| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: BOINC Agent Support
Thread: Bogus Bulldozer BOINC Benchmarks - blooking blor Badvice  |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 38
|
|
| Author |
|
|
kateiacy
Veteran Cruncher USA Joined: Jan 23, 2010 Post Count: 1027 Status: Offline Project Badges: 
|
Very, very interesting, Dave! In my recent experience, C4CW WUs usually end up getting around 46 points or so, so I think the 1055T is underclaiming (no idea why).
----------------------------------------I'll be very interested to see your results under Xubuntu. Btw, I read that there was a Windows patch that did (or did not, depending on whom you read) make a big difference in BD performance. Were you using it when running under Windows? Kate  |
|||
|
||||
|
David Autumns
Ace Cruncher UK Joined: Nov 16, 2004 Post Count: 11062 Status: Offline Project Badges:
|
c4cw_ target05_ 262658982_ 1-- 641 Valid 29/04/12 13:41:18 29/04/12 17:53:15 2.41 45.8 / 45.9
----------------------------------------c4cw_ target05_ 262658982_ 0-- 641 Valid 29/04/12 13:17:14 29/04/12 20:17:18 1.74 39.5 / 45.9 Here's the first valid result Mines the one requesting 39.5 points I'm now at a loss to explain why we both ended up with, you are right, almost 46 points for a C4CW result. It just gets worse  --- Xubuntu is now fully installed (on the FX8150) with the latest version 3 kernel. It's currently crunching away through it's first batch of C4CW WU's - I'll let you know. the latest Benchmark result 2538 and 12411 - compare that with 1620 and 6702 on Windows Server 2008 R2 !  |
|||
|
||||
|
David Autumns
Ace Cruncher UK Joined: Nov 16, 2004 Post Count: 11062 Status: Offline Project Badges:
|
kateiacy
----------------------------------------I think the patch is to improve the way the scheduler allocates tasks to the BD Cores Normally the first task should go to core 1 and the 2nd to 2 and so on However to get the best results from the BD, as the FPU is shared between the 2 cores, they have to be allocated as below The first task should go to 1 the second to 3 3rd to 5 and 4th to 7 then the 5th to 2 and so on .. this only matters if there are still unused cores just twidddling their thumbs waiting for work to do - throw in turbo core and it's even more confusing. But as you know unused cores. and consequently turbo core, are an anathema to a WCG Cruncher  [Edit 2 times, last edit by David Autumns at Apr 30, 2012 6:39:31 AM] |
|||
|
||||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
My 8150 benchmarks a little higher but using W7-64 SP1 and processor at stock speed.
/21/2012 11:27:16 AM | | Processor: 8 AuthenticAMD AMD FX(tm)-8150 Eight-Core Processor [Family 21 Model 1 Stepping 2] 4/21/2012 11:27:16 AM | | Processor: 2.00 MB cache 4/21/2012 11:27:16 AM | | Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 htt pni ssse3 cx16 sse4_1 sse4_2 syscall nx lm svm sse4a osvw ibs xop skinit wdt lwp fma4 page1gb rdtscp 4/21/2012 11:27:16 AM | | OS: Microsoft Windows 7: Professional x64 Edition, Service Pack 1, (06.01.7601.00) 4/21/2012 11:27:16 AM | | Memory: 15.95 GB physical, 31.89 GB virtual 4/21/2012 11:27:16 AM | | Disk: 443.23 GB total, 385.88 GB free 4/21/2012 11:27:16 AM | | Local time is UTC -5 hours 4/21/2012 11:27:16 AM | | ATI GPU 0: Cayman (CAL version 1.4.1703, 2048MB, 2016MB available, 5914 GFLOPS peak) 4/21/2012 11:27:16 AM | | OpenCL: ATI GPU 0: Cayman (driver version CAL 1.4.1703 (VM), device version OpenCL 1.1 