Does AP get priority for download??

Message boards : Number crunching : Does AP get priority for download??
Message board moderation

To post messages, you must log in.

1 · 2 · Next

AuthorMessage
Dave Stegner
Volunteer tester
Avatar

Send message
Joined: 20 Oct 04
Posts: 540
Credit: 65,583,328
RAC: 27
United States
Message 886141 - Posted: 18 Apr 2009, 0:19:08 UTC

3 weeks ago or so I decided to try AP again.

I set up 12 of my 22 machines to accept MB and AP ver 503. The others accept MB only or Ap only. I am using opti app and for AP only have Ap503 in my app_info.

After 3 weeks the 12 machines have nothing but Ap. Seems a bit strange as my MB only machines have not come close to running out of work and the last 3 weeks work has been reasonably available.

Anyone else seen this??


Dave

ID: 886141 · Report as offensive
Profile popandbob
Volunteer tester

Send message
Joined: 19 Mar 05
Posts: 551
Credit: 4,673,015
RAC: 0
Canada
Message 886147 - Posted: 18 Apr 2009, 0:33:41 UTC

You need to include MB in the app info.


Do you Good Search for Seti@Home? http://www.goodsearch.com/?charityid=888957
Or Good Shop? http://www.goodshop.com/?charityid=888957
ID: 886147 · Report as offensive
Cosmic_Ocean
Avatar

Send message
Joined: 23 Dec 00
Posts: 3027
Credit: 13,516,867
RAC: 13
United States
Message 886192 - Posted: 18 Apr 2009, 3:51:27 UTC

This has been discussed numerous times, and I'm usually the one mentioning or re-affirming the point, but I believe there is some broken logic on the scheduler that tends to give you AP if you have that selected as an app you want work for.

Switching any of my hosts to MB-only have no problems getting the daily quota of tasks in a short period of time, when the queue is set to NNT and drained first. A few weeks ago, I enabled new tasks on my 2p dual-core rig and requested 2.4M seconds of work, and got 157 MBs with just 10 or 11 requests (20 at a time for most, less than 20 for others).

So there is no shortage of MB out there, but the scheduler just doesn't seem to want to give out anything else.
Linux laptop:
record uptime: 1511d 20h 19m (ended due to the power brick giving-up)
ID: 886192 · Report as offensive
Profile KW2E
Avatar

Send message
Joined: 18 May 99
Posts: 346
Credit: 104,396,190
RAC: 34
United States
Message 886194 - Posted: 18 Apr 2009, 3:56:46 UTC - in response to Message 886192.  

I can also confirm this. On the vast majority of machines I have AP on, that's all they get. They are set to accept everything. Ever so often I will see a 603 or two mixed in there, but nothing in any great quantity.

Rob
ID: 886194 · Report as offensive
Profile [B^S] madmac
Volunteer tester
Avatar

Send message
Joined: 9 Feb 04
Posts: 1175
Credit: 4,754,897
RAC: 0
United Kingdom
Message 886218 - Posted: 18 Apr 2009, 6:41:15 UTC
Last modified: 18 Apr 2009, 6:41:28 UTC

I too can confirm this, I have got MB in my app file and all I got was APs. Perhaps something can be done so that we get a mixture and I will happily take 2 days to crunch an AP whilst my MB take about 2Hrs
ID: 886218 · Report as offensive
Profile arkayn
Volunteer tester
Avatar

Send message
Joined: 14 May 99
Posts: 4438
Credit: 55,006,323
RAC: 0
United States
Message 886492 - Posted: 19 Apr 2009, 22:14:52 UTC

I can chime in and also confirm this.

I just change my venue between one that allows AP and one that does not just to keep a little sanity on my machines.

ID: 886492 · Report as offensive
Dave Stegner
Volunteer tester
Avatar

Send message
Joined: 20 Oct 04
Posts: 540
Credit: 65,583,328
RAC: 27
United States
Message 886496 - Posted: 19 Apr 2009, 22:30:16 UTC
Last modified: 19 Apr 2009, 22:32:01 UTC

PopandBob
For these 12 machines I do have MB and AP in App_info.

Maybe the scheduler is broken or maybe it is doing exactly what someone wants it to do.
Dave

ID: 886496 · Report as offensive
Josef W. Segur
Volunteer developer
Volunteer tester

Send message
Joined: 30 Oct 99
Posts: 4504
Credit: 1,414,761
RAC: 0
United States
Message 886534 - Posted: 20 Apr 2009, 2:18:51 UTC - in response to Message 886496.  

