| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: Website Support
Thread: Pre-set device profiles |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 23
|
|
| Author |
|
|
robertmiles
Senior Cruncher US Joined: Apr 16, 2008 Post Count: 445 Status: Offline Project Badges: 
|
The change saved, and I used the Update function. I even saw the web-site change on some of the other BOINC projects I participate in, just no sign that the CPU use actually decreased. Of course you have to tell BOINC to run based on preferences in the activity menu. You can confirm the settings copied to your client by opening the global_prefs.xml file and look in the section that has the name of the device profile associated with the device. Added: Of course the start up message log of BOINC shows which profile is in use and when last updated e.g. 2/7/2009 11:28:44 PM||General prefs: from World Community Grid (last modified 05-Feb-2009 23:52:58) Does this give sufficient information about the settings? 2/10/2009 1:29:48 PM||Starting BOINC client version 6.2.28 for windows_intelx86 2/10/2009 1:29:48 PM||log flags: task, file_xfer, sched_ops 2/10/2009 1:29:48 PM||Libraries: libcurl/7.19.0 OpenSSL/0.9.8i zlib/1.2.3 2/10/2009 1:29:48 PM||Running as a daemon 2/10/2009 1:29:48 PM||Data directory: C:\ProgramData\BOINC 2/10/2009 1:29:48 PM||Running under account boinc_master 2/10/2009 1:29:48 PM|boincsimap|Found app_info.xml; using anonymous platform 2/10/2009 1:29:53 PM||Processor: 2 AuthenticAMD AMD Athlon(tm) 64 X2 Dual Core Processor 3600+ [x86 Family 15 Model 107 Stepping 1] 2/10/2009 1:29:53 PM||Processor features: fpu tsc pae nx sse sse2 pni 3dnow mmx 2/10/2009 1:29:53 PM||OS: Microsoft Windows Vista: Home Premium x86 Editon, Service Pack 1, (06.00.6001.00) 2/10/2009 1:29:53 PM||Memory: 1.87 GB physical, 4.62 GB virtual 2/10/2009 1:29:53 PM||Disk: 224.09 GB total, 163.44 GB free 2/10/2009 1:29:53 PM||Local time is UTC -6 hours 2/10/2009 1:29:53 PM|rosetta@home|URL: http://boinc.bakerlab.org/rosetta/; Computer ID: 835364; location: home; project prefs: home 2/10/2009 1:29:53 PM|boincsimap|URL: http://boinc.bio.wzw.tum.de/boincsimap/; Computer ID: 115552; location: home; project prefs: home 2/10/2009 1:29:53 PM|Poem@Home|URL: http://boinc.fzk.de/poem/; Computer ID: 27564; location: (none); project prefs: default 2/10/2009 1:29:53 PM|Cels@Home|URL: http://cels-at-home-dev.dyndns.org/cels/; Computer ID: 6267; location: (none); project prefs: default 2/10/2009 1:29:53 PM|Predictor @ Home|URL: http://predictor.chem.lsa.umich.edu/; Computer ID: 370762; location: home; project prefs: home 2/10/2009 1:29:53 PM|ralph@home|URL: http://ralph.bakerlab.org/; Computer ID: 16985; location: (none); project prefs: default 2/10/2009 1:29:53 PM|malariacontrol.net|URL: http://www.malariacontrol.net/; Computer ID: 98632; location: (none); project prefs: default 2/10/2009 1:29:53 PM|World Community Grid|URL: http://www.worldcommunitygrid.org/; Computer ID: 574785; location: home; project prefs: home 2/10/2009 1:29:53 PM||General prefs: from World Community Grid (last modified 10-Feb-2009 11:31:06) 2/10/2009 1:29:53 PM||Computer location: home 2/10/2009 1:29:53 PM||General prefs: using separate prefs for home 2/10/2009 1:29:53 PM||Reading preferences override file 2/10/2009 1:29:53 PM||Preferences limit memory usage when active to 1438.32MB 2/10/2009 1:29:53 PM||Preferences limit memory usage when idle to 1725.99MB 2/10/2009 1:29:53 PM||Preferences limit disk usage to 27.94GB 2/10/2009 1:29:54 PM||Suspending network activity - user request However, I finally got a working procedure over at RALPH@home. Advanced menu/preferences ... make the setting ... click "Ok" button ... Changing the setting there from 100% to 90% finally worked. I don't remember ever using that function before, so I suspect that one of the BOINC projects I participate in made that setting without telling me, and without providing any way to undo it. I also got a procedure for going back to the web settings; untested so far. Advanced menu/preferences ... click "Clear" button ... I'll try that when I no longer need the 90% setting for RALPH@home tests. |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
robertmiles,
