Low points

eOn code for long time scale dynamics

Moderator: moderators

chill
Posts: 96
Joined: Tue Jul 28, 2009 9:04 pm

Re: Low points

Post by chill »

modesti wrote:[Prepared on Fri, 29th July - waiting for registration on the forum]

Hello,

this morning I found several WUs of V3.09 on my i7 860 (Win7 64b) that were running for nearly 5 hours. The least problem is that – due to the short deadline – they were/are all in high priority.

The bigger problem I found were the granted credits.
http://eon.ices.utexas.edu/eon2/results ... =0&state=3

Code: Select all

103782879 	112944808 	29 Jul 2011 0:59:07 UTC 	29 Jul 2011 5:56:48 UTC 	Completed and validated 	17,846.34 	16,896.33 	10.17 	eOn Client v3.09
103782828 	112944757 	29 Jul 2011 0:57:59 UTC 	29 Jul 2011 6:42:03 UTC 	Completed and validated 	20,629.32 	19,779.48 	11.48 	eOn Client v3.09
103781280 	112943209 	28 Jul 2011 22:21:38 UTC 	29 Jul 2011 4:06:22 UTC 	Completed and validated 	20,322.76 	19,161.26 	6.80 	eOn Client v3.09
103781011 	112942940 	28 Jul 2011 22:27:32 UTC 	29 Jul 2011 4:22:58 UTC 	Completed and validated 	21,311.43 	20,113.44 	7.55 	eOn Client v3.09
103780733 	112942662 	28 Jul 2011 21:47:55 UTC 	29 Jul 2011 2:49:15 UTC 	Completed and validated 	18,065.38 	16,848.31 	4.36 	eOn Client v3.09
103780566 	112942495 	28 Jul 2011 21:44:37 UTC 	29 Jul 2011 4:48:49 UTC 	Completed and validated 	25,249.45 	24,074.46 	8.22 	eOn Client v3.09
103776483 	112938412 	28 Jul 2011 19:22:54 UTC 	28 Jul 2011 22:22:42 UTC 	Completed and validated 	10,713.89 	9,632.20 	25.04 	eOn Client v3.09
103776389 	112938318 	28 Jul 2011 19:18:53 UTC 	28 Jul 2011 22:28:37 UTC 	Completed and validated 	11,079.95 	9,968.34 	26.81 	eOn Client v3.09
103776127 	112938056 	28 Jul 2011 18:56:41 UTC 	28 Jul 2011 21:49:00 UTC 	Completed and validated 	10,263.77 	9,908.15 	22.45 	eOn Client v3.09
103776122 	112938051 	28 Jul 2011 18:56:41 UTC 	28 Jul 2011 21:45:44 UTC 	Completed and validated 	10,066.81 	8,940.96 	21.27 	eOn Client v3.09
103772938 	112934867 	28 Jul 2011 16:22:56 UTC 	28 Jul 2011 19:23:59 UTC 	Completed and validated 	10,786.43 	9,777.10 	23.42 	eOn Client v3.09
103772907 	112934836 	28 Jul 2011 16:17:11 UTC 	28 Jul 2011 19:20:00 UTC 	Completed and validated 	10,557.59 	9,562.10 	22.06 	eOn Client v3.09
103769507 	112931436 	28 Jul 2011 13:18:17 UTC 	28 Jul 2011 16:24:02 UTC 	Completed and validated 	3,398.22 	3,006.67 	7.36 	eOn Client v3.09
103768542 	112930471 	28 Jul 2011 12:35:45 UTC 	28 Jul 2011 16:18:18 UTC 	Completed and validated 	11,949.59 	10,887.73 	24.71 	eOn Client v3.09
103765723 	112927652 	28 Jul 2011 9:33:43 UTC 	28 Jul 2011 11:19:47 UTC 	Completed and validated 	3,191.34 	3,042.11 	9.19 	eOn Client v3.09
103765721 	112927650 	28 Jul 2011 9:33:43 UTC 	28 Jul 2011 10:43:35 UTC 	Completed and validated 	2,938.59 	2,783.03 	8.32 	eOn Client v3.09
103765695 	112927624 	28 Jul 2011 9:32:23 UTC 	28 Jul 2011 13:18:17 UTC 	Completed and validated 	9,393.52 	8,843.18 	20.91 	eOn Client v3.09
103765694 	112927623 	28 Jul 2011 9:32:23 UTC 	28 Jul 2011 12:35:45 UTC 	Completed and validated 	9,389.88 	8,950.84 	26.71 	eOn Client v3.09
103765665 	112927594 	28 Jul 2011 9:31:10 UTC 	28 Jul 2011 10:20:04 UTC 	Completed and validated 	2,918.15 	2,626.90 	8.20 	eOn Client v3.09
103765657 	112927586 	28 Jul 2011 9:31:10 UTC 	28 Jul 2011 11:50:55 UTC 	Completed and validated 	8,340.40 	7,848.19 	23.73 	eOn Client v3.09 
If you compare the last line with the first one, you get 10.855 cr/h vs. 2.167 cr/h.
Usually, I don't care very much about credits, but I won't let my computer run for as less as 2 cr/h. Even Biology projects pay better.
Maybe you could look into the matter?