PopandBob
For these 12 machines I do have MB and AP in App_info.

Maybe the scheduler is broken or maybe it is doing exactly what someone wants it to do.

Or maybe the project staff is overworked and underpaid so haven't had time to adjust settings which control how much of each kind of work the Feeder tells the Scheduler about.

What has worked for me is to have only SETI@home Enhanced checked for the three types, but also allow "work for other applications". I usually get MB, but sometimes there isn't any and I get AP_v5. The balance seems reasonable so far, about 25 MBs for each AP_v5.
                                                                 Joe
ID: 886534 · Report as offensive
Profile Kinguni
Volunteer tester
Avatar

Send message
Joined: 15 Feb 00
Posts: 239
Credit: 9,043,007
RAC: 0
Canada
Message 886544 - Posted: 20 Apr 2009, 2:46:04 UTC - in response to Message 886534.  

I was getting too much MB for a while after the big database crash so I set my preferences only to do AP unless no AP work is available. Longer wait for the credit but better credit per hour. Before the big database failure though I was seeing what everyone else was seeing - nothing but AP.
Join Team Starfire
BOINC Chat

ID: 886544 · Report as offensive
Profile KW2E
Avatar

Send message
Joined: 18 May 99
Posts: 346
Credit: 104,396,190
RAC: 34
United States
Message 886551 - Posted: 20 Apr 2009, 3:15:05 UTC - in response to Message 886544.  

I wasn't complaining. I want all my machines running AP with the exception of my slower machines. AP with the opt apps gives great RAC. :)

Rob
ID: 886551 · Report as offensive
Profile Timi
Volunteer tester

Send message
Joined: 7 Oct 99
Posts: 25
Credit: 6,533,108
RAC: 0
Greece
Message 886566 - Posted: 20 Apr 2009, 5:22:04 UTC

I also confirm that I only got AP_v5.03 workunits on 3 machines, which also have the MB section on app_info.xml file. The machines' capability of AP workunits is 12hrs, 25hrs and 51hrs (if it has to do something with processing speed).
ID: 886566 · Report as offensive
Andy Williams
Volunteer tester
Avatar

Send message
Joined: 11 May 01
Posts: 187
Credit: 112,464,820
RAC: 0
United States
Message 886568 - Posted: 20 Apr 2009, 5:25:38 UTC - in response to Message 886566.  

Same on all my machines.
--
Classic 82353 WU / 400979 h
ID: 886568 · Report as offensive
Profile Fred J. Verster
Volunteer tester
Avatar

Send message
Joined: 21 Apr 04
Posts: 3252
Credit: 31,903,643
RAC: 0
Netherlands
Message 886612 - Posted: 20 Apr 2009, 11:20:11 UTC - in response to Message 886568.  

Mostly > 90%, AP WU's, wander where all the MB WU's are?

BOINC 6.6.20. keeps better track of recourse settings, then some of the previous versions.
If that's the reason, why so little MB WU's are available, I don't know.

ID: 886612 · Report as offensive
Profile mr.mac52
Avatar

Send message
Joined: 18 Mar 03
Posts: 67
Credit: 245,882,461
RAC: 0
United States
Message 886625 - Posted: 20 Apr 2009, 12:39:19 UTC

For what its worth, my mix of MB 6.08 to AP 5.03 is about 10 to 1 with Raistmer's optimized apps. This is from monitoring it for about the past week or so.
ID: 886625 · Report as offensive
1mp0£173
Volunteer tester

Send message
Joined: 3 Apr 99
Posts: 8423
Credit: 356,897
RAC: 0
United States
Message 886822 - Posted: 20 Apr 2009, 23:29:38 UTC - in response to Message 886612.  

Mostly > 90%, AP WU's, wander where all the MB WU's are?

BOINC 6.6.20. keeps better track of recourse settings, then some of the previous versions.
If that's the reason, why so little MB WU's are available, I don't know.

They aren't here. I crunch only AP.
ID: 886822 · Report as offensive
Cosmic_Ocean
Avatar

Send message
Joined: 23 Dec 00
Posts: 3027
Credit: 13,516,867
RAC: 13
United States
Message 886868 - Posted: 21 Apr 2009, 2:53:28 UTC - in response to Message 886534.  

