| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: BOINC Agent Support
Thread: Problem with profile preferences |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 6
|
|
| Author |
|
|
nanoprobe
Master Cruncher Classified Joined: Aug 29, 2008 Post Count: 2998 Status: Offline Project Badges: 
|
I have several machines running other GPU projects that don't get any GPU betas even though I have the preferences set to receive them. It seems like the Allow research to run on my CPU? setting is being ignored. When I set it to no and hit save it still shows yes. I have my local preferences set to Connect to network about every 0.05 and cache set to 0.02 extra days of work but BOINC never polls the website for work. I looked at the messages log for yesterday and never once did BM ask WCG for work. If I'm not sitting at home when the betas go out and manually hit update I get no betas on those 2 machines. Currently running 7.0.31 but it did the same thing with7.0.28 as well. Currently using the default profile setting.
----------------------------------------EDIT: Even though the CPU preference says yes I get no CPU work either so maybe the no setting is working even though the preferences says yes.
In 1969 I took an oath to defend and protect the U S Constitution against all enemies, both foreign and Domestic. There was no expiration date.
----------------------------------------  [Edit 2 times, last edit by nanoprobe at Sep 14, 2012 1:44:38 PM] |
||
|
|
Jim1348
Veteran Cruncher USA Joined: Jul 13, 2009 Post Count: 1066 Status: Offline Project Badges: 
|
EDIT: Even though the CPU preference says yes I get no CPU work either so maybe the no setting is working even though the preferences says yes. That is the way it is for me too. But I had it originally set to "no", and so maybe it just stayed that way. The real question is what happens if someone had it set to "yes". Could they now change it to "no", even though it says "yes"? |
||
|
|
mikey
Veteran Cruncher Joined: May 10, 2009 Post Count: 826 Status: Offline Project Badges: 
|
I have several machines running other GPU projects that don't get any GPU betas even though I have the preferences set to receive them. It seems like the Allow research to run on my CPU? setting is being ignored. When I set it to no and hit save it still shows yes. I have my local preferences set to Connect to network about every 0.05 and cache set to 0.02 extra days of work but BOINC never polls the website for work. I looked at the messages log for yesterday and never once did BM ask WCG for work. If I'm not sitting at home when the betas go out and manually hit update I get no betas on those 2 machines. Currently running 7.0.31 but it did the same thing with7.0.28 as well. Currently using the default profile setting. EDIT: Even though the CPU preference says yes I get no CPU work either so maybe the no setting is working even though the preferences says yes. I think the problem is your cache settings, you have it so short that a single cached unit could be enough to satisfy the 0.5 and the 0.2 is tiny too. Try changing the 0.2 to 0.75 and then hit update and see if you don't get a small handful of units. Boinc ver 7 cache settings are more LITERAL than previous versions. Meaning 0.5 is the MINIMUM amount of estimated time cached before asking for more work, so if you have more than 12 hours of work Boinc thinks you are fine and will NOT ask for more units. The 0.2 setting is for how much EXTRA you get when it does connect. If you crunch multiple projects within WCG or projects outside of WCG you could easily be not asking for more work except one unit at a time. On the pc I am typ8ing this one I have 0.5 in the first cache box and 0.75 ibn the 2nd box and get around 20 HCC units that each take about 1.5 hours, this is a 6 core cpu so it goes thru 6 in that 1.5 hours. This happened this morning: 9/15/2012 9:18:22 AM | World Community Grid | Reporting 2 completed tasks 9/15/2012 9:18:22 AM | World Community Grid | Requesting new tasks for CPU and ATI 9/15/2012 9:18:25 AM | World Community Grid | Scheduler request completed: got 27 new tasks I got all HCC tasks, no Beta units came thru at that time. Although on another pc I did get ONE gpu Beta this morning. I tend to work one project at a time, although I do have some projects that are VERY irregular in their work distribution such as Dengue also selected. As far as getting gpu units, there are thousands of us ASKING for units but only so many units available, so it is purely a hit or miss thing whether you get the gpu Beta units or not. The more pc's you have that can get them the more chance you have to get some, but luck of the draw is how it works. Being in the right place at the right time is the only way. Now if they dropped 1 bazillion gpu units on us, we would ALL get some, but the time between send them out, getting them back, analyzing the results and then sending out another batch could be VERY LONG! Personally I don't mind the 'being lucky' part as if not this time then next time I have an equal chance again.   [Edit 1 times, last edit by mikey159b at Sep 15, 2012 1:32:43 PM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Keep reading about client caching issues, and if it is client caching issues with 7.0.xx, then read the post I made in the W8 thread on 7.0.36 and follow change log link. Specifically there's a fix for work requesting when the GPU is asking to couple this with a CPU work request i.e. a combine of RPC. With 7.0.xx, if the profile were set to not send CPU's that would be printed in the 7.0.28 log. If it is not printed, then either no work is requested, or there is a bug.
|
||
|
|
nanoprobe
Master Cruncher Classified Joined: Aug 29, 2008 Post Count: 2998 Status: Offline Project Badges:
|
I think the problem is your cache settings, you have it so short that a single cached unit could be enough to satisfy the 0.5 and the 0.2 is tiny too. Try changing the 0.2 to 0.75 and then hit update and see if you don't get a small handful of units. I'm only looking to get GPU betas or GPU tasks when the HCC project goes live on these machines. Changing the cache settings and hitting update won't help if there are no tasks available. The client should be periodically asking WCG for work even if there is none available. It never does that and I have tried different cache settings.
In 1969 I took an oath to defend and protect the U S Constitution against all enemies, both foreign and Domestic. There was no expiration date.
|
||
|
|
mikey
Veteran Cruncher Joined: May 10, 2009 Post Count: 826 Status: Offline Project Badges:
|
I think the problem is your cache settings, you have it so short that a single cached unit could be enough to satisfy the 0.5 and the 0.2 is tiny too. Try changing the 0.2 to 0.75 and then hit update and see if you don't get a small handful of units. I'm only looking to get GPU betas or GPU tasks when the HCC project goes live on these machines. Changing the cache settings and hitting update won't help if there are no tasks available. The client should be periodically asking WCG for work even if there is none available. It never does that and I have tried different cache settings. What I do is check the box to yes accept gpu work, then I also have the settings set to yes receive Beta tasks, then I just wait. With multiple machines I am connecting every hour or so so get whatever I need or is available from my list of what I want. YES I also get Beta tasks for other projects that are not just gpu units as they become available, but it seems to be a one setting fits all thing. |
||
|
|
|