----------------------------------------That is a well discussed option here and reflected in the log by the "2/10/2009 1:29:53 PM||Reading preferences override file" line which are the locally entered settings. It's often causing conflict. The Clear button operation ensures that the website prefs are used and works well to restore sanity to the controls. Your log does say your web pref modifications were copied over by: 2/10/2009 1:29:53 PM||General prefs: from World Community Grid (last modified 10-Feb-2009 11:31:06) 2/10/2009 1:29:53 PM||Computer location: home 2/10/2009 1:29:53 PM||General prefs: using separate prefs for home In brief, if you want to manage multiple machines from a single profile you do it from the Device Manager on My Grid. If, when running multiple devices, you want specific changes applied to a single machine you do it from local preferences. If that is not understood you get the seemingly "prefs not sticking" experience. The foregoing is assuming that the website screens work according expectation and the modified 10-Feb-2009 11:31:06 confirms this. Was the Activity Menu option "Run based on Preferences" selected? Added: The CPU setting of 90% of time means the client will run 100% for 9 seconds and 0% for 1 second. Effectively a rather useless control which is why I use ThreadMaster GUI. That free tool for W2K and up allows managing the science use of the CPU so e.g. 80% is really 80%, all the time. Then the 20% is permanently available without causing jitter when watching a video (on machines that suffer that issue). It works on Vista/W2K8 and probably the upcoming Windows 7 too, if you need it.
WCG
----------------------------------------Please help to make the Forums an enjoyable experience for All! [Edit 1 times, last edit by Sekerob at Feb 11, 2009 1:41:58 PM] |
||
|
|
robertmiles
Senior Cruncher US Joined: Apr 16, 2008 Post Count: 445 Status: Offline Project Badges:
|
robertmiles, Was the Activity Menu option "Run based on Preferences" selected? No, I only recently noticed that possibility. |
||
|
|
robertmiles
Senior Cruncher US Joined: Apr 16, 2008 Post Count: 445 Status: Offline Project Badges:
|
Run based on Preferences worked for the last few months, but it now looks like I don't need the different setting for that machine any more so I'm about to remove its preference file.
----------------------------------------[Edit 1 times, last edit by robertmiles at Sep 8, 2009 5:45:42 AM] |
||
|
|
robertmiles
Senior Cruncher US Joined: Apr 16, 2008 Post Count: 445 Status: Offline Project Badges:
|
Could this web page be modified to include a link to instructions on how to do device-specific project settings?
----------------------------------------https://secure.worldcommunitygrid.org/ms/device/viewProfiles.do Could the device-specific project settings page be modified to allow not setting the same computer location for both WCG projects and all other BOINC projects? I now want to use frequent connections for WCG to encourage getting Beta Testing work, but less frequent connections for some other BOINC projects at sites that want less frequent connections due to running low on their server capacities. Using the local preferences method for each machine (at least in BOINC 6.6.36 for Windows) currently doesn't seem to offer any option for allowing different locations for different projects, or different connection schedules for WCG and non-WCG projects. Could the device-specific settings page be modified to allow preservation of any settings to allow use of the GPU for non-WCG projects, such as GPUGRID? Feel free to allow it to include an automatic setting not to allow use of the GPU, but for WCG only, if this is possible, until WCG has a project that can actually use a GPU, since this should reduce the load on your server. [Edit 1 times, last edit by robertmiles at Sep 22, 2009 3:24:53 PM] |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
From the top:
----------------------------------------1. You can associate device profiles (officially 4, unsupported, you can create as many as you want), to specific devices, but there is no such thing as device specific project settings. 2. You can specify in the device manager which profile to associate with what device, to include the unsupported profiles you can name anything you like. It's unofficial/unsupported, thus you wont find an FAQ on how to. 3. No, a device is associate with one profile for all projects. You can't have different profiles for different projects. That creates an impossible conflict for the continued operation of BOINC. Each project only controls the project weight/resource share, but all else is pretty much on the basis of: Make a change in a profile at any project and they apply to all projects i.e. the last change overrules all profiles of all other projects i.e. as the name says "global_prefs" with local mods named "global_prefs_override". 4. Don't think WCG profiles would effect specific GPU options set by other projects. You can presently enable/disable GPU crunching in the cc_config.xml of each device. Maybe I missed something, but then if you draw a diagram to help visualize what you have in mind it may clarify for me and others.