PopandBob
For these 12 machines I do have MB and AP in App_info.

Maybe the scheduler is broken or maybe it is doing exactly what someone wants it to do.

Or maybe the project staff is overworked and underpaid so haven't had time to adjust settings which control how much of each kind of work the Feeder tells the Scheduler about.

What has worked for me is to have only SETI@home Enhanced checked for the three types, but also allow "work for other applications". I usually get MB, but sometimes there isn't any and I get AP_v5. The balance seems reasonable so far, about 25 MBs for each AP_v5.
                                                                 Joe

Completely understandable, but I have all for checkboxes selected in my prefs (mb, ap, ap_v5, allow for others), and I get 100% AP..always, no matter what.

The only way I can get MB is to go to my MB-only venue.

I used to get the same as you, ~25 MBs for every AP, but then all this CUDA stuff came out and the 6.6.x family of clients with the fancy scheduling, and now the scheduler won't give me MB unless I force it.
Linux laptop:
record uptime: 1511d 20h 19m (ended due to the power brick giving-up)
ID: 886868 · Report as offensive
Profile popandbob
Volunteer tester

Send message
Joined: 19 Mar 05
Posts: 551
Credit: 4,673,015
RAC: 0
Canada
Message 886898 - Posted: 21 Apr 2009, 6:38:20 UTC - in response to Message 886868.  

but then all this CUDA stuff came out


With the servers set to dish out equal amounts of work the CUDA clients would get all the MB the server has and CPU's the AP.

That makes sense now... Its an unintentional side effect from fixing another problem and releasing cuda.

We didn't notice before due to there was never any AP ready to send (before v5) so MB were abundant.

~Bob


Do you Good Search for Seti@Home? http://www.goodsearch.com/?charityid=888957
Or Good Shop? http://www.goodshop.com/?charityid=888957
ID: 886898 · Report as offensive
Ianab
Volunteer tester

Send message
Joined: 11 Jun 08
Posts: 732
Credit: 20,635,586
RAC: 5
New Zealand
Message 886907 - Posted: 21 Apr 2009, 7:32:31 UTC

I would suggest that there are four groups of crunchers.

People with powerfull CUDA machines. These grab lots of MB units.

People that have powerfull PCs but are scared or bored of AP units and have selected not to do them

People that have old or unreliable PCs. The smart ones have turned of AP, and the others just return errors, or time out.

The fourth group are the ones that have PCs with enough grunt to run AP, AND choose to do them, preferably with the optimised apps.

So if you are in group 4, you get the AP units, because there are lots of other people requesting MB, and probably a shortage of machines requesting AP. Given the choice, the scheduler sends everyone in group 4 AP units to clear the backlog.

Thats just my thinking anyway. I have switched most of my more powerfull machines to MB + AP, and they are getting AP units as others have reported. But with the optimised apps I'm earning mmore credits. Even an old PIII can do an optimised AP unit in about 10days.

Just relax and let the scheduler give out the work that needs to be done :-)

Ian
ID: 886907 · Report as offensive
Profile Virtual Boss*
Volunteer tester
Avatar

Send message
Joined: 4 May 08
Posts: 417
Credit: 6,440,287
RAC: 0
Australia
Message 886913 - Posted: 21 Apr 2009, 8:54:14 UTC

In the long term the overall ratio of AP to MB will be stable at X:Y.

There are X number of AP and Y number of MB per "tape".

Its just each cruncher that has a different ratio, but they all balance out.
ID: 886913 · Report as offensive
daysteppr Project Donor

Send message
Joined: 22 Mar 05
Posts: 80
Credit: 19,575,419
RAC: 53
United States
Message 886929 - Posted: 21 Apr 2009, 12:04:14 UTC

I seem to get nothing but AP. I got some MB WU's like 2 weeks ago ( a total of 6) and its been nothing but AP since March I think.

I have all 3 selected, and use the r112 optimized app ( thanks to whoever wrote that btw!) but for me? Its AP, all the time

Than again, I have my cache set at 1+2 days. Maybe thats the reason why.

Sincerely,
Daysteppr
ID: 886929 · Report as offensive
1 · 2 · Next

Message boards : Number crunching : Does AP get priority for download??


 
©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.