The cache & downloading WUs?

Message boards : Number crunching : The cache & downloading WUs?
Message board moderation

To post messages, you must log in.

AuthorMessage
Profile John Clark
Volunteer tester
Avatar

Send message
Joined: 29 Sep 99
Posts: 16515
Credit: 4,418,829
RAC: 0
United Kingdom
Message 715225 - Posted: 19 Feb 2008, 10:49:05 UTC
Last modified: 19 Feb 2008, 10:51:04 UTC

I have a C2Q which recently seems to have got a mind of it's own, and my fiddling has not cured it.

I am currently running the recommended BOINC Manager (5.10.30). Before this (yesterday) I was running 5.10.20, and in both cases it crunched WUs OK.

Where my problem comes is for the size of the cache of WUs ready to start.

A week ago it was running with 5 days cache, which were crunched but not replaced. This went on and the rig ran as if there was no cache (what it was crunching plus a spare WU for each core).

I tried upping the cache to 8 days, and I get this in the BOINC Manager messages tab -

19/02/2008 10:35:59|SETI@home|Computation for task 29no06aa.5993.19295.10.7.201_0 finished
19/02/2008 10:35:59|SETI@home|Starting 29no06aa.5993.19295.10.7.163_0
19/02/2008 10:35:59|SETI@home|Starting task 29no06aa.5993.19295.10.7.163_0 using setiathome_enhanced version 527
19/02/2008 10:36:01|SETI@home|Started upload of 29no06aa.5993.19295.10.7.201_0_0
19/02/2008 10:36:04|SETI@home|Finished upload of 29no06aa.5993.19295.10.7.201_0_0
19/02/2008 10:36:15|SETI@home|Sending scheduler request: Requested by user. Requesting 0 seconds of work, reporting 1 completed tasks
19/02/2008 10:36:20|SETI@home|Scheduler request succeeded: got 0 new tasks
19/02/2008 10:36:30|SETI@home|Sending scheduler request: Requested by user. Requesting 0 seconds of work, reporting 0 completed tasks
19/02/2008 10:36:35|SETI@home|Scheduler request succeeded: got 0 new tasks


Yesterday, when the WUs were nearly cleared I detached from SETI. Cleared all data from the BOINC folder and the projects folder and installed 5.10.30, then reattached.

The system downloaded about 150 WUs OK. But since the number of cached WUs is dropping and no new WUs are being downloaded, as per Messages tab.

Is there anything I can do to return the rig to normal, and working with a reasonable cache like the rest of my rigs?
It's good to be back amongst friends and colleagues



ID: 715225 · Report as offensive
W-K 666 Project Donor
Volunteer tester

Send message
Joined: 18 May 99
Posts: 19078
Credit: 40,757,560
RAC: 67
United Kingdom
Message 715230 - Posted: 19 Feb 2008, 11:09:49 UTC

Had similar problem on one of mine recently, the clock had gone forward a couple of months for some unknown reason. But the net result was that the line;

While BOINC running, % of time work is allowed 99.9922 %

had reduced to about 23.nnnn%

The line reporting this is in client_state.xml file on line;
<cpu_efficiency>0.999922</cpu_efficiency>

If this is case stop BOINC, edit to reasonable figure, in format shown, and restart BOINC.
ID: 715230 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14653
Credit: 200,643,578
RAC: 874
United Kingdom
Message 715232 - Posted: 19 Feb 2008, 11:15:09 UTC - in response to Message 715230.  

Had similar problem on one of mine recently, the clock had gone forward a couple of months for some unknown reason. But the net result was that the line;

While BOINC running, % of time work is allowed 99.9922 %

had reduced to about 23.nnnn%

The line reporting this is in client_state.xml file on line;
<cpu_efficiency>0.999922</cpu_efficiency>

If this is case stop BOINC, edit to reasonable figure, in format shown, and restart BOINC.