AMD-APP (898.1), 2048MB, 2016MB available) 4/21/2012 11:27:16 AM | | Version change (7.0.3 -> 7.0.25) 4/21/2012 11:27:16 AM | World Community Grid | URL http://www.worldcommunitygrid.org/; Computer ID 1760669; resource share 100 4/21/2012 11:27:16 AM | World Community Grid | General prefs: from World Community Grid (last modified 21-Apr-2012 08:50:43) 4/21/2012 11:27:16 AM | World Community Grid | Computer location: work 4/21/2012 11:27:16 AM | | General prefs: using separate prefs for work 4/21/2012 11:27:16 AM | | Reading preferences override file 4/21/2012 11:27:16 AM | | Preferences: 4/21/2012 11:27:16 AM | | max memory usage when active: 16328.28MB 4/21/2012 11:27:16 AM | | max memory usage when idle: 16328.28MB 4/21/2012 11:27:16 AM | | max disk usage: 10.00GB 4/21/2012 11:27:16 AM | | (to change preferences, visit the web site of an attached project, or select Preferences in the Manager) 4/21/2012 11:27:16 AM | | Not using a proxy 4/21/2012 11:27:17 AM | | Running CPU benchmarks 4/21/2012 11:27:17 AM | | Suspending computation - CPU benchmarks in progress 4/21/2012 11:27:48 AM | | Benchmark results: 4/21/2012 11:27:48 AM | | Number of CPUs: 8 4/21/2012 11:27:48 AM | | 2741 floating point MIPS (Whetstone) per CPU 4/21/2012 11:27:48 AM | | 10428 integer MIPS (Dhrystone) per CPU |
|||
|
||||
|
David Autumns
Ace Cruncher UK Joined: Nov 16, 2004 Post Count: 11062 Status: Offline Project Badges:
|
Cheers Doneske
---------------------------------------- I hope you are loving the 8150 as much as I am - comparing them to the PhenomII and the small sample of results we have above that is pretty close to what you might expect in Windows 7 for that processor at stock. So that is reassuring. Here's the fully patched up and updated Xubuntu results c4cw_ target05_ 263381590_ 1-- 641 Valid 29/04/12 22:36:53 30/04/12 04:13:47 1.92 46.7 / 45.5 c4cw_ target05_ 263381590_ 0-- 640 Valid 29/04/12 22:35:27 30/04/12 05:47:06 2.30 46.1 / 45.5 Mines the one that ran for 2.3 hours (compare with 1.49 hours in 2008 R2) and ended up with 45.5 points. At first glance it would appear that "lightweight and efficient" Xubuntu Linux is less effective than the "behemoth bloatware" Windows 2008 Server - who would have thought  and it would appear the Benchmark results count for nothing (certainly in C4CW) Stunned silence falls across the nations ...and look there it is....the Tumbleweed. Make it a great day Dave [Edit 2 times, last edit by David Autumns at Apr 30, 2012 6:27:58 AM] |
|||
|
||||
|
kateiacy
Veteran Cruncher USA Joined: Jan 23, 2010 Post Count: 1027 Status: Offline Project Badges:
|
Just to muddy the water still more, here are the most recent results from my AMD 910e on C4CW running under Xubuntu 12.04:
----------------------------------------c4cw_ target05_ 261468117_ 0--Name Valid 4/29/12 00:21:53 4/29/12 21:50:29 2.07 45.0 / 45.6 c4cw_ target05_ 261466882_ 0-- Name Valid 4/29/12 00:21:51 4/29/12 21:50:29 2.06 44.9 / 45.6 c4cw_ target05_ 261602580_ 0-- Name Valid 4/29/12 01:13:02 4/29/12 21:50:29 2.39 60.1 / 45.6 c4cw_ target05_ 261616461_ 0--Name Valid 4/29/12 01:13:01 4/29/12 21:50:29 2.35 58.9 / 45.6 c4cw_ target05_ 261606864_ 0--Name Valid 4/29/12 01:13:01 4/29/12 21:50:29 2.36 59.1 / 45.6 c4cw_ target05_ 261612545_ 0-- Name Valid 4/29/12 01:13:02 4/29/12 21:50:29 2.36 59.1 / 45.6 It looks as if the machine thinks there are 2 different sizes of C4CW WUs, one running .3 hrs longer and producing a bigger points claim than the other. I wonder whether your BD is doing the same. It's hit or miss which sciences run faster under Windows vs Linux. I think it has more to do with compiler optimization at this point than with the efficiency of the OS. |