added: The last WUs ran over 6.5 hours and paid about 2.26-2.27 cr/h :-/
I am currently in touch with the BOINC developers about our credit as compared to other projects. It is unfortunate that the credit calculation is so complicated. I really have little to do, as far as I know, with how much credit our users get for their effort. I hope that we can bring our project on par with the credit "norm". Thanks for volunteering!
Tex1954
Posts: 24
Joined: Fri May 27, 2011 9:47 pm

Re: Low points

Post by Tex1954 »

by chill » Sat Jul 30, 2011 2:19 am
I am currently in touch with the BOINC developers about our credit as compared to other projects. It is unfortunate that the credit calculation is so complicated. I really have little to do, as far as I know, with how much credit our users get for their effort. I hope that we can bring our project on par with the credit "norm". Thanks for volunteering!
Seems to me you are missing something here... 3.08 ran fine! Credits for 3.08 we even and scaled to the task duration as shown in the following spreadsheet entries I keep.

Look at the values I was getting before it started erroring out and I finally deleted the tasks!!! Notice the PPS on 3.08 vs. 3.09!

http://i.imgur.com/hH7KA.jpg

Clearly something is greatly changed in the clients...

And look at the latest values for points all over the place:

103783550 112945479 16,695.67 16,585.23 12.76 eOn Client v3.09
103783279 112945208 15,043.16 14,587.26 24.91 eOn Client v3.09
103783223 112945152 13,782.73 13,355.73 24.13 eOn Client v3.09
103774323 112936252 7,605.32 5,590.39 69.19 eOn Client v3.09
103774908 112936837 3,360.93 1,806.18 15.24 eOn Client v3.09
103783550 112945479 16,695.67 16,585.23 12.76 eOn Client v3.09
103783549 112945478 17,039.58 16,785.03 30.72 eOn Client v3.09

One might think a "minor" looking revision change would mean a minor impact (if any) on the tasks, but that isn't the case.

Someone put together these tasks and it seems they are not tolerant of being suspended and restarted and the WU's I received would error out 50% of the time.

So, whomever developed these tasks did much more than a "minor" change. They totally screwed up the points given and the runtime properties. I don't really care about the overall "low" credit as long as it is the same for all, but as it is, it's totally unpredictable! Points from 12 to 69 with no sense of what's what. It's nice to have tasks that have a narrower range of runtime deviation so one can predict and plan other project tasks besides yours.

Someone scewed up... please fix the problem or give an in depth technical explanation of why one WU gets 69 points and another MUCH longer WU gets 12 points.

As of now, I won't run this on my main system because it is intolerant of start/stop operation. It runs fine on my other systems that do only BOINC tasks, but not on my main system where I stop and start BOINC 3 or 4 times a day.

Until these HUGE LONG work units get evened out points wise, ya'll are not paying "fairly" across the board to all participants! One person or task gets tons more points than another that runs longer on a faster system... that is just not right. I've halted all EON work until I see something done about this.