I think that particular error would be in

<active_frac>0.999938</active_frac>

but the principle is good - check all the metrics in <time_stats>

While you're in there, perhaps also check

<duration_correction_factor>0.276953</duration_correction_factor>

(there's a separate one for each project). Anything more than 0.500000 for SETI is suspect - with an optimised application, it'll probably be below 0.300000, as here.
ID: 715232 · Report as offensive
Profile John Clark
Volunteer tester
Avatar

Send message
Joined: 29 Sep 99
Posts: 16515
Credit: 4,418,829
RAC: 0
United Kingdom
Message 715259 - Posted: 19 Feb 2008, 12:45:28 UTC
Last modified: 19 Feb 2008, 12:46:22 UTC

Thanks

I need to go out for a couple of hours, but will look over your suggested changes.

I hope they give the clues needed, and I can get the cache working normally. That way I can cover the Planned Outrage at 6.00 pm UTC tonight, and it will help reduce my growing pending credit pile (now nearly 20,000 cobblestones).
It's good to be back amongst friends and colleagues



ID: 715259 · Report as offensive
Profile JDWhale
Volunteer tester
Avatar

Send message
Joined: 6 Apr 99
Posts: 921
Credit: 21,935,817
RAC: 3
United States
Message 715264 - Posted: 19 Feb 2008, 12:55:40 UTC - in response to Message 715225.  


I tried upping the cache to 8 days,



So many ways to control the cache, so many different behaviors.

I've gone with setting "local preferences" to:
    Connect about every 0.0 days  Additional work buffer 5.0 days 

Setting connect interval to 5 days causes download havoc with the flood of "shorties" coming down the pipe due to the short deadlines.

YMMV
ID: 715264 · Report as offensive
Profile John Clark
Volunteer tester
Avatar

Send message
Joined: 29 Sep 99
Posts: 16515
Credit: 4,418,829
RAC: 0
United Kingdom
Message 715312 - Posted: 19 Feb 2008, 15:19:26 UTC - in response to Message 715264.  


I tried upping the cache to 8 days,



So many ways to control the cache, so many different behaviors.

I've gone with setting "local preferences" to:
    Connect about every 0.0 days  Additional work buffer 5.0 days 

Setting connect interval to 5 days causes download havoc with the flood of "shorties" coming down the pipe due to the short deadlines.

YMMV


Very true. I think I am getting a flood of shorties now that WUs are now coming down in droves.
It's good to be back amongst friends and colleagues



ID: 715312 · Report as offensive
Profile John Clark
Volunteer tester
Avatar

Send message
Joined: 29 Sep 99
Posts: 16515
Credit: 4,418,829
RAC: 0
United Kingdom
Message 715318 - Posted: 19 Feb 2008, 15:29:04 UTC - in response to Message 715232.  

Had similar problem on one of mine recently, the clock had gone forward a couple of months for some unknown reason. But the net result was that the line;

While BOINC running, % of time work is allowed 99.9922 %

had reduced to about 23.nnnn%

The line reporting this is in client_state.xml file on line;
<cpu_efficiency>0.999922</cpu_efficiency>

If this is case stop BOINC, edit to reasonable figure, in format shown, and restart BOINC.

I think that particular error would be in

<active_frac>0.999938</active_frac>

but the principle is good - check all the metrics in <time_stats>

While you're in there, perhaps also check

<duration_correction_factor>0.276953</duration_correction_factor>

(there's a separate one for each project). Anything more than 0.500000 for SETI is suspect - with an optimised application, it'll probably be below 0.300000, as here.



Thanks both to WinterKnight and Richard Haselgrove I now have a flood of WUs to tide me over the Planned Outrage in a few hours.

I went in to both files - client_state.xml and the client_state_previous.xml and found the statement for the <active frac> was given as 0.070707 and 0.070167.

As soon as I amended these to 0.999938 and rebooted BOINC the WUs flooded down, with a request for 1,294,796 seconds of work.

I will let the rig crunch like this for a few days, then reduce my cache in the General Preferences to one third the setting I currently demand.

Many thanks, and I hope to see WUs coming down as work is crunched.
It's good to be back amongst friends and colleagues



ID: 715318 · Report as offensive
Profile Dr. C.E.T.I.
Avatar

Send message
Joined: 29 Feb 00
Posts: 16019
Credit: 794,685
RAC: 0
United States
Message 715322 - Posted: 19 Feb 2008, 15:36:44 UTC


@ JC - got a whole bunch of those shorties - crunched in like 32 min on dv9060us Duo Core 2 . . . returned and waitin' , , ,

found that i was crunchin (Wingman) with Mark Sattler ;))

Hope you are well and all is fine w/ You & Family Sir!

richard
BOINC Wiki . . .

