More Work !!! Please :)
Message boards :
Number crunching :
More Work !!! Please :)
Message board moderation
| Author | Message |
|---|---|
 Crunch3r Send message Joined: 17 Feb 08 Posts: 363 Credit: 258,227,990 RAC: 0   |
Howdy folks! there are things that bug me.... first thing is that i can only get 20 wus on my V8... that suckz... Can we get rid of that limiting factor ? second thing IF #1 fails(getting more wus) is that wus are way to short ! my V8 is computing one WU in about 310 sec... can we increase that to about 4 times the time ?  Join Support science! Joinc Team BOINC United now! |
Dave Przybylo Send message Joined: 5 Feb 08 Posts: 236 Credit: 49,648 RAC: 0 
|
Howdy folks! well....there is a coded maximum of 20 wus per cycle. and we have to think about slower systems when setting that factor. i'll talk with travis and see what he thinks if he doesn't see this thread. Dave Przybylo MilkyWay@home Developer Department of Computer Science Rensselaer Polytechnic Institute |
Jayargh Send message Joined: 8 Oct 07 Posts: 289 Credit: 3,690,838 RAC: 0 
|
This question has been answered in numerous threads,Any more than 20 at a time slows down the searches as it is an evolving process... excess work runs the risk of being outdated before it is returned.The amount of time each unit takes will get progressively longer in time as the volume of stars used increases. We are of course here to help the project not to have the project design suit our needs. Travis is trying to get the time delay of 20 minutes between rpc calls reduced to help dead zones on fast 8 cores and 16 cores from running out of work. Seems at the moment the server is sending no work. |
|
Dave Przybylo Send message Joined: 5 Feb 08 Posts: 236 Credit: 49,648 RAC: 0
|
This question has been answered in numerous threads,Any more than 20 at a time slows down the searches as it is an evolving process... excess work runs the risk of being outdated before it is returned.The amount of time each unit takes will get progressively longer in time as the volume of stars used increases. This is correct. I think the optimum solution would be to reduce rpc call time which we should have implemented fairly soon. Dave Przybylo MilkyWay@home Developer Department of Computer Science Rensselaer Polytechnic Institute |
|
Swordfish Send message Joined: 19 Nov 07 Posts: 4 Credit: 82,330,797 RAC: 0 
|
looks like we drained the WU dry,cant get any on my systems |
|
Crunch3r Send message Joined: 17 Feb 08 Posts: 363 Credit: 258,227,990 RAC: 0 |
This question has been answered in numerous threads,Any more than 20 at a time slows down the searches as it is an evolving process... excess work runs the risk of being outdated before it is returned.The amount of time each unit takes will get progressively longer in time as the volume of stars used increases. Well if that's the case... any chance to triple the WU "size" ? I mean 310 sec for one wu is a bit low.... Need more work :) Join Support science! Joinc Team BOINC United now! |
|
Jayargh Send message Joined: 8 Oct 07 Posts: 289 Credit: 3,690,838 RAC: 0
|
... any chance to triple the WU "size" ? I mean 310 sec for one wu is a bit low.... Its all perspective ...slow computers can take 1/2 hour to 1 hour and longer....you weren't here when the 1st work would have taken about 100 sec on your host :) I think the studies need to be completed before volumes of stars are increased...it might take a while...but I don't have those answers ;) |
|
Send message Joined: 10 Nov 07 Posts: 96 Credit: 29,931,027 RAC: 0 
|
Well if that's the case... any chance to triple the WU "size" ? I mean 310 sec for one wu is a bit low.... We’ve already had one such tripling. It’s my impression that the plan is to gradually ‘ramp up’ the size & complexity of tasks as the admins gain experience about how the number of stars modelled and whatever other inputs they provide will affect crunching times & throughput. I imagine the relation is more or less exponential, so small increases in the ‘input size’ could greatly increase the work involved. A couple of years ago SzTAKI Desktop Grid made a “minimal†increase in the size of its tasks that resulted in running times something like two orders of magnitude (IIRC) longer than they had been, and wildly unpredictable to boot. A rash of “maximum CPU time exceeded†errors ensued, to the dismay of many participants … Of course it’s a very different kind of app over there, but the same principles may apply. As this project is still at the “testing†stage, as opposed to “productionâ€, I can understand the developer’s desire to proceed incrementally.  |
|
Travis Send message Joined: 30 Aug 07 Posts: 2046 Credit: 26,480 RAC: 0  |
are you guys still not getting work? it looks like milkyway was down for some time which might have caused that... |
|
Travis Send message Joined: 30 Aug 07 Posts: 2046 Credit: 26,480 RAC: 0 |
ok there should be work available soon. the assimilator had crashed and thats why no new WUs were being generated. |
|
Crunch3r Send message Joined: 17 Feb 08 Posts: 363 Credit: 258,227,990 RAC: 0 |
ok there should be work available soon. the assimilator had crashed and thats why no new WUs were being generated. Thanks. it's working again. Join Support science! Joinc Team BOINC United now! |
|
Crunch3r Send message Joined: 17 Feb 08 Posts: 363 Credit: 258,227,990 RAC: 0 |
|
|
Send message Joined: 7 Oct 07 Posts: 5 Credit: 72,480 RAC: 0
|
heeeeeeeeeelp out of work!!!!! |
|
Travis Send message Joined: 30 Aug 07 Posts: 2046 Credit: 26,480 RAC: 0 |
heeeeeeeeeelp out of work!!!!! should be a bunch more work waiting for you guys now :) 
|
|
Crunch3r Send message Joined: 17 Feb 08 Posts: 363 Credit: 258,227,990 RAC: 0 |
Results ready to send 0 :(  Join Support science! Joinc Team BOINC United now! |
|
Travis Send message Joined: 30 Aug 07 Posts: 2046 Credit: 26,480 RAC: 0 |
it should be genrating more work, you guys getting any yet?
|
|
Crunch3r Send message Joined: 17 Feb 08 Posts: 363 Credit: 258,227,990 RAC: 0 |
no... 07.03.2008 18:31:47|Milkyway@home|Sending scheduler request: Requested by project 07.03.2008 18:31:47|Milkyway@home|Requesting 407189 seconds of new work 07.03.2008 18:31:52|Milkyway@home|Scheduler RPC succeeded [server version 511] 07.03.2008 18:31:52|Milkyway@home|Deferring communication for 3 hr 36 min 0 sec 07.03.2008 18:31:52|Milkyway@home|Reason: no work from project Join Support science! Joinc Team BOINC United now! |
ChertseyAl Send message Joined: 31 Aug 07 Posts: 66 Credit: 1,002,668 RAC: 0 
|
Yes! Got some for 2 hosts. The second host didn't get the full amount requested - I guess I drained you dry ;) Al. |
|
Travis Send message Joined: 30 Aug 07 Posts: 2046 Credit: 26,480 RAC: 0 |
yeah i think theres a lot of people trying to get work right now so it might take awhile for the queue to get filled up
|
|
Crunch3r Send message Joined: 17 Feb 08 Posts: 363 Credit: 258,227,990 RAC: 0 |
|
©2024 Astroinformatics Group