927866 nps 4910000 tbhits 0 cpuload 497
info time 16062 nodes 79449327 nps 4946000 tbhits 0 cpuload 503
info time 17067 nodes 84609736 nps 4957000 tbhits 0 cpuload 492
info time 18079 nodes 89793451 nps 4966000 tbhits 0 cpuload 497
info time 19080 nodes 95027723 nps 4980000 tbhits 0 cpuload 496
info time 20085 nodes 100395143 nps 4998000 tbhits 0 cpuload 504
info time 21089 nodes 105058651 nps 4981000 tbhits 0 cpuload 497
info time 22095 nodes 110579339 nps 5004000 tbhits 0 cpuload 503
info time 23097 nodes 116450631 nps 5041000 tbhits 0 cpuload 494
info time 24101 nodes 122013628 nps 5062000 tbhits 0 cpuload 493
info time 25104 nodes 127102526 nps 5063000 tbhits 0 cpuload 495
info time 26108 nodes 132601944 nps 5078000 tbhits 0 cpuload 501
info time 27113 nodes 137189214 nps 5059000 tbhits 0 cpuload 492
info time 28115 nodes 142414150 nps 5065000 tbhits 0 cpuload 492
info time 29121 nodes 148082081 nps 5085000 tbhits 0 cpuload 503
Split_Depth 16, Position 3: 5092 kN/s
info time 1003 nodes 2819495 nps 2811000 tbhits 0 cpuload 497
info time 2009 nodes 7377604 nps 3672000 tbhits 0 cpuload 500
info time 3012 nodes 11795372 nps 3916000 tbhits 0 cpuload 491
info time 4020 nodes 16919571 nps 4208000 tbhits 0 cpuload 502
info time 5022 nodes 21543510 nps 4289000 tbhits 0 cpuload 496
info time 6024 nodes 26297575 nps 4365000 tbhits 0 cpuload 496
info time 7027 nodes 32252893 nps 4589000 tbhits 0 cpuload 503
info time 8029 nodes 37457091 nps 4665000 tbhits 0 cpuload 498
info time 9033 nodes 42946068 nps 4754000 tbhits 0 cpuload 495
info time 10036 nodes 48373475 nps 4819000 tbhits 0 cpuload 493
info time 11038 nodes 54192940 nps 4909000 tbhits 0 cpuload 504
info time 12045 nodes 59712345 nps 4957000 tbhits 0 cpuload 495
info time 13046 nodes 65269286 nps 5003000 tbhits 0 cpuload 494
info time 14053 nodes 70268947 nps 5000000 tbhits 0 cpuload 499
info time 15054 nodes 75871620 nps 5039000 tbhits 0 cpuload 498
info time 16055 nodes 81601408 nps 5082000 tbhits 0 cpuload 496
info time 17059 nodes 87117236 nps 5106000 tbhits 0 cpuload 504
info time 18062 nodes 93002177 nps 5149000 tbhits 0 cpuload 493
info time 19067 nodes 98512999 nps 5166000 tbhits 0 cpuload 492
info time 20068 nodes 103754814 nps 5170000 tbhits 0 cpuload 496
info time 21072 nodes 109333307 nps 5188000 tbhits 0 cpuload 504
info time 22077 nodes 114616086 nps 5191000 tbhits 0 cpuload 492
info time 23081 nodes 119927800 nps 5195000 tbhits 0 cpuload 504
info time 24088 nodes 124780075 nps 5180000 tbhits 0 cpuload 493
info time 25091 nodes 130193968 nps 5188000 tbhits 0 cpuload 493
info time 26094 nodes 135493614 nps 5192000 tbhits 0 cpuload 499
info time 27100 nodes 141219162 nps 5211000 tbhits 0 cpuload 496
info time 28104 nodes 147257581 nps 5239000 tbhits 0 cpuload 499
info time 29106 nodes 153496322 nps 5273000 tbhits 0 cpuload 494
Split_Depth 16, Position 4: 5289 kN/s
info time 1003 nodes 3066306 nps 3057000 tbhits 0 cpuload 493
info time 2012 nodes 7479778 nps 3717000 tbhits 0 cpuload 502
info time 3013 nodes 12122603 nps 4023000 tbhits 0 cpuload 494
info time 4015 nodes 16316755 nps 4063000 tbhits 0 cpuload 498
info time 5023 nodes 20494910 nps 4080000 tbhits 0 cpuload 502
info time 6029 nodes 25212157 nps 4181000 tbhits 0 cpuload 488
info time 7032 nodes 30493297 nps 4336000 tbhits 0 cpuload 503
info time 8036 nodes 36118161 nps 4494000 tbhits 0 cpuload 497
info time 9040 nodes 41310746 nps 4569000 tbhits 0 cpuload 495
info time 10044 nodes 46029470 nps 4582000 tbhits 0 cpuload 503
info time 11049 nodes 51181996 nps 4632000 tbhits 0 cpuload 494
info time 12052 nodes 56239483 nps 4666000 tbhits 0 cpuload 505
info time 13054 nodes 60673470 nps 4647000 tbhits 0 cpuload 496
info time 14055 nodes 65773378 nps 4679000 tbhits 0 cpuload 496
info time 15058 nodes 71106042 nps 4722000 tbhits 0 cpuload 497
info time 16062 nodes 76711792 nps 4775000 tbhits 0 cpuload 501
info time 17065 nodes 81814542 nps 4794000 tbhits 0 cpuload 497
info time 18069 nodes 86789778 nps 4803000 tbhits 0 cpuload 497
info time 19072 nodes 92441202 nps 4846000 tbhits 0 cpuload 503
info time 20076 nodes 98160326 nps 4889000 tbhits 0 cpuload 497
info time 21081 nodes 103590012 nps 4913000 tbhits 0 cpuload 490
info time 22083 nodes 108710964 nps 4922000 tbhits 0 cpuload 498
info time 23088 nodes 114542143 nps 4961000 tbhits 0 cpuload 496
info time 24093 nodes 119776448 nps 4971000 tbhits 0 cpuload 494
info time 25094 nodes 124665603 nps 4967000 tbhits 0 cpuload 506
info time 26095 nodes 130059852 nps 4984000 tbhits 0 cpuload 496
info time 27097 nodes 135639577 nps 5005000 tbhits 0 cpuload 496
info time 28099 nodes 141245043 nps 5026000 tbhits 0 cpuload 494
info time 29102 nodes 146404576 nps 5030000 tbhits 0 cpuload 497
Split_Depth 18, Position 1: 5046 kN/s
info time 1004 nodes 3750646 nps 3735000 tbhits 0 cpuload 495
info time 2013 nodes 8235333 nps 4091000 tbhits 0 cpuload 500
info time 3014 nodes 12584027 nps 4175000 tbhits 0 cpuload 490
info time 4017 nodes 17955561 nps 4469000 tbhits 0 cpuload 493
info time 5025 nodes 23428673 nps 4662000 tbhits 0 cpuload 497
info time 6030 nodes 28865753 nps 4787000 tbhits 0 cpuload 492
info time 7035 nodes 34221077 nps 4864000 tbhits 0 cpuload 500