|||
|
||||
|
mikey
Veteran Cruncher Joined: May 10, 2009 Post Count: 824 Status: Offline Project Badges:
|
FX8150 + Buckets of DDR3-1600 RAM OC'ed to 4Ghz (Windows Server 2008 R2) 1620 FP MIPS and 6702 Integer MIPS exactly the same hardware on Xubuntu 2542 FP MIPS and 11236 Integer MIPS How bizarre ! I wonder what I would get if I paid for another Windows 7 We'll never know..... Do you have Win7 set on Performance? Start, type "power settings" no quotes, check high performance. These are from my 1090T 6 core: 4/30/2012 10:20:17 AM | | Running CPU benchmarks 4/30/2012 10:20:18 AM | | Suspending computation - CPU benchmarks in progress 4/30/2012 10:20:49 AM | | Benchmark results: 4/30/2012 10:20:49 AM | | Number of CPUs: 6 4/30/2012 10:20:49 AM | | 2647 floating point MIPS (Whetstone) per CPU 4/30/2012 10:20:49 AM | | 8956 integer MIPS (Dhrystone) per CPU 4/30/2012 10:20:51 AM | | Resuming computation and these are from my FX-6100: 4/30/2012 10:20:17 AM | | Running CPU benchmarks 4/30/2012 10:20:18 AM | | Suspending computation - CPU benchmarks in progress 4/30/2012 10:20:49 AM | | Benchmark results: 4/30/2012 10:20:49 AM | | Number of CPUs: 6 4/30/2012 10:20:49 AM | | 2628 floating point MIPS (Whetstone) per CPU 4/30/2012 10:20:49 AM | | 9474 integer MIPS (Dhrystone) per CPU 4/30/2012 10:20:51 AM | | Resuming computation BOTH are done with the settings at High Performance.   [Edit 1 times, last edit by mikey159b at Apr 30, 2012 2:26:12 PM] |
|||
|
||||
|
David Autumns
Ace Cruncher UK Joined: Nov 16, 2004 Post Count: 11062 Status: Offline Project Badges:
|
It was on balanced so I have set it to Performance and...
----------------------------------------30/04/2012 17:29:05 Running CPU benchmarks 30/04/2012 17:29:06 Suspending computation - running CPU benchmarks 30/04/2012 17:29:36 Benchmark results: 30/04/2012 17:29:36 Number of CPUs: 8 30/04/2012 17:29:36 1620 floating point MIPS (Whetstone) per CPU 30/04/2012 17:29:36 6995 integer MIPS (Dhrystone) per CPU still have my missing mips  |
|||
|
||||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
OS: Centos 6.0
Client: Linux 6.10 AMD FX8120 8 Cores Motherboard: ASUS M5A78L-M LX V2 With a FX 8120 it is possible to raise easily 3.7 Ghz and the results would be these: Date Message 01/05/2012 11:00:52 Resuming computation 01/05/2012 11:00:51 10818 integer MIPS (Dhrystone) per CPU 01/05/2012 11:00:51 2353 floating point MIPS (Whetstone) per CPU 01/05/2012 11:00:51 Number of CPUs: 8 01/05/2012 11:00:51 Benchmark results: 01/05/2012 11:00:20 Suspending computation - running CPU benchmarks 01/05/2012 11:00:20 Running CPU benchmarks |
|||
|
||||
|
mikey
Veteran Cruncher Joined: May 10, 2009 Post Count: 824 Status: Offline Project Badges:
|
It was on balanced so I have set it to Performance and... 30/04/2012 17:29:05 Running CPU benchmarks 30/04/2012 17:29:06 Suspending computation - running CPU benchmarks 30/04/2012 17:29:36 Benchmark results: 30/04/2012 17:29:36 Number of CPUs: 8 30/04/2012 17:29:36 1620 floating point MIPS (Whetstone) per CPU 30/04/2012 17:29:36 6995 integer MIPS (Dhrystone) per CPU still have my missing mips How much ram does Windows say you have in the machines? [Edit 1 times, last edit by mikey159b at May 1, 2012 1:01:00 PM] |
|||
|
||||
|
|