Tex1954
modesti
Posts: 3
Joined: Fri Jul 29, 2011 7:34 am

Re: Low points

Post by modesti »

chill wrote: I am currently in touch with the BOINC developers about our credit as compared to other projects. It is unfortunate that the credit calculation is so complicated. I really have little to do, as far as I know, with how much credit our users get for their effort. I hope that we can bring our project on par with the credit "norm". Thanks for volunteering!
Thank you for your quick reply and action.
The new batch seems to have come back to "normal" runtime (about 47-50 min on my i7) and the credits are.... how shall I put that? ... a "little high"? Even compared to before the extra-long WUs. Is it meant as compensation for the credit loss? ;-) Well, thank you - I surely won't complain about that :-))
Saenger
Posts: 29
Joined: Thu Sep 02, 2010 4:23 pm
Contact:

Re: Low points

Post by Saenger »

I definitely can't confirm that credits per time get down, it's the other way around for my machine. I get more credits with newer versions (exception was the v3.04, that was excessive).

Perhaps it's just that the newer apps are not that effective on windows machines any more, and thus don't deserve that much credits?
Tex1954
Posts: 24
Joined: Fri May 27, 2011 9:47 pm

Re: Low points

Post by Tex1954 »

Saenger wrote:I definitely can't confirm that credits per time get down, it's the other way around for my machine. I get more credits with newer versions (exception was the v3.04, that was excessive).

Perhaps it's just that the newer apps are not that effective on windows machines any more, and thus don't deserve that much credits?
The tasks I posted are available for anyone to see... points were all over the place... as were times and errors.

I read modesti's post today and decided to give it a try again. So far, it's working well now. Seems the WU's are fairly even now at around 1.75 hrs runtime, no errors yet this morning... points more even and improved, but still all over the place... still, they run fine now without error...

If it continues to run this way, then no complaint from me. But, for a while there, I was pulling my hair out! LOL!

8-)

Tex1954
modesti
Posts: 3
Joined: Fri Jul 29, 2011 7:34 am

Re: Low points

Post by modesti »

After having posted yesterday, I noticed that the "duration correction factor" was over 43 on my i7. I reset the project last night and it's running fine now.
Then I attached a new computer last night and this morning its duration correction factor had increased to 33. Can't reset for the moment because there's a WU running.
Just checked on the i7 again, and it is 2.2944.
Strange that this duration correction factor changes so quickly on this project...

I still don't know exactly how it works, but from all I've read on different forums I know it has /can have an influence on the granted credits.
Nflight
Posts: 8
Joined: Fri Oct 22, 2010 5:10 pm
Contact:

Re: Low points

Post by Nflight »

Still seeing very little points values. This is two days of nearly 2 to 3 credits per hour, a change is needed to bring the masses back to crunch this project. As project admin you repeatedly exclaim you can't see why people are so distressed about points values. The other side of the equation does this because as nerds our realm of dominance in achieving a level of greatness which comes from who obtains the highest point values. Us nerds are not really excited about the results of each project but, the actual point values we can obtain to rise above our fellow nerds. I personally participate in over 90 projects all to help my team and myself to rise in stardom. EON is a nice project or at least before you came to BOINC it surely was an enjoyable project to run. You are going through growing pains but points values really do need to change in the upward rise to keep all those hard working nerds ( Sorry if I offend anyone with that wording) crunching your data.

Hope this Helps
GaryWilson
Posts: 12
Joined: Fri Dec 03, 2010 2:13 am

Re: Low points

Post by GaryWilson »

Well, I suspect you may realize by now, based on the # of computers reported on your project home page, that crunchers would appear to be losing interest in your project due to the low points as there used to be over 800 computers at one point. I have a couple machines that seem to be doing "ok", but I use that loosely. I have removed my other computers as they were generating less than a point per credit hour. I am closely watching the remaining two but will probably remove them at some point as for some reason, the longer I crunch on a new project executable, the lower the cr/hour gets.

I know this is less of your issue as this problem seems to constantly comes up on other projects as well. But it seems much worse here than on others for some reason. Hopefully you will find a resolution to it and get your crunchers back.
Tex1954
Posts: 24
Joined: Fri May 27, 2011 9:47 pm

Re: Low points

Post by Tex1954 »

Opened up the tasks again yesterday and finally got some good results.

http://eon.ices.utexas.edu/eon2/results ... =0&state=3

http://i.imgur.com/AgJdC.jpg

Looks like the tasks are scaled properly again, none of the wildness showing...

Okay, I'm happy now so long as it stays this way.

Still, would be nice if somebody "in the know" would talk to us... I know most are super busy... so as long as it stays this way, I'll play.