info time 8041 nodes 39960820 nps 4969000 tbhits 0 cpuload 490
info time 9047 nodes 45622796 nps 5042000 tbhits 0 cpuload 500
info time 10051 nodes 50697367 nps 5044000 tbhits 0 cpuload 495
info time 11052 nodes 54980900 nps 4974000 tbhits 0 cpuload 496
info time 12055 nodes 60562711 nps 5023000 tbhits 0 cpuload 503
info time 13057 nodes 65813348 nps 5040000 tbhits 0 cpuload 494
info time 14062 nodes 71265743 nps 5067000 tbhits 0 cpuload 490
info time 15065 nodes 76246219 nps 5061000 tbhits 0 cpuload 499
info time 16069 nodes 80831277 nps 5030000 tbhits 0 cpuload 497
info time 17074 nodes 85897914 nps 5030000 tbhits 0 cpuload 494
info time 18078 nodes 90709890 nps 5017000 tbhits 0 cpuload 501
info time 19083 nodes 96381935 nps 5050000 tbhits 0 cpuload 490
info time 20088 nodes 101457881 nps 5050000 tbhits 0 cpuload 504
info time 21094 nodes 105632192 nps 5007000 tbhits 0 cpuload 494
info time 22099 nodes 111009209 nps 5023000 tbhits 0 cpuload 504
info time 23102 nodes 116607564 nps 5047000 tbhits 0 cpuload 497
info time 24107 nodes 121799644 nps 5052000 tbhits 0 cpuload 496
info time 25109 nodes 127566677 nps 5080000 tbhits 0 cpuload 498
info time 26110 nodes 133175811 nps 5100000 tbhits 0 cpuload 502
info time 27112 nodes 139087063 nps 5130000 tbhits 0 cpuload 498
info time 28117 nodes 144815287 nps 5150000 tbhits 0 cpuload 494
info time 29122 nodes 149962431 nps 5149000 tbhits 0 cpuload 498
Split_Depth 18, Position 2: 5121 kN/s
info time 1001 nodes 2534757 nps 2532000 tbhits 0 cpuload 496
info time 2007 nodes 6604282 nps 3290000 tbhits 0 cpuload 500
info time 3013 nodes 10853049 nps 3602000 tbhits 0 cpuload 494
info time 4017 nodes 14795889 nps 3683000 tbhits 0 cpuload 503
info time 5023 nodes 19658043 nps 3913000 tbhits 0 cpuload 496
info time 6027 nodes 25065048 nps 4158000 tbhits 0 cpuload 495
info time 7032 nodes 30446427 nps 4329000 tbhits 0 cpuload 502
info time 8039 nodes 35913541 nps 4467000 tbhits 0 cpuload 495
info time 9040 nodes 41017589 nps 4537000 tbhits 0 cpuload 504
info time 10081 nodes 45874002 nps 4550000 tbhits 0 cpuload 488
info time 11103 nodes 50737841 nps 4569000 tbhits 0 cpuload 501
info time 12109 nodes 55876920 nps 4614000 tbhits 0 cpuload 494
info time 13116 nodes 60431772 nps 4607000 tbhits 0 cpuload 499
info time 14121 nodes 65796360 nps 4659000 tbhits 0 cpuload 496
info time 15128 nodes 70664079 nps 4671000 tbhits 0 cpuload 503
info time 16131 nodes 75979925 nps 4710000 tbhits 0 cpuload 491
info time 17134 nodes 81734509 nps 4770000 tbhits 0 cpuload 503
info time 18138 nodes 87034578 nps 4798000 tbhits 0 cpuload 497
info time 19143 nodes 92575114 nps 4835000 tbhits 0 cpuload 496
info time 20148 nodes 98172625 nps 4872000 tbhits 0 cpuload 500
info time 21155 nodes 104002910 nps 4916000 tbhits 0 cpuload 495
info time 22160 nodes 108937322 nps 4915000 tbhits 0 cpuload 504
info time 23166 nodes 114061197 nps 4923000 tbhits 0 cpuload 492
info time 24167 nodes 119771381 nps 4955000 tbhits 0 cpuload 498
info time 25168 nodes 124190418 nps 4934000 tbhits 0 cpuload 496
info time 26169 nodes 129412248 nps 4945000 tbhits 0 cpuload 500
info time 27173 nodes 135127891 nps 4972000 tbhits 0 cpuload 491
info time 28174 nodes 140578632 nps 4989000 tbhits 0 cpuload 496
info time 29182 nodes 146315137 nps 5013000 tbhits 0 cpuload 502
Split_Depth 18, Position 3: 5036 kN/s
info time 1003 nodes 2271882 nps 2265000 tbhits 0 cpuload 495
info time 2008 nodes 4828984 nps 2404000 tbhits 0 cpuload 502
info time 3010 nodes 7892128 nps 2621000 tbhits 0 cpuload 494
info time 4011 nodes 12387733 nps 3088000 tbhits 0 cpuload 494
info time 5016 nodes 17345733 nps 3458000 tbhits 0 cpuload 498
info time 6018 nodes 23372620 nps 3883000 tbhits 0 cpuload 494
info time 7022 nodes 28480228 nps 4055000 tbhits 0 cpuload 495
info time 8028 nodes 33228282 nps 4139000 tbhits 0 cpuload 500
info time 9034 nodes 38696181 nps 4283000 tbhits 0 cpuload 494
info time 10041 nodes 43381949 nps 4320000 tbhits 0 cpuload 503
info time 11043 nodes 48133176 nps 4358000 tbhits 0 cpuload 494
info time 12055 nodes 52417408 nps 4348000 tbhits 0 cpuload 500
info time 13057 nodes 57616702 nps 4412000 tbhits 0 cpuload 498
info time 14062 nodes 62876605 nps 4471000 tbhits 0 cpuload 500
info time 15069 nodes 67994881 nps 4512000 tbhits 0 cpuload 493
info time 16073 nodes 73000535 nps 4541000 tbhits 0 cpuload 491
info time 17077 nodes 78717987 nps 4609000 tbhits 0 cpuload 503
info time 18081 nodes 84385195 nps 4667000 tbhits 0 cpuload 493
info time 19083 nodes 89738871 nps 4702000 tbhits 0 cpuload 498
info time 20087 nodes 94741807 nps 4716000 tbhits 0 cpuload 503
info time 21093 nodes 100644691 nps 4771000 tbhits 0 cpuload 494
info time 22098 nodes 106639172 nps 4825000 tbhits 0 cpuload 502
info time 23105 nodes 112493416 nps 4868000 tbhits 0 cpuload 487
info time 24110 nodes 118000019 nps 4894000 tbhits 0 cpuload 502
info time 25118 nodes 123453220 nps 4914000 tbhits 0 cpuload 493
info time 26124 nodes 128238479 nps 4908000 tbhits 0 cpuload 502
info time 27131 nodes 133517456 nps 4921000 tbhits 0 cpuload 495
info time 28137 nodes 138646118 nps 4927000 tbhits 0 cpuload 502
info time 29141 nodes 142961026 nps 4905000 tbhits 0 cpuload 491
Split_Depth 18, Position 4: 4924 kN/s