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
Ingleside
Veteran Cruncher Norway Joined: Nov 19, 2005 Post Count: 974 Status: Offline Project Badges: 
|
I now want to use frequent connections for WCG to encourage getting Beta Testing work, but less frequent connections for some other BOINC projects with less server capacity but often with no workunits available. The connection-settings is global, you can't have separate settings for individual projects. Also, BOINC won't make extra connections asking for work if the cache is full with work, so changing your cache-settings has little or no effect on how often you asks for work. Setting a small "Connect about every..." don't mean client will connect example once every 14.4 minutes even if this setting is set to 0.01 days, and actually if you've got a permanent connection the recommended setting is zero. For anyone permanently connected, the cache-size should be specified by using "Additional..." instead. As for the other projects, if they're constantly out of work, they'll after some time be asked once per day, so shouldn't be a problem. Could the device-specific settings page be modified to allow preservation of any settings to allow use of the GPU for non-WCG projects, such as GPUGRID? Feel free to allow it to include an automatic setting not to allow use of the GPU, but for WCG only, if this is possible, until WCG has a project that can actually use a GPU, since this should reduce the load on your server. The BOINC-client will ask for work for all hardware-devices it detects, as long as it's not specifically disabled by either an addition in cc_config.xml, or for GPU is disabled by running BOINC as a service on vista or win7. Any project-specific settings of application-type to use or not use, like example only do Aids-work or do a combination of Aids and Cancer, has no meaning to the BOINC-client. It's the Scheduling-server in combination with the users project-specific choises that decides this, and there's no reason to change this behaviour for non-cpu-work. As for work-request for a hardware-type project either doesn't have any application for, or user has disabled in the project-specific settings, project can ask v6.10.5 and later clients to not ask again for this hardware-type (including cpu) for upto 4 weeks. If project haven't specified anything or old client, the normal rules of deferral is used, there after some tries there will be one connection-attempt per day for v6.6.xx-clients, and more frequently for older clients.  "I make so many mistakes. But then just think of all the mistakes I don't make, although I might." |
||
|
|
robertmiles
Senior Cruncher US Joined: Apr 16, 2008 Post Count: 445 Status: Offline Project Badges:
|
From the top: 1. You can associate device profiles (officially 4, unsupported, you can create as many as you want), to specific devices, but there is no such thing as device specific project settings. 2. You can specify in the device manager which profile to associate with what device, to include the unsupported profiles you can name anything you like. It's unofficial/unsupported, thus you wont find an FAQ on how to. 3. No, a device is associate with one profile for all projects. You can't have different profiles for different projects. That creates an impossible conflict for the continued operation of BOINC. Each project only controls the project weight/resource share, but all else is pretty much on the basis of: Make a change in a profile at any project and they apply to all projects i.e. the last change overrules all profiles of all other projects i.e. as the name says "global_prefs" with local mods named "global_prefs_override". 4. Don't think WCG profiles would effect specific GPU options set by other projects. You can presently enable/disable GPU crunching in the cc_config.xml of each device. Maybe I missed something, but then if you draw a diagram to help visualize what you have in mind it may clarify for me and others. 1. I probably should have written profile-specific project settings instead. I haven't found any information on how to create profiles other than the official 4. 2. OK. 3. If so, the log from BOINC startup should say which one. As you can see earlier in this thread, for at least some BOINC versions, it can now show that projects from different web sites have different location settings, and BOINC doesn't seem to say it picked just one of them to use unless you create a local preferences file on that machine; it then implies it as as part of the preferences file settings. Also, my experience seems to show that it uses these different profiles for projects from different sites to give them different workunit request policies, if a local preferences file is NOT used. I have no objection if it chooses just one profile to use for the CPU scheduling and GPU scheduling, but I'd like it to be clear about which one. I haven't found any need to use different profiles for different projects when both projects are from WCG; only when they are from different BOINC project sites. For example, Rosetta@home wants less frequent server access for all users than WCG seems to prefer for beta testers. 4. Not the method I've been using; I've been setting GPU access in the local preferences file instead. I'll try your method if I can find either that machine's cc_config.xml file or adequate instructions on how to create one for a machine that doesn't already have one. One problem with my method is that it stops working if the participant decides to stop using a local preferences file on that machine, as I did. |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Your message log says every time the client fetches a profile change from the project website with a newer than last received timestamp and who's project they're being used from:
----------------------------------------22/09/2009 00:30:57 General prefs: from World Community Grid (last modified 18-Aug-2009 13:25:43) 22/09/2009 21:25:48 Computer location: work It will also tell you when the local prefs are changed and that they have been overridden by local prefs: 22/09/2009 21:25:48 Reading preferences override file At each project you can associate the host to a different profile, but it will only use the last one changed, so if on Rosetta the client is associated with home and you make a change at WCG where it's associated with work, work will be the one that's written to the global_prefs. xml file on next client-WCG server contact. 'adequate' instructions you find via the Start Here Forum Index shown as global thread at top of each forum, with explanations for each line that can be added to the cc_config.xml.