:)

Tex1954
GaryWilson
Posts: 12
Joined: Fri Dec 03, 2010 2:13 am

Re: Low points

Post by GaryWilson »

Actually, what I have observed is that sometimes if you detach and then reattach, the points get better or if you stop working on the project a bit and then start again, they are OK. But return a few days worth of results and suddenly the points drop in half. Really strange.
upstatelabs
Posts: 20
Joined: Wed Oct 27, 2010 2:07 pm

Re: Low points

Post by upstatelabs »

I am now getting really low points for the work my computers are doing on this project, compared to say, SETI, or WCG. By a large percentage. Perhaps something can be done to fix the issue. It's an interesting project you have here, I'd like to keep participating....
Saenger
Posts: 29
Joined: Thu Sep 02, 2010 4:23 pm
Contact:

Re: Low points

Post by Saenger »

upstatelabs wrote:I am now getting really low points for the work my computers are doing on this project, compared to say, SETI, or WCG. By a large percentage. Perhaps something can be done to fix the issue. It's an interesting project you have here, I'd like to keep participating....
Can you please link your computers or state your system setup here?
I get about 60% better credits as with WCG, I don't do Seti, so I can't say anything here. WCG gives me ~25 C/h, eOn gives ~40 C/h
Here's my WUs and my computer.
Tex1954
Posts: 24
Joined: Fri May 27, 2011 9:47 pm

Windows-7 vs. Linux on same machines

Post by Tex1954 »

chill wrote:
Stef42 wrote:Still, the points difference between Windows and Linux is way too huge, just for the same runtime.
I'd be interested in seeing the difference in the benchmark on windows and linux for the same computer. I remember reading the the BOINC client benchmark isn't as optimized on linux as it is on windows.
Okay, here it is off two different machines for your pleasure...
----------------------------------------------------------------------------------------------------------------------------------------------------

Linux vs. Windows 7 Benchmarks: By Tex1954 8/13/2011

Two machines, same Asus M4N75TD installed in both machines.
Both machines have Corsair F90 SSD's as master boot drives, split partition 60/40 Windows7/Linux.
Ran benchmark test three times in a row to copy/past under the startup info.


**********************************************************************************
1090T machine: 3.75GHz AMD-1090T, 8Gig DDR3, Win7/Linux dual boot.
**********************************************************************************
Linux-1090T: Ubuntu 11.04-64b, BOINC 6.10.59

1 8/13/2011 9:04:55 PM Starting BOINC client version 6.10.59 for x86_64-pc-linux-gnu
2 8/13/2011 9:04:55 PM Config: GUI RPC allowed from:
3 8/13/2011 9:04:55 PM Config: 192.168.1.100
4 8/13/2011 9:04:55 PM log flags: file_xfer, sched_ops, task
5 8/13/2011 9:04:55 PM Libraries: libcurl/7.21.3 OpenSSL/0.9.8o zlib/1.2.3.4 libidn/1.18
6 8/13/2011 9:04:55 PM Data directory: /var/lib/boinc-client
7 8/13/2011 9:04:55 PM Processor: 6 AuthenticAMD AMD Phenom(tm) II X6 1090T Processor [Family 16 Model 10 Stepping 0]
8 8/13/2011 9:04:55 PM Processor: 512.00 KB cache
9 8/13/2011 9:04:55 PM Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm 3dnowext 3dnow constant_tsc rep_good nopl nonstop_tsc extd_apicid aperfmperf pni monitor
10 8/13/2011 9:04:55 PM OS: Linux: 2.6.38-8-generic
11 8/13/2011 9:04:55 PM Memory: 7.81 GB physical, 255.99 MB virtual
12 8/13/2011 9:04:55 PM Disk: 25.71 GB total, 21.49 GB free
13 8/13/2011 9:04:55 PM Local time is UTC -5 hours
14 8/13/2011 9:04:55 PM NVIDIA GPU 0: GeForce GTX 460 (driver version unknown, CUDA version 4000, compute capability 2.1, 1024MB, 605 GFLOPS peak)
15 8/13/2011 9:04:55 PM NVIDIA GPU 1: GeForce GTX 460 (driver version unknown, CUDA version 4000, compute capability 2.1, 1023MB, 605 GFLOPS peak)
...
59 8/13/2011 9:19:57 PM Running CPU benchmarks
60 8/13/2011 9:20:29 PM Benchmark results:
61 8/13/2011 9:20:29 PM Number of CPUs: 6
62 8/13/2011 9:20:29 PM 3530 floating point MIPS (Whetstone) per CPU
63 8/13/2011 9:20:29 PM 17709 integer MIPS (Dhrystone) per CPU
64 8/13/2011 9:21:30 PM Running CPU benchmarks
65 8/13/2011 9:22:02 PM Benchmark results:
66 8/13/2011 9:22:02 PM Number of CPUs: 6
67 8/13/2011 9:22:02 PM 3536 floating point MIPS (Whetstone) per CPU
68 8/13/2011 9:22:02 PM 17019 integer MIPS (Dhrystone) per CPU
69 8/13/2011 9:22:23 PM Running CPU benchmarks
70 8/13/2011 9:22:54 PM Benchmark results:
71 8/13/2011 9:22:54 PM Number of CPUs: 6
72 8/13/2011 9:22:54 PM 3552 floating point MIPS (Whetstone) per CPU
73 8/13/2011 9:22:54 PM 17146 integer MIPS (Dhrystone) per CPU
74 8/13/2011 9:23:02 PM Running CPU benchmarks
75 8/13/2011 9:23:33 PM Benchmark results:
76 8/13/2011 9:23:33 PM Number of CPUs: 6
77 8/13/2011 9:23:33 PM 3529 floating point MIPS (Whetstone) per CPU
78 8/13/2011 9:23:33 PM 17085 integer MIPS (Dhrystone) per CPU