Summary:
=======================================================
Split_Depth 10, Position 1: 5343 kN/s
Split_Depth 10, Position 2: 5448 kN/s
Split_Depth 10, Position 3: 5252 kN/s
Split_Depth 10, Position 4: 5349 kN/s
Split_Depth 10 - Average speed: 5348 kN/s
=======================================================
Split_Depth 12, Position 1: 5318 kN/s
Split_Depth 12, Position 2: 5389 kN/s
Split_Depth 12, Position 3: 5270 kN/s
Split_Depth 12, Position 4: 5347 kN/s
Split_Depth 12 - Average speed: 5331 kN/s
=======================================================
Split_Depth 14, Position 1: 5395 kN/s
Split_Depth 14, Position 2: 5231 kN/s
Split_Depth 14, Position 3: 5166 kN/s
Split_Depth 14, Position 4: 5405 kN/s
Split_Depth 14 - Average speed: 5299 kN/s
=======================================================
Split_Depth 16, Position 1: 5120 kN/s
Split_Depth 16, Position 2: 5377 kN/s
Split_Depth 16, Position 3: 5092 kN/s
Split_Depth 16, Position 4: 5289 kN/s
Split_Depth 16 - Average speed: 5219 kN/s
=======================================================
Split_Depth 18, Position 1: 5046 kN/s
Split_Depth 18, Position 2: 5121 kN/s
Split_Depth 18, Position 3: 5036 kN/s
Split_Depth 18, Position 4: 4924 kN/s
Split_Depth 18 - Average speed: 5032 kN/s
=======================================================
Average CPU time per thread: 597.106 s
CPU-corrected average speed: 5271 kN/s
=======================================================


