Index  | Recent Threads  | Unanswered Threads  | Who's Active  | Guidelines  | Search
 

Quick Go »
No member browsing this thread
Thread Status: Active
Total posts in this thread: 15
Posts: 15   Pages: 2   [ Previous Page | 1 2 ]
[ Jump to Last Post ]
Post new Thread
Author
Previous Thread This topic has been viewed 1891 times and has 14 replies Next Thread
Sekerob
Ace Cruncher
Joined: Jul 24, 2005
Post Count: 20043
Status: Offline
Reply to this Post  Reply with Quote 
Re: WUs on AMD quad running very slow

RMAU,

would you please post the start up message log of BOINC so we can see the base configuration. Before doing so also do the Advanced menu, benchmark, so we see that too. Though the benchmark is run at normal priority, whilst the tasks run at idle priority, it's another indicator. Also, go to the data_dir (shown in the start up log), open the client_state.xml and find the <project> section of WCG and the Duration Correction Factor. That's an indicator of real performance versus benchmark. Normal devices have it near the 1.000000 value, but it can swing if job lengths have been incorrectly computed. This happens for projects like HCMD2 and CEP with high duration variation.
----------------------------------------
WCG Global & Research > Make Proposal Help: Start Here!
Please help to make the Forums an enjoyable experience for All!
[May 29, 2009 7:48:00 AM]   Link   Report threatening or abusive post: please login first  Go to top 
twilyth
Master Cruncher
US
Joined: Mar 30, 2007
Post Count: 2130
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: WUs on AMD quad running very slow

Hello, I have run into a similar problem. for some reason the work units are taking longer to do. However I suspected it was processor based. I suspect you are talking about phenoms. If this is true there may be a reason for this slowdown.
"AMD considers the quad core Phenoms to be the first "true" quad core design, as these processors are a monolithic multi-core design"
From my challenges from server motherboards that truly have 2 separate processers I can say this: The more processers you have the more memory each has to fight for. This is one of the reasons of lesser returns from more processers. Is this from the monolithic design? I may never know. Also I run Linux because of a bug in the TLB that is the look-aside buffer. Linux implements a kernel patch to avoid the bug that does not impact performance. However on windows machines, the "patch" makes a 20-50% performance hit. If you are stuck with windows then it is up to you on to disable it or not. From what I have seen it is about as common as if you suddenly won the lottery 10 times. But with these science apps I have no idea. Anyway Phenom gen 1s had some bumps and gen 2 is supposed to be so much better. Just my 2 cents i really have no idea what your problem is. I will say that on my core2 it takes a cretin project about half as long. However this is not true about HCC WUs (maybe due to 64 bit?) Hope this helps a bit.

Please check to see if you have this patch. This seems like the most likely explanation other than a virus or other malware.

However, even in the worst case of a 50% performance decrease, I would expect the quad to be on par with the dual cores, so there may be more than one issue.

I would suggest running Malwarebytes, Avira and/or other virus/malware software. Malwarebytes is free for the scan and you can get Avira for a 30 day trial.

Also, I highly recommend Whatsrunning. This should show you ALL of your processes and what their cpu usage is. Just right click in each tabbed window to get a list of all the parameters you can view.

If you bought your motherboard after 1/2008, the odds are that it already had the TLB patch in the bios, so you would probably have to go back to an older bios version - if that is even an option for your motherboard.

I had 2 Phenom 9600's with the TLB bug and neither had a problem with BOINC, but both came with pre/unpatched bios.
----------------------------------------


----------------------------------------
[Edit 1 times, last edit by twilyth at May 29, 2009 11:05:59 AM]
[May 29, 2009 11:02:29 AM]   Link   Report threatening or abusive post: please login first  Go to top 
RMau
Cruncher
Joined: Feb 6, 2008
Post Count: 44
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: WUs on AMD quad running very slow

Sekerob,

Here is the Duration Correction Factor for WCG. A lot larger than 1.000000.

<duration_correction_factor>3.243494</duration_correction_factor>

And, here is the message log.