Phenom-1090T Windows 7-64b, BOINC 6.12.33

1 8/13/2011 9:25:45 PM Starting BOINC client version 6.12.33 for windows_x86_64
2 8/13/2011 9:25:45 PM Config: GUI RPC allowed from:
3 8/13/2011 9:25:45 PM Config: 192.168.1.100
4 8/13/2011 9:25:45 PM log flags: file_xfer, sched_ops, task
5 8/13/2011 9:25:45 PM Libraries: libcurl/7.19.7 OpenSSL/0.9.8l zlib/1.2.5
6 8/13/2011 9:25:45 PM Data directory: D:\BOINC Data
7 8/13/2011 9:25:45 PM Running under account Owner
8 8/13/2011 9:25:45 PM Processor: 6 AuthenticAMD AMD Phenom(tm) II X6 1090T Processor [Family 16 Model 10 Stepping 0]
9 8/13/2011 9:25:45 PM Processor: 512.00 KB cache
10 8/13/2011 9:25:45 PM 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 cx16 syscall nx lm svm sse4a osvw ibs skinit wdt page1gb rdtscp 3dnowext 3dnow
11 8/13/2011 9:25:45 PM OS: Microsoft Windows 7: Professional x64 Edition, Service Pack 1, (06.01.7601.00)
12 8/13/2011 9:25:45 PM Memory: 8.00 GB physical, 16.00 GB virtual
13 8/13/2011 9:25:45 PM Disk: 244.13 GB total, 231.98 GB free
14 8/13/2011 9:25:45 PM Local time is UTC -5 hours
15 8/13/2011 9:25:45 PM NVIDIA GPU 0: GeForce GTX 460 (driver version 26658, CUDA version 3020, compute capability 2.1, 993MB, 605 GFLOPS peak)
16 8/13/2011 9:25:45 PM NVIDIA GPU 1: GeForce GTX 460 (driver version 26658, CUDA version 3020, compute capability 2.1, 994MB, 605 GFLOPS peak)
...
82 8/13/2011 9:29:23 PM Benchmark results:
83 8/13/2011 9:29:23 PM Number of CPUs: 6
84 8/13/2011 9:29:23 PM 3083 floating point MIPS (Whetstone) per CPU
85 8/13/2011 9:29:23 PM 9409 integer MIPS (Dhrystone) per CPU
86 8/13/2011 9:29:50 PM Running CPU benchmarks
87 8/13/2011 9:30:22 PM Benchmark results:
88 8/13/2011 9:30:22 PM Number of CPUs: 6
89 8/13/2011 9:30:22 PM 3083 floating point MIPS (Whetstone) per CPU
90 8/13/2011 9:30:22 PM 9481 integer MIPS (Dhrystone) per CPU
91 8/13/2011 9:30:29 PM Running CPU benchmarks
92 8/13/2011 9:31:00 PM Benchmark results:
93 8/13/2011 9:31:00 PM Number of CPUs: 6
94 8/13/2011 9:31:00 PM 3083 floating point MIPS (Whetstone) per CPU
95 8/13/2011 9:31:00 PM 9417 integer MIPS (Dhrystone) per CPU