8 Threads Houdini 1.5 64bit

info time 8117 nodes 35452327 nps 4367000 tbhits 0 cpuload 972
info time 9124 nodes 41124676 nps 4507000 tbhits 0 cpuload 968
info time 10125 nodes 47501917 nps 4691000 tbhits 0 cpuload 995
info time 11130 nodes 53948942 nps 4847000 tbhits 0 cpuload 970
info time 12136 nodes 60209879 nps 4961000 tbhits 0 cpuload 975
info time 13145 nodes 64529027 nps 4909000 tbhits 0 cpuload 977
info time 14147 nodes 69510283 nps 4913000 tbhits 0 cpuload 984
info time 15151 nodes 75581874 nps 4988000 tbhits 0 cpuload 978
info time 16153 nodes 81472323 nps 5043000 tbhits 0 cpuload 975
info time 17156 nodes 86250915 nps 5027000 tbhits 0 cpuload 968
info time 18164 nodes 91279719 nps 5025000 tbhits 0 cpuload 978
info time 19171 nodes 97221958 nps 5071000 tbhits 0 cpuload 964
info time 20177 nodes 104009402 nps 5154000 tbhits 0 cpuload 982
info time 21190 nodes 110285440 nps 5204000 tbhits 0 cpuload 977
info time 22201 nodes 116131160 nps 5230000 tbhits 0 cpuload 979
info time 23210 nodes 120439002 nps 5189000 tbhits 0 cpuload 977
info time 24223 nodes 126979239 nps 5242000 tbhits 0 cpuload 974
info time 25237 nodes 132391739 nps 5245000 tbhits 0 cpuload 986
info time 26238 nodes 138976768 nps 5296000 tbhits 0 cpuload 968
info time 27249 nodes 144496734 nps 5302000 tbhits 0 cpuload 979
info time 28255 nodes 149382886 nps 5286000 tbhits 0 cpuload 978
info time 29259 nodes 155462706 nps 5313000 tbhits 0 cpuload 967
Split_Depth 14, Position 4: 5344 kN/s
info time 1006 nodes 2407222 nps 2392000 tbhits 0 cpuload 895
info time 2007 nodes 6160241 nps 3069000 tbhits 0 cpuload 979
info time 3011 nodes 9965842 nps 3309000 tbhits 0 cpuload 986
info time 4023 nodes 15063145 nps 3744000 tbhits 0 cpuload 969
info time 5031 nodes 19499164 nps 3875000 tbhits 0 cpuload 976
info time 6045 nodes 24818938 nps 4105000 tbhits 0 cpuload 984
info time 7059 nodes 30040041 nps 4255000 tbhits 0 cpuload 976
info time 8061 nodes 34665064 nps 4300000 tbhits 0 cpuload 975
info time 9067 nodes 40317189 nps 4446000 tbhits 0 cpuload 990
info time 10068 nodes 45308594 nps 4500000 tbhits 0 cpuload 972
info time 11069 nodes 51101876 nps 4616000 tbhits 0 cpuload 974
info time 12079 nodes 56633195 nps 4688000 tbhits 0 cpuload 975
info time 13080 nodes 61198068 nps 4678000 tbhits 0 cpuload 977
info time 14085 nodes 66532646 nps 4723000 tbhits 0 cpuload 975
info time 15096 nodes 71218927 nps 4717000 tbhits 0 cpuload 991
info time 16105 nodes 77239687 nps 4796000 tbhits 0 cpuload 964
info time 17125 nodes 82955811 nps 4844000 tbhits 0 cpuload 976
info time 18131 nodes 88253500 nps 4867000 tbhits 0 cpuload 994
info time 19140 nodes 91215984 nps 4765000 tbhits 0 cpuload 916
info time 20152 nodes 97623026 nps 4844000 tbhits 0 cpuload 975
info time 21159 nodes 104194018 nps 4924000 tbhits 0 cpuload 985
info time 22162 nodes 110501729 nps 4986000 tbhits 0 cpuload 975
info time 23166 nodes 116343050 nps 5022000 tbhits 0 cpuload 988
info time 24169 nodes 121722496 nps 5036000 tbhits 0 cpuload 974
info time 25186 nodes 126968345 nps 5041000 tbhits 0 cpuload 979
info time 26195 nodes 132594840 nps 5061000 tbhits 0 cpuload 985
info time 27205 nodes 139131845 nps 5114000 tbhits 0 cpuload 971
info time 28206 nodes 145546368 nps 5160000 tbhits 0 cpuload 987
info time 29214 nodes 151834890 nps 5197000 tbhits 0 cpuload 975
Split_Depth 16, Position 1: 5218 kN/s
info time 1001 nodes 2046365 nps 2044000 tbhits 0 cpuload 911
info time 2008 nodes 6824957 nps 3398000 tbhits 0 cpuload 958
info time 3017 nodes 11599020 nps 3844000 tbhits 0 cpuload 977
info time 4020 nodes 16039073 nps 3989000 tbhits 0 cpuload 958
info time 5027 nodes 21925199 nps 4361000 tbhits 0 cpuload 989
info time 6037 nodes 27988027 nps 4636000 tbhits 0 cpuload 976
info time 7048 nodes 32836573 nps 4658000 tbhits 0 cpuload 974
info time 8057 nodes 39098244 nps 4852000 tbhits 0 cpuload 983
info time 9063 nodes 43477584 nps 4797000 tbhits 0 cpuload 990
info time 10070 nodes 48689442 nps 4835000 tbhits 0 cpuload 977
info time 11073 nodes 53551672 nps 4836000 tbhits 0 cpuload 964
info time 12080 nodes 59631885 nps 4936000 tbhits 0 cpuload 985
info time 13088 nodes 65095038 nps 4973000 tbhits 0 cpuload 980
info time 14090 nodes 69707116 nps 4947000 tbhits 0 cpuload 980
info time 15093 nodes 75145584 nps 4978000 tbhits 0 cpuload 970
info time 16100 nodes 81566110 nps 5066000 tbhits 0 cpuload 989
info time 17106 nodes 87480411 nps 5114000 tbhits 0 cpuload 973
info time 18110 nodes 91931002 nps 5076000 tbhits 0 cpuload 982
info time 19113 nodes 98227791 nps 5139000 tbhits 0 cpuload 964
info time 20116 nodes 104826214 nps 5211000 tbhits 0 cpuload 972
info time 21117 nodes 111381994 nps 5274000 tbhits 0 cpuload 979
info time 22122 nodes 117902770 nps 5329000 tbhits 0 cpuload 981
info time 23131 nodes 124480970 nps 5381000 tbhits 0 cpuload 972
info time 24137 nodes 130241503 nps 5395000 tbhits 0 cpuload 982
info time 25142 nodes 134733877 nps 5358000 tbhits 0 cpuload 981
info time 26143 nodes 140332063 nps 5367000 tbhits 0 cpuload 966
info time 27150 nodes 145000093 nps 5340000 tbhits 0 cpuload 991
info time 28153 nodes 151487669 nps 5380000 tbhits 0 cpuload 974
info time 29159 nodes 157804592 nps 5411000 tbhits 0 cpuload 978
Split_Depth 16, Position 2: 5436 kN/s
info time 1004 nodes 2289418 nps 2280000 tbhits 0 cpuload 947
info time 2006 nodes 6523758 nps 3252000 tbhits 0 cpuload 994
info time 3011 nodes 10650657 nps 3537000 tbhits 0 cpuload 970
info time 4016 nodes 15806509 nps 3935000 tbhits 0 cpuload 974