5/28/2009 10:26:41 AM||Starting BOINC client version 6.2.28 for windows_intelx86
5/28/2009 10:26:41 AM||log flags: task, file_xfer, sched_ops
5/28/2009 10:26:41 AM||Libraries: libcurl/7.19.0 OpenSSL/0.9.8i zlib/1.2.3
5/28/2009 10:26:41 AM||Data directory: C:\ProgramData\BOINC
5/28/2009 10:26:42 AM||Running under account RMau
5/28/2009 10:26:42 AM||Processor: 4 AuthenticAMD AMD Phenom(tm) 9500 Quad-Core Processor [AMD64 Family 16 Model 2 Stepping 2]
5/28/2009 10:26:42 AM||Processor features: fpu tsc pae nx sse sse2 pni 3dnow mmx
5/28/2009 10:26:42 AM||OS: Microsoft Windows Vista: Home Premium x64 Editon, Service Pack 1, (06.00.6001.00)
5/28/2009 10:26:42 AM||Memory: 5.00 GB physical, 10.14 GB virtual
5/28/2009 10:26:42 AM||Disk: 586.47 GB total, 395.74 GB free
5/28/2009 10:26:42 AM||Local time is UTC -4 hours
5/28/2009 10:26:55 AM|lhcathome|URL: http://lhcathome.cern.ch/lhcathome/; Computer ID: 9708476; location: (none); project prefs: default
5/28/2009 10:26:55 AM|World Community Grid|URL: http://www.worldcommunitygrid.org/; Computer ID: 602169; location: (none); project prefs: default
5/28/2009 10:26:55 AM||General prefs: from World Community Grid (last modified 22-May-2009 07:13:49)
5/28/2009 10:26:56 AM||Host location: none
5/28/2009 10:26:56 AM||General prefs: using your defaults
5/28/2009 10:26:56 AM||Reading preferences override file
5/28/2009 10:26:56 AM||Preferences limit memory usage when active to 2558.79MB
5/28/2009 10:26:56 AM||Preferences limit memory usage when idle to 3838.18MB
5/28/2009 10:26:56 AM||Preferences limit disk usage to 18.63GB
5/28/2009 10:28:42 AM|World Community Grid|Restarting task faah6822_ZINC00861607_xmdEq_1HSI_00_0 using faah version 607
5/28/2009 10:28:42 AM|World Community Grid|Restarting task faah6822_ZINC00865972_xmdEq_1HSI_03_0 using faah version 607
5/28/2009 10:28:42 AM|World Community Grid|Restarting task dddt1402f0324_100165_0 using dddt version 607
5/28/2009 10:28:42 AM|World Community Grid|Restarting task HFCC_t1_00735204_TrkB_0002_0 using hfcc version 610
5/28/2009 11:57:08 AM||General prefs: from World Community Grid (last modified 22-May-2009 07:13:49)
5/28/2009 11:57:08 AM||Host location: none
5/28/2009 11:57:08 AM||General prefs: using your defaults
5/28/2009 11:57:08 AM||Reading preferences override file
5/28/2009 11:57:08 AM||Preferences limit memory usage when active to 2558.79MB
5/28/2009 11:57:08 AM||Preferences limit memory usage when idle to 3838.18MB
5/28/2009 11:57:08 AM||Preferences limit disk usage to 18.63GB
5/28/2009 4:54:26 PM|World Community Grid|Computation for task dddt1402f0324_100165_0 finished
5/28/2009 4:54:27 PM|World Community Grid|Starting ml458_00000_14
5/28/2009 4:54:27 PM|World Community Grid|Starting task ml458_00000_14 using hpf2 version 603
5/28/2009 4:54:29 PM|World Community Grid|Started upload of dddt1402f0324_100165_0_0
5/28/2009 4:54:29 PM|World Community Grid|Started upload of dddt1402f0324_100165_0_1
5/28/2009 4:54:35 PM|World Community Grid|Finished upload of dddt1402f0324_100165_0_0
5/28/2009 4:54:35 PM|World Community Grid|Started upload of dddt1402f0324_100165_0_2
5/28/2009 4:54:37 PM|World Community Grid|Finished upload of dddt1402f0324_100165_0_2
5/28/2009 4:54:37 PM|World Community Grid|Started upload of dddt1402f0324_100165_0_3
5/28/2009 4:54:39 PM|World Community Grid|Finished upload of dddt1402f0324_100165_0_3
5/28/2009 4:54:47 PM|World Community Grid|Finished upload of dddt1402f0324_100165_0_1
5/28/2009 4:59:43 PM|World Community Grid|Computation for task faah6822_ZINC00861607_xmdEq_1HSI_00_0 finished
5/28/2009 4:59:43 PM|World Community Grid|Starting HFCC_t1_00734815_TrkB_0000_0
5/28/2009 4:59:43 PM|World Community Grid|Starting task HFCC_t1_00734815_TrkB_0000_0 using hfcc version 610
5/28/2009 4:59:45 PM|World Community Grid|Started upload of faah6822_ZINC00861607_xmdEq_1HSI_00_0_0
5/28/2009 4:59:45 PM|World Community Grid|Started upload of faah6822_ZINC00861607_xmdEq_1HSI_00_0_1
5/28/2009 4:59:48 PM|World Community Grid|Finished upload of faah6822_ZINC00861607_xmdEq_1HSI_00_0_0
5/28/2009 4:59:48 PM|World Community Grid|Started upload of faah6822_ZINC00861607_xmdEq_1HSI_00_0_2
5/28/2009 4:59:49 PM|World Community Grid|Finished upload of faah6822_ZINC00861607_xmdEq_1HSI_00_0_1