**********************************************************************************
955 machine: 3.93GHz AMD-955BE, 8Gig DDR3, Win7/Linux dual boot.
**********************************************************************************
Linux-955: Ubuntu 11.04-64b, BOINC 6.10.59

1 8/13/2011 9:46:16 PM Starting BOINC client version 6.10.59 for x86_64-pc-linux-gnu
2 8/13/2011 9:46:16 PM Config: GUI RPC allowed from:
3 8/13/2011 9:46:16 PM Config: 192.168.1.100
4 8/13/2011 9:46:16 PM Config: 192.168.1.102
5 8/13/2011 9:46:16 PM log flags: file_xfer, sched_ops, task
6 8/13/2011 9:46:16 PM Libraries: libcurl/7.21.3 OpenSSL/0.9.8o zlib/1.2.3.4 libidn/1.18
7 8/13/2011 9:46:16 PM Data directory: /var/lib/boinc-client
8 8/13/2011 9:46:16 PM Processor: 4 AuthenticAMD AMD Phenom(tm) II X4 955 Processor [Family 16 Model 4 Stepping 3]
9 8/13/2011 9:46:16 PM Processor: 512.00 KB cache
10 8/13/2011 9:46:16 PM Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm 3dnowext 3dnow constant_tsc rep_good nopl nonstop_tsc extd_apicid pni monitor cx16 popcnt
11 8/13/2011 9:46:16 PM OS: Linux: 2.6.38-10-generic
12 8/13/2011 9:46:16 PM Memory: 7.82 GB physical, 255.99 MB virtual
13 8/13/2011 9:46:16 PM Disk: 18.98 GB total, 14.57 GB free
14 8/13/2011 9:46:16 PM Local time is UTC -5 hours
15 8/13/2011 9:46:16 PM NVIDIA GPU 0: GeForce 9800 GT (driver version unknown, CUDA version 4000, compute capability 1.1, 512MB, 336 GFLOPS peak)
16 8/13/2011 9:46:16 PM NVIDIA GPU 1: GeForce 9800 GT (driver version unknown, CUDA version 4000, compute capability 1.1, 511MB, 336 GFLOPS peak)
...
37 8/13/2011 9:47:42 PM Running CPU benchmarks
38 8/13/2011 9:48:14 PM Benchmark results:
39 8/13/2011 9:48:14 PM Number of CPUs: 4
40 8/13/2011 9:48:14 PM 3724 floating point MIPS (Whetstone) per CPU
41 8/13/2011 9:48:14 PM 17628 integer MIPS (Dhrystone) per CPU
42 8/13/2011 9:48:22 PM Running CPU benchmarks
43 8/13/2011 9:48:53 PM Benchmark results:
44 8/13/2011 9:48:53 PM Number of CPUs: 4
45 8/13/2011 9:48:53 PM 3711 floating point MIPS (Whetstone) per CPU
46 8/13/2011 9:48:53 PM 17888 integer MIPS (Dhrystone) per CPU
47 8/13/2011 9:49:05 PM Running CPU benchmarks
48 8/13/2011 9:49:36 PM Benchmark results:
49 8/13/2011 9:49:36 PM Number of CPUs: 4
50 8/13/2011 9:49:36 PM 3725 floating point MIPS (Whetstone) per CPU
51 8/13/2011 9:49:36 PM 17696 integer MIPS (Dhrystone) per CPU