info time 5026 nodes 20784628 nps 4135000 tbhits 0 cpuload 975
info time 6033 nodes 25525467 nps 4230000 tbhits 0 cpuload 983
info time 7039 nodes 30383910 nps 4316000 tbhits 0 cpuload 959
info time 8051 nodes 35859748 nps 4454000 tbhits 0 cpuload 986
info time 9053 nodes 40515007 nps 4475000 tbhits 0 cpuload 980
info time 10066 nodes 45511021 nps 4521000 tbhits 0 cpuload 975
info time 11074 nodes 50502956 nps 4560000 tbhits 0 cpuload 998
info time 12096 nodes 54762562 nps 4527000 tbhits 0 cpuload 973
info time 13107 nodes 60198722 nps 4592000 tbhits 0 cpuload 987
info time 14112 nodes 66157484 nps 4688000 tbhits 0 cpuload 981
info time 15136 nodes 71142112 nps 4700000 tbhits 0 cpuload 973
info time 16142 nodes 76818875 nps 4758000 tbhits 0 cpuload 976
info time 17148 nodes 82352132 nps 4802000 tbhits 0 cpuload 975
info time 18157 nodes 87640058 nps 4826000 tbhits 0 cpuload 981
info time 19191 nodes 93757064 nps 4885000 tbhits 0 cpuload 975
info time 20223 nodes 99490743 nps 4919000 tbhits 0 cpuload 982
info time 21230 nodes 105593298 nps 4973000 tbhits 0 cpuload 983
info time 22235 nodes 110617942 nps 4974000 tbhits 0 cpuload 979
info time 23244 nodes 115178242 nps 4955000 tbhits 0 cpuload 985
info time 24254 nodes 120829693 nps 4981000 tbhits 0 cpuload 988
info time 25266 nodes 126400713 nps 5002000 tbhits 0 cpuload 976
info time 26272 nodes 132833501 nps 5056000 tbhits 0 cpuload 969
info time 27291 nodes 138175433 nps 5063000 tbhits 0 cpuload 972
info time 28299 nodes 143877690 nps 5084000 tbhits 0 cpuload 984
info time 29317 nodes 149745688 nps 5107000 tbhits 0 cpuload 971
Split_Depth 16, Position 3: 5097 kN/s
info time 1001 nodes 1559702 nps 1558000 tbhits 0 cpuload 905
info time 2006 nodes 4356097 nps 2171000 tbhits 0 cpuload 952
info time 3011 nodes 8927234 nps 2964000 tbhits 0 cpuload 993
info time 4015 nodes 13887537 nps 3458000 tbhits 0 cpuload 971
info time 5024 nodes 17656829 nps 3514000 tbhits 0 cpuload 987
info time 6026 nodes 21107356 nps 3502000 tbhits 0 cpuload 976
info time 7028 nodes 26184931 nps 3725000 tbhits 0 cpuload 980
info time 8036 nodes 32364128 nps 4027000 tbhits 0 cpuload 990
info time 9051 nodes 37886140 nps 4185000 tbhits 0 cpuload 966
info time 10056 nodes 44127804 nps 4388000 tbhits 0 cpuload 977
info time 11068 nodes 50037067 nps 4520000 tbhits 0 cpuload 971
info time 12076 nodes 55018449 nps 4556000 tbhits 0 cpuload 986
info time 13081 nodes 60884998 nps 4654000 tbhits 0 cpuload 972
info time 14082 nodes 66203053 nps 4701000 tbhits 0 cpuload 979
info time 15085 nodes 72840654 nps 4828000 tbhits 0 cpuload 997
info time 16095 nodes 79238257 nps 4923000 tbhits 0 cpuload 976
info time 17103 nodes 85698839 nps 5010000 tbhits 0 cpuload 961
info time 18111 nodes 91394180 nps 5046000 tbhits 0 cpuload 982
info time 19117 nodes 97863778 nps 5119000 tbhits 0 cpuload 959
info time 20126 nodes 103877891 nps 5161000 tbhits 0 cpuload 970
info time 21134 nodes 110016350 nps 5205000 tbhits 0 cpuload 990
info time 22135 nodes 115802464 nps 5231000 tbhits 0 cpuload 964
info time 23142 nodes 122624143 nps 5298000 tbhits 0 cpuload 960
info time 24146 nodes 129159894 nps 5349000 tbhits 0 cpuload 969
info time 25148 nodes 135534720 nps 5389000 tbhits 0 cpuload 978
info time 26153 nodes 142183075 nps 5436000 tbhits 0 cpuload 970
info time 27162 nodes 148833883 nps 5479000 tbhits 0 cpuload 985
info time 28168 nodes 154892361 nps 5498000 tbhits 0 cpuload 965
info time 29174 nodes 159579881 nps 5469000 tbhits 0 cpuload 973
Split_Depth 16, Position 4: 5482 kN/s
info time 1021 nodes 1808865 nps 1771000 tbhits 0 cpuload 966
info time 2034 nodes 4780173 nps 2350000 tbhits 0 cpuload 970
info time 3035 nodes 7800376 nps 2570000 tbhits 0 cpuload 985
info time 4044 nodes 11595462 nps 2867000 tbhits 0 cpuload 970
info time 5053 nodes 17137016 nps 3391000 tbhits 0 cpuload 983
info time 6055 nodes 22651727 nps 3740000 tbhits 0 cpuload 975
info time 7058 nodes 26974451 nps 3821000 tbhits 0 cpuload 987
info time 8076 nodes 32604136 nps 4037000 tbhits 0 cpuload 978
info time 9079 nodes 35837017 nps 3947000 tbhits 0 cpuload 890
info time 10090 nodes 40333723 nps 3997000 tbhits 0 cpuload 981
info time 11101 nodes 46644352 nps 4201000 tbhits 0 cpuload 981
info time 12111 nodes 51780901 nps 4275000 tbhits 0 cpuload 971
info time 13117 nodes 56655433 nps 4319000 tbhits 0 cpuload 990
info time 14121 nodes 61541219 nps 4358000 tbhits 0 cpuload 971
info time 15136 nodes 66381615 nps 4385000 tbhits 0 cpuload 958
info time 16142 nodes 70799055 nps 4386000 tbhits 0 cpuload 986
info time 17145 nodes 77113737 nps 4497000 tbhits 0 cpuload 979
info time 18151 nodes 83446043 nps 4597000 tbhits 0 cpuload 990
info time 19163 nodes 88749957 nps 4631000 tbhits 0 cpuload 959
info time 20171 nodes 94661974 nps 4692000 tbhits 0 cpuload 986
info time 21186 nodes 99221200 nps 4683000 tbhits 0 cpuload 972
info time 22197 nodes 105403144 nps 4748000 tbhits 0 cpuload 972
info time 23200 nodes 111592436 nps 4810000 tbhits 0 cpuload 977
info time 24219 nodes 117470512 nps 4850000 tbhits 0 cpuload 987
info time 25246 nodes 124155204 nps 4917000 tbhits 0 cpuload 972
info time 26258 nodes 130203205 nps 4958000 tbhits 0 cpuload 978
info time 27268 nodes 136425202 nps 5003000 tbhits 0 cpuload 986
info time 28277 nodes 142061459 nps 5023000 tbhits 0 cpuload 970
info time 29278 nodes 148226451 nps 5062000 tbhits 0 cpuload 983
Split_Depth 18, Position 1: 5102 kN/s
info time 1006 nodes 2302938 nps 2289000 tbhits 0 cpuload 895
info time 2017 nodes 6570554 nps 3257000 tbhits 0 cpuload 975