WCG
----------------------------------------Please help to make the Forums an enjoyable experience for All! [Edit 1 times, last edit by Sekerob at Sep 22, 2009 8:10:58 PM] |
||
|
|
robertmiles
Senior Cruncher US Joined: Apr 16, 2008 Post Count: 445 Status: Offline Project Badges:
|
I now want to use frequent connections for WCG to encourage getting Beta Testing work, but less frequent connections for some other BOINC projects with less server capacity but often with no workunits available. The connection-settings is global, you can't have separate settings for individual projects. Also, BOINC won't make extra connections asking for work if the cache is full with work, so changing your cache-settings has little or no effect on how often you asks for work. Setting a small "Connect about every..." don't mean client will connect example once every 14.4 minutes even if this setting is set to 0.01 days, and actually if you've got a permanent connection the recommended setting is zero. For anyone permanently connected, the cache-size should be specified by using "Additional..." instead. As for the other projects, if they're constantly out of work, they'll after some time be asked once per day, so shouldn't be a problem. Maybe that's why Rosetta@home seems to have gone to some effort to allow participants to select expected workunit lengths, so that at least the machines with fewer problems do not have to connect to the server as often. Could the device-specific settings page be modified to allow preservation of any settings to allow use of the GPU for non-WCG projects, such as GPUGRID? Feel free to allow it to include an automatic setting not to allow use of the GPU, but for WCG only, if this is possible, until WCG has a project that can actually use a GPU, since this should reduce the load on your server. The BOINC-client will ask for work for all hardware-devices it detects, as long as it's not specifically disabled by either an addition in cc_config.xml, or for GPU is disabled by running BOINC as a service on vista or win7. Any project-specific settings of application-type to use or not use, like example only do Aids-work or do a combination of Aids and Cancer, has no meaning to the BOINC-client. It's the Scheduling-server in combination with the users project-specific choises that decides this, and there's no reason to change this behaviour for non-cpu-work According to what I've seen over at GPUGRID, not completely true for versions of BOINC more recent than about 6.6.20. These versions assume that use of the GPU is disabled unless some effort has been made to enable it. So preserving a setting to enable it, even when changing settings on a site that does not plan to offer workunits that use the GPU any time soon, should be sufficient. As for work-request for a hardware-type project either doesn't have any application for, or user has disabled in the project-specific settings, project can ask v6.10.5 and later clients to not ask again for this hardware-type (including cpu) for upto 4 weeks. If project haven't specified anything or old client, the normal rules of deferral is used, there after some tries there will be one connection-attempt per day for v6.6.xx-clients, and more frequently for older clients. Does not quite work that way for requests for GPU workunits sent to WCG, at least from BOINC 6.6.36 with GPU use enabled by the local preferences file. Such requests stay about as frequent as requests for CPU workunits, and often become frequent again if the cache size is decreased in the local preferences file. Also, the GPUGRID project seems to need some information on what to do at their end to reduce requests for CPU-only workunits, which they currently aren't offering for all the operating systems they offer GPU workunits for. |
||
|
|
|