5/28/2009 4:59:49 PM|World Community Grid|Started upload of faah6822_ZINC00861607_xmdEq_1HSI_00_0_3
5/28/2009 4:59:50 PM|World Community Grid|Finished upload of faah6822_ZINC00861607_xmdEq_1HSI_00_0_2
5/28/2009 4:59:51 PM|World Community Grid|Finished upload of faah6822_ZINC00861607_xmdEq_1HSI_00_0_3
5/28/2009 6:58:22 PM|World Community Grid|Computation for task faah6822_ZINC00865972_xmdEq_1HSI_03_0 finished
5/28/2009 6:58:22 PM|World Community Grid|Starting HFCC_t1_00736683_TrkB_0000_0
5/28/2009 6:58:23 PM|World Community Grid|Starting task HFCC_t1_00736683_TrkB_0000_0 using hfcc version 610
5/28/2009 6:58:25 PM|World Community Grid|Started upload of faah6822_ZINC00865972_xmdEq_1HSI_03_0_0
5/28/2009 6:58:25 PM|World Community Grid|Started upload of faah6822_ZINC00865972_xmdEq_1HSI_03_0_1
5/28/2009 6:58:27 PM|World Community Grid|Finished upload of faah6822_ZINC00865972_xmdEq_1HSI_03_0_0
5/28/2009 6:58:27 PM|World Community Grid|Started upload of faah6822_ZINC00865972_xmdEq_1HSI_03_0_2
5/28/2009 6:58:28 PM|World Community Grid|Finished upload of faah6822_ZINC00865972_xmdEq_1HSI_03_0_1
5/28/2009 6:58:28 PM|World Community Grid|Started upload of faah6822_ZINC00865972_xmdEq_1HSI_03_0_3
5/28/2009 6:58:30 PM|World Community Grid|Finished upload of faah6822_ZINC00865972_xmdEq_1HSI_03_0_2
5/28/2009 6:58:32 PM|World Community Grid|Finished upload of faah6822_ZINC00865972_xmdEq_1HSI_03_0_3
5/28/2009 7:17:31 PM|World Community Grid|Sending scheduler request: Requested by user. Requesting 0 seconds of work, reporting 3 completed tasks
5/28/2009 7:17:36 PM|World Community Grid|Scheduler request succeeded: got 0 new tasks
5/28/2009 11:05:57 PM|World Community Grid|Computation for task HFCC_t1_00735204_TrkB_0002_0 finished
5/28/2009 11:05:59 PM|World Community Grid|Started upload of HFCC_t1_00735204_TrkB_0002_0_0
5/28/2009 11:05:59 PM|World Community Grid|Started upload of HFCC_t1_00735204_TrkB_0002_0_1
5/28/2009 11:06:05 PM|World Community Grid|Finished upload of HFCC_t1_00735204_TrkB_0002_0_0
5/28/2009 11:06:05 PM|World Community Grid|Started upload of HFCC_t1_00735204_TrkB_0002_0_2
5/28/2009 11:06:06 PM|World Community Grid|Finished upload of HFCC_t1_00735204_TrkB_0002_0_1
5/28/2009 11:06:06 PM|World Community Grid|Started upload of HFCC_t1_00735204_TrkB_0002_0_3
5/28/2009 11:06:07 PM|World Community Grid|Finished upload of HFCC_t1_00735204_TrkB_0002_0_2
5/28/2009 11:06:08 PM|World Community Grid|Finished upload of HFCC_t1_00735204_TrkB_0002_0_3
5/28/2009 11:30:26 PM|World Community Grid|Sending scheduler request: Requested by user. Requesting 0 seconds of work, reporting 1 completed tasks
5/28/2009 11:30:31 PM|World Community Grid|Scheduler request succeeded: got 0 new tasks
5/29/2009 9:10:22 AM|World Community Grid|Computation for task ml458_00000_14 finished
5/29/2009 9:10:24 AM|World Community Grid|Started upload of ml458_00000_14_0
5/29/2009 9:10:29 AM|World Community Grid|Finished upload of ml458_00000_14_0
5/29/2009 1:48:10 PM|World Community Grid|Computation for task HFCC_t1_00734815_TrkB_0000_0 finished
5/29/2009 1:48:12 PM|World Community Grid|Started upload of HFCC_t1_00734815_TrkB_0000_0_0
5/29/2009 1:48:12 PM|World Community Grid|Started upload of HFCC_t1_00734815_TrkB_0000_0_1
5/29/2009 1:48:14 PM|World Community Grid|Finished upload of HFCC_t1_00734815_TrkB_0000_0_0
5/29/2009 1:48:14 PM|World Community Grid|Started upload of HFCC_t1_00734815_TrkB_0000_0_2
5/29/2009 1:48:15 PM|World Community Grid|Finished upload of HFCC_t1_00734815_TrkB_0000_0_1
5/29/2009 1:48:15 PM|World Community Grid|Started upload of HFCC_t1_00734815_TrkB_0000_0_3
5/29/2009 1:48:16 PM|World Community Grid|Finished upload of HFCC_t1_00734815_TrkB_0000_0_2
5/29/2009 1:48:17 PM|World Community Grid|Finished upload of HFCC_t1_00734815_TrkB_0000_0_3
5/29/2009 4:41:35 PM||Running CPU benchmarks
5/29/2009 4:41:36 PM||Suspending computation - running CPU benchmarks
5/29/2009 4:42:07 PM||Benchmark results:
5/29/2009 4:42:07 PM|| Number of CPUs: 4
5/29/2009 4:42:07 PM|| 2176 floating point MIPS (Whetstone) per CPU
5/29/2009 4:42:07 PM|| 4805 integer MIPS (Dhrystone) per CPU
5/29/2009 4:42:08 PM||Resuming computation