info time 3019 nodes 11559561 nps 3828000 tbhits 0 cpuload 978
info time 4026 nodes 17598393 nps 4371000 tbhits 0 cpuload 991
info time 5035 nodes 22221580 nps 4413000 tbhits 0 cpuload 977
info time 6042 nodes 26863279 nps 4446000 tbhits 0 cpuload 970
info time 7054 nodes 31621416 nps 4482000 tbhits 0 cpuload 967
info time 8060 nodes 36787933 nps 4564000 tbhits 0 cpuload 973
info time 9067 nodes 42669351 nps 4706000 tbhits 0 cpuload 981
info time 10075 nodes 48364323 nps 4800000 tbhits 0 cpuload 980
info time 11082 nodes 54159211 nps 4887000 tbhits 0 cpuload 968
info time 12092 nodes 60082768 nps 4968000 tbhits 0 cpuload 988
info time 13098 nodes 66538614 nps 5080000 tbhits 0 cpuload 963
info time 14106 nodes 72986110 nps 5174000 tbhits 0 cpuload 980
info time 15108 nodes 79387994 nps 5254000 tbhits 0 cpuload 973
info time 16115 nodes 84963379 nps 5272000 tbhits 0 cpuload 985
info time 17120 nodes 91140266 nps 5323000 tbhits 0 cpuload 975
info time 18132 nodes 97539839 nps 5379000 tbhits 0 cpuload 971
info time 19133 nodes 103961867 nps 5433000 tbhits 0 cpuload 974
info time 20136 nodes 110215596 nps 5473000 tbhits 0 cpuload 987
info time 21142 nodes 116646302 nps 5517000 tbhits 0 cpuload 976
info time 22158 nodes 123106816 nps 5555000 tbhits 0 cpuload 971
info time 23168 nodes 129488829 nps 5589000 tbhits 0 cpuload 990
info time 24178 nodes 136011574 nps 5625000 tbhits 0 cpuload 978
info time 25188 nodes 142757488 nps 5667000 tbhits 0 cpuload 978
info time 26196 nodes 149223308 nps 5696000 tbhits 0 cpuload 984
info time 27199 nodes 155546277 nps 5718000 tbhits 0 cpuload 966
info time 28209 nodes 162069485 nps 5745000 tbhits 0 cpuload 975
info time 29212 nodes 168441990 nps 5766000 tbhits 0 cpuload 966
Split_Depth 18, Position 2: 5781 kN/s
info time 1033 nodes 1769954 nps 1713000 tbhits 0 cpuload 972
info time 2045 nodes 5030055 nps 2459000 tbhits 0 cpuload 982
info time 3052 nodes 8697804 nps 2849000 tbhits 0 cpuload 968
info time 4108 nodes 12312463 nps 2997000 tbhits 0 cpuload 980
info time 5124 nodes 15325114 nps 2990000 tbhits 0 cpuload 965
info time 6125 nodes 19292987 nps 3149000 tbhits 0 cpuload 981
info time 7133 nodes 23284947 nps 3264000 tbhits 0 cpuload 986
info time 8136 nodes 29377409 nps 3610000 tbhits 0 cpuload 962
info time 9137 nodes 34196283 nps 3742000 tbhits 0 cpuload 979
info time 10140 nodes 39082022 nps 3854000 tbhits 0 cpuload 981
info time 11153 nodes 43855663 nps 3932000 tbhits 0 cpuload 987
info time 12158 nodes 47986592 nps 3946000 tbhits 0 cpuload 968
info time 13165 nodes 54077734 nps 4107000 tbhits 0 cpuload 985
info time 14166 nodes 58879324 nps 4156000 tbhits 0 cpuload 975
info time 15296 nodes 62914031 nps 4113000 tbhits 0 cpuload 973
info time 16317 nodes 67545894 nps 4139000 tbhits 0 cpuload 975
info time 17321 nodes 73628444 nps 4250000 tbhits 0 cpuload 967
info time 18330 nodes 79260257 nps 4324000 tbhits 0 cpuload 987
info time 19337 nodes 85411914 nps 4417000 tbhits 0 cpuload 966
info time 20344 nodes 91601070 nps 4502000 tbhits 0 cpuload 989
info time 21346 nodes 97521928 nps 4568000 tbhits 0 cpuload 971
info time 22355 nodes 103611225 nps 4634000 tbhits 0 cpuload 993
info time 23362 nodes 108973005 nps 4664000 tbhits 0 cpuload 972
info time 24370 nodes 115237222 nps 4728000 tbhits 0 cpuload 992
info time 25373 nodes 121463421 nps 4787000 tbhits 0 cpuload 975
info time 26379 nodes 126431967 nps 4792000 tbhits 0 cpuload 978
info time 27388 nodes 130854244 nps 4777000 tbhits 0 cpuload 975
info time 28399 nodes 135572047 nps 4773000 tbhits 0 cpuload 966
info time 29410 nodes 141980515 nps 4827000 tbhits 0 cpuload 979
Split_Depth 18, Position 3: 4853 kN/s
info time 1001 nodes 1687681 nps 1685000 tbhits 0 cpuload 956
info time 2008 nodes 3894986 nps 1939000 tbhits 0 cpuload 974
info time 3021 nodes 7829980 nps 2591000 tbhits 0 cpuload 972
info time 4031 nodes 11843311 nps 2938000 tbhits 0 cpuload 978
info time 5040 nodes 17997678 nps 3570000 tbhits 0 cpuload 977
info time 6043 nodes 23621362 nps 3908000 tbhits 0 cpuload 975
info time 7053 nodes 28522591 nps 4044000 tbhits 0 cpuload 986
info time 8054 nodes 34872271 nps 4329000 tbhits 0 cpuload 979
info time 9060 nodes 40560636 nps 4476000 tbhits 0 cpuload 976
info time 10062 nodes 45432935 nps 4515000 tbhits 0 cpuload 969
info time 11070 nodes 50139839 nps 4529000 tbhits 0 cpuload 980
info time 12075 nodes 56789620 nps 4703000 tbhits 0 cpuload 977
info time 13079 nodes 61857215 nps 4729000 tbhits 0 cpuload 984
info time 14080 nodes 68108796 nps 4837000 tbhits 0 cpuload 974
info time 15082 nodes 74921261 nps 4967000 tbhits 0 cpuload 980
info time 16090 nodes 81700197 nps 5077000 tbhits 0 cpuload 971
info time 17097 nodes 86840410 nps 5079000 tbhits 0 cpuload 970
info time 18100 nodes 91446219 nps 5052000 tbhits 0 cpuload 991
info time 19105 nodes 97438794 nps 5100000 tbhits 0 cpuload 970
info time 20107 nodes 103936425 nps 5169000 tbhits 0 cpuload 976
info time 21112 nodes 110453838 nps 5231000 tbhits 0 cpuload 987
info time 22119 nodes 117176407 nps 5297000 tbhits 0 cpuload 975
info time 23122 nodes 123806793 nps 5354000 tbhits 0 cpuload 981
info time 24126 nodes 129489973 nps 5367000 tbhits 0 cpuload 973
info time 25135 nodes 134259294 nps 5341000 tbhits 0 cpuload 983
info time 26140 nodes 139973598 nps 5354000 tbhits 0 cpuload 972
info time 27145 nodes 144632053 nps 5328000 tbhits 0 cpuload 981
info time 28148 nodes 149595347 nps 5314000 tbhits 0 cpuload 979
info time 29155 nodes 155449521 nps 5331000 tbhits 0 cpuload 972
Split_Depth 18, Position 4: 5360 kN/s

