| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: FightAIDS@Home
Thread: FightAIDS@Home Estimated Completion Date ( June 4, 2015 ) ESTIMATE!!! |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 325
|
|
| Author |
|
|
Rickjb
Veteran Cruncher Australia Joined: Sep 17, 2006 Post Count: 666 Status: Offline Project Badges: 
|
@Sekerob, re. app_config
AFAIK, the set of max_concurrent values that you specify in your app_config.xml has no effect whatsoever on the project mix of the WUs that the BOINC server sends you. Our BOINC clients can't specify the mix of WUs received - that's done at the server end, by your device profile, project priority and feeder contents at the time your client fetches. If all of the WUs in your work cache are for projects on which you have restricted the no of cores, cores can sit idle! Use