Science Status Page . . .
ID: 715322 · Report as offensive
Profile John Clark
Volunteer tester
Avatar

Send message
Joined: 29 Sep 99
Posts: 16515
Credit: 4,418,829
RAC: 0
United Kingdom
Message 715323 - Posted: 19 Feb 2008, 15:42:07 UTC

I am good and you will get rid of the shorties soon. Be interesting to time your shorty WU time with Mark's.

Shorties for me take between 14 minutes (on the new rig) and 20 on the older one.
It's good to be back amongst friends and colleagues



ID: 715323 · Report as offensive
Profile Dr. C.E.T.I.
Avatar

Send message
Joined: 29 Feb 00
Posts: 16019
Credit: 794,685
RAC: 0
United States
Message 715329 - Posted: 19 Feb 2008, 16:10:12 UTC - in response to Message 715323.  

I am good and you will get rid of the shorties soon. Be interesting to time your shorty WU time with Mark's.

Shorties for me take between 14 minutes (on the new rig) and 20 on the older one.


i believe that the last one i finished was 32 minutes on the Core 2 Duo . . .

Mark hasn't started that one yet - as far as i know

< Q. is your System the new Dual you mentioned to mi on Skype?


BOINC Wiki . . .

Science Status Page . . .
ID: 715329 · Report as offensive
Profile John Clark
Volunteer tester
Avatar

Send message
Joined: 29 Sep 99
Posts: 16515
Credit: 4,418,829
RAC: 0
United Kingdom
Message 715333 - Posted: 19 Feb 2008, 16:20:29 UTC - in response to Message 715329.  

I am good and you will get rid of the shorties soon. Be interesting to time your shorty WU time with Mark's.

Shorties for me take between 14 minutes (on the new rig) and 20 on the older one.


i believe that the last one i finished was 32 minutes on the Core 2 Duo . . .

Mark hasn't started that one yet - as far as i know

< Q. is your System the new Dual you mentioned to mi on Skype?



No! It's a Penryn Quad with air cooling, running at 4.0GHz
It's good to be back amongst friends and colleagues



ID: 715333 · Report as offensive
Profile Dr. C.E.T.I.
Avatar

Send message
Joined: 29 Feb 00
Posts: 16019
Credit: 794,685
RAC: 0
United States
Message 715336 - Posted: 19 Feb 2008, 16:28:34 UTC - in response to Message 715333.  

I am good and you will get rid of the shorties soon. Be interesting to time your shorty WU time with Mark's.

Shorties for me take between 14 minutes (on the new rig) and 20 on the older one.


i believe that the last one i finished was 32 minutes on the Core 2 Duo . . .

Mark hasn't started that one yet - as far as i know

< Q. is your System the new Dual you mentioned to mi on Skype?



No! It's a Penryn Quad with air cooling, running at 4.0GHz


YUMMY


BOINC Wiki . . .

Science Status Page . . .
ID: 715336 · Report as offensive
Profile SATAN
Avatar

Send message
Joined: 27 Aug 06
Posts: 835
Credit: 2,129,006
RAC: 0
United Kingdom
Message 715339 - Posted: 19 Feb 2008, 16:43:09 UTC

No fair i wanna over clock.
ID: 715339 · Report as offensive
Profile John Clark
Volunteer tester
Avatar

Send message
Joined: 29 Sep 99
Posts: 16515
Credit: 4,418,829
RAC: 0
United Kingdom
Message 715351 - Posted: 19 Feb 2008, 17:21:55 UTC - in response to Message 715339.  

No fair i wanna over clock.


Not competitive with your V8 Mac though!
It's good to be back amongst friends and colleagues



ID: 715351 · Report as offensive
Profile SATAN
Avatar

Send message
Joined: 27 Aug 06
Posts: 835
Credit: 2,129,006
RAC: 0
United Kingdom
Message 715353 - Posted: 19 Feb 2008, 17:33:21 UTC

So, I still want to over clock it!

Maybe i'll figure a way one day.
ID: 715353 · Report as offensive

Message boards : Number crunching : The cache & downloading WUs?


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.