Summary:
=======================================================
Split_Depth 10, Position 1: 5609 kN/s
Split_Depth 10, Position 2: 5887 kN/s
Split_Depth 10, Position 3: 5446 kN/s
Split_Depth 10, Position 4: 5584 kN/s
Split_Depth 10 - Average speed: 5631 kN/s
=======================================================
Split_Depth 12, Position 1: 5577 kN/s
Split_Depth 12, Position 2: 5764 kN/s
Split_Depth 12, Position 3: 5596 kN/s
Split_Depth 12, Position 4: 5425 kN/s
Split_Depth 12 - Average speed: 5590 kN/s
=======================================================
Split_Depth 14, Position 1: 5486 kN/s
Split_Depth 14, Position 2: 5916 kN/s
Split_Depth 14, Position 3: 5455 kN/s
Split_Depth 14, Position 4: 5344 kN/s
Split_Depth 14 - Average speed: 5550 kN/s
=======================================================
Split_Depth 16, Position 1: 5218 kN/s
Split_Depth 16, Position 2: 5436 kN/s
Split_Depth 16, Position 3: 5097 kN/s
Split_Depth 16, Position 4: 5482 kN/s
Split_Depth 16 - Average speed: 5308 kN/s
=======================================================
Split_Depth 18, Position 1: 5102 kN/s
Split_Depth 18, Position 2: 5781 kN/s
Split_Depth 18, Position 3: 4853 kN/s
Split_Depth 18, Position 4: 5360 kN/s
Split_Depth 18 - Average speed: 5274 kN/s
=======================================================
Average CPU time per thread: 585.425 s
CPU-corrected average speed: 5607 kN/s
=======================================================