Phenom-955: Windows 7-64b, BOINC 6.12.33

1 8/13/2011 9:37:46 PM Starting BOINC client version 6.12.33 for windows_x86_64
2 8/13/2011 9:37:46 PM Config: GUI RPC allowed from:
3 8/13/2011 9:37:46 PM Config: 192.168.1.100
4 8/13/2011 9:37:46 PM Config: 192.168.1.101
5 8/13/2011 9:37:46 PM Config: 192.168.1.102
6 8/13/2011 9:37:46 PM log flags: file_xfer, sched_ops, task
7 8/13/2011 9:37:46 PM Libraries: libcurl/7.19.7 OpenSSL/0.9.8l zlib/1.2.5
8 8/13/2011 9:37:46 PM Data directory: D:\BOINC Data
9 8/13/2011 9:37:46 PM Running under account Owner
10 8/13/2011 9:37:46 PM Processor: 4 AuthenticAMD AMD Phenom(tm) II X4 955 Processor [Family 16 Model 4 Stepping 3]
11 8/13/2011 9:37:46 PM Processor: 512.00 KB cache
12 8/13/2011 9:37:46 PM 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 cx16 syscall nx lm svm sse4a osvw ibs skinit wdt page1gb rdtscp 3dnowext 3dnow
13 8/13/2011 9:37:46 PM OS: Microsoft Windows 7: Professional x64 Edition, Service Pack 1, (06.01.7601.00)
14 8/13/2011 9:37:46 PM Memory: 8.00 GB physical, 16.00 GB virtual
15 8/13/2011 9:37:46 PM Disk: 241.45 GB total, 230.86 GB free
16 8/13/2011 9:37:46 PM Local time is UTC -5 hours
17 8/13/2011 9:37:46 PM NVIDIA GPU 0: GeForce 9800 GT (driver version 26658, CUDA version 3020, compute capability 1.1, 488MB, 336 GFLOPS peak)
18 8/13/2011 9:37:46 PM NVIDIA GPU 1: GeForce 9800 GT (driver version 26658, CUDA version 3020, compute capability 1.1, 488MB, 336 GFLOPS peak)
...
54 8/13/2011 9:38:48 PM Running CPU benchmarks
55 8/13/2011 9:39:19 PM Benchmark results:
56 8/13/2011 9:39:19 PM Number of CPUs: 4
57 8/13/2011 9:39:19 PM 3223 floating point MIPS (Whetstone) per CPU
58 8/13/2011 9:39:19 PM 10091 integer MIPS (Dhrystone) per CPU
59 8/13/2011 9:40:40 PM Running CPU benchmarks
60 8/13/2011 9:41:12 PM Benchmark results:
61 8/13/2011 9:41:12 PM Number of CPUs: 4
62 8/13/2011 9:41:12 PM 3221 floating point MIPS (Whetstone) per CPU
63 8/13/2011 9:41:12 PM 10150 integer MIPS (Dhrystone) per CPU
64 8/13/2011 9:41:23 PM Running CPU benchmarks
65 8/13/2011 9:41:55 PM Benchmark results:
66 8/13/2011 9:41:55 PM Number of CPUs: 4
67 8/13/2011 9:41:55 PM 3227 floating point MIPS (Whetstone) per CPU
68 8/13/2011 9:41:55 PM 10164 integer MIPS (Dhrystone) per CPU
**********************************************************************************

I would say that was a pretty big difference...

Tex1954
upstatelabs
Posts: 20
Joined: Wed Oct 27, 2010 2:07 pm

Re: Low points

Post by upstatelabs »

Saenger wrote:
upstatelabs wrote:I am now getting really low points for the work my computers are doing on this project, compared to say, SETI, or WCG. By a large percentage. Perhaps something can be done to fix the issue. It's an interesting project you have here, I'd like to keep participating....
Can you please link your computers or state your system setup here?
I get about 60% better credits as with WCG, I don't do Seti, so I can't say anything here. WCG gives me ~25 C/h, eOn gives ~40 C/h
Here's my WUs and my computer.
The credit thing here has gotten way out of whack. This WU gave me 5.38 credits for 3.89 hrs runtime on a p4 3.2gHz computer running winxp. That works out to about 1.38 credits/ hr. That computer''s RAC is 49 and falling. As a comparison, I have a 3.0GHz P4 running with very similar conditions and usage, and is running WCG with a RAC of 221.

And here's the thing, it's not just one computer, it's all of them (well, all of mine anyway).
Well, it'll be what's left after I reassign elsewhere, because this is ridiculous.
Saenger
Posts: 29
Joined: Thu Sep 02, 2010 4:23 pm
Contact:

Re: Low points

Post by Saenger »

If I look at this graph, it seems the current app is not good suited for Win-computers:
http://wuprop.boinc-af.org/results/grap ... l&cpuid=17

It's my Linux-machine against a similar Windows one.
Post Reply