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

----------------------------------------
[Edit 1 times, last edit by RMau at May 30, 2009 12:29:47 PM]
[May 29, 2009 8:52:12 PM]   Link   Report threatening or abusive post: please login first  Go to top 
RMau
Cruncher
Joined: Feb 6, 2008
Post Count: 44
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: WUs on AMD quad running very slow

twilyth,

Thank you for the suggestion. I'll look in to this patch and your suggested trouble shooting programs. I've used similar ones that have not found anything yet, but one program won't catch everything.
----------------------------------------

[May 29, 2009 8:54:42 PM]   Link   Report threatening or abusive post: please login first  Go to top 
twilyth
Master Cruncher
US
Joined: Mar 30, 2007
Post Count: 2130
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: WUs on AMD quad running very slow

I'm not doing faah or dddt right now, but my times on a Phenom 9600 (slightly over stock) for HFCC are from 4.74 hrs to 8.31 hrs with most falling in the 6.5-7.0 range. So from the one job in your listing that I can compare, you seem to be right on par.

also, just a helpful hint, when you have a lot of text you want to post but don't want it to take up most of a page, use the code /code BBC tags and you will get a scrolling text box - like a quote but with a scroll bar. What you posted wasn't really long, but if you ever need to post something that is, that's a good way to do it without having to worry about trying to edit it.
----------------------------------------


----------------------------------------
[Edit 1 times, last edit by twilyth at May 30, 2009 1:15:15 AM]
[May 30, 2009 1:14:08 AM]   Link   Report threatening or abusive post: please login first  Go to top 
Posts: 15   Pages: 2   [ Previous Page | 1 2 ]
[ Jump to Last Post ]
Post new Thread