User avatar
AdminX
Posts: 5613
Joined: Mon Mar 13, 2006 1:34 pm
Location: Acworth, GA
Contact:

Re: i7 Hyper-threading

Post by AdminX » Fri Dec 24, 2010 2:12 pm

Houdini wrote:
mwyoung wrote:After loading my new i7 laptop with a few chess engines. I am finding Rybka 4, Fritz 12, and Houdini are running faster using 8 logical cores. This is true with NPS and timed test positions. As this is my first i7 cpu, does Hyper-threading with the i7 help instead of hurt chess engines performance. This was not true in the past. In my testing I am clearly faster with Hyper-threading turned on running my chess engines.

Example Fritz 12

Fritz benchmark 4 cores. 5380 Nps 11.21

Fritz benchmark with HT 8 logical cores. 6995 Nps 14.57
Mark,

I've never experienced any useful improvement from hyper-threading for Houdini. Even if the nps is slightly higher, the overhead of the additional threads could very well reduce the actual playing strength.

Could you try the "autotune" feature of Houdini? This feature was intended for picking the best Split_Depth parameter, but it also functions as an accurate benchmark for multi-thread node speeds.

Double-click on the Houdini executable to open it in a console window.
Then enter the following commands to "autotune" for 4 threads.

Code: Select all

setoption name threads value 4
autotune
Houdini will run for about 10 minutes analyzing a number of positions for different values of split_depth. Make sure you're doing nothing else on your computer, just leave it running for 10 minutes. At the end you'll get a summary of the node speeds.

Then repeat the procedure for 8 threads.

Code: Select all

setoption name threads value 8
autotune
What results do you get?
Here are my results from a Intel (R) Core (TM)2 Quad Q6700 @2.66 GHz 2.67GHz with 8GB Ram running Windows 7 Pro (64-bit). What should Split_Depth be set to? 10, 14, or Other ....

Summary:
=======================================================
Split_Depth 10, Position 1: 5859 kN/s
Split_Depth 10, Position 2: 5866 kN/s
Split_Depth 10, Position 3: 5881 kN/s
Split_Depth 10, Position 4: 5982 kN/s
Split_Depth 10 - Average speed: 5897 kN/s
=======================================================
Split_Depth 12, Position 1: 5926 kN/s
Split_Depth 12, Position 2: 5961 kN/s
Split_Depth 12, Position 3: 5505 kN/s
Split_Depth 12, Position 4: 5897 kN/s
Split_Depth 12 - Average speed: 5822 kN/s
=======================================================
Split_Depth 14, Position 1: 5752 kN/s
Split_Depth 14, Position 2: 5897 kN/s
Split_Depth 14, Position 3: 5784 kN/s
Split_Depth 14, Position 4: 5994 kN/s
Split_Depth 14 - Average speed: 5857 kN/s
=======================================================
Split_Depth 16, Position 1: 5710 kN/s
Split_Depth 16, Position 2: 5821 kN/s
Split_Depth 16, Position 3: 5652 kN/s
Split_Depth 16, Position 4: 5554 kN/s
Split_Depth 16 - Average speed: 5684 kN/s
=======================================================
Split_Depth 18, Position 1: 5546 kN/s
Split_Depth 18, Position 2: 5585 kN/s
Split_Depth 18, Position 3: 5377 kN/s
Split_Depth 18, Position 4: 5309 kN/s
Split_Depth 18 - Average speed: 5454 kN/s
=======================================================
Average CPU time per thread: 588.436 s
CPU-corrected average speed: 5856 kN/s
=======================================================
Last edited by AdminX on Fri Dec 24, 2010 2:15 pm, edited 2 times in total.
"Good decisions come from experience, and experience comes from bad decisions."
__________________________________________________________________
Ted Summers

mwyoung
Posts: 2727
Joined: Wed May 12, 2010 8:00 pm

Re: i7 Hyper-threading

Post by mwyoung » Fri Dec 24, 2010 2:14 pm

Houdini wrote:
mwyoung wrote:After loading my new i7 laptop with a few chess engines. I am finding Rybka 4, Fritz 12, and Houdini are running faster using 8 logical cores. This is true with NPS and timed test positions. As this is my first i7 cpu, does Hyper-threading with the i7 help instead of hurt chess engines performance. This was not true in the past. In my testing I am clearly faster with Hyper-threading turned on running my chess engines.

Example Fritz 12

Fritz benchmark 4 cores. 5380 Nps 11.21

Fritz benchmark with HT 8 logical cores. 6995 Nps 14.57
Mark,

I've never experienced any useful improvement from hyper-threading for Houdini. Even if the nps is slightly higher, the overhead of the additional threads could very well reduce the actual playing strength.

Could you try the "autotune" feature of Houdini? This feature was intended for picking the best Split_Depth parameter, but it also functions as an accurate benchmark for multi-thread node speeds.

Double-click on the Houdini executable to open it in a console window.
Then enter the following commands to "autotune" for 4 threads.

Code: Select all

setoption name threads value 4
autotune
Houdini will run for about 10 minutes analyzing a number of positions for different values of split_depth. Make sure you're doing nothing else on your computer, just leave it running for 10 minutes. At the end you'll get a summary of the node speeds.

Then repeat the procedure for 8 threads.

Code: Select all

setoption name threads value 8
autotune
What results do you get?
Looks like houdini is not hurt by HT with the i7 and even helped a bit. I see no reason not to use HT with any of the chess program I have tried with the i7 Mobile CPU. It does help some chess programs more then others, but no chess program has been slowed down by using HT.

User avatar
Houdini
Posts: 1471
Joined: Mon Mar 15, 2010 11:00 pm
Contact:

Re: i7 Hyper-threading

Post by Houdini » Fri Dec 24, 2010 2:40 pm

mwyoung wrote:Looks like houdini is not hurt by HT with the i7 and even helped a bit. I see no reason not to use HT with any of the chess program I have tried with the i7 Mobile CPU. It does help some chess programs more then others, but no chess program has been slowed down by using HT.
As the numbers in this topic show, and also in my experience, the HT will indeed increase the node speed by about 5% to 15%. But hardly any of this is going to translate into real strength, the general efficiency of the SMP alpha-beta algorithm is not very high - the problem is that we're squeezing a serial algorithm into a parallel scheme.

Given the choice between a 4-thread engine running at 8,000 kN/s, and a 8-thread engine running at 9,000 kN/s, I expect that the 4-thread will be stronger despite the lower node count. The only way to know for sure would be to run a test match between a Houdini 4T and Houdini 4T+4HT.

Then there are other considerations like the power consumption and the capacity for overclocking that are much more favorable without HT.

Robert

User avatar
Houdini
Posts: 1471
Joined: Mon Mar 15, 2010 11:00 pm
Contact:

Re: i7 Hyper-threading

Post by Houdini » Fri Dec 24, 2010 2:45 pm

AdminX wrote:Here are my results from a Intel (R) Core (TM)2 Quad Q6700 @2.66 GHz 2.67GHz with 8GB Ram running Windows 7 Pro (64-bit). What should Split_Depth be set to? 10, 14, or Other ....
Ted, for your configuration the default value of 10 appears to be best.
In my experience the best value for 4 cores is nearly always 10, but for 8 cores usually 12 will be better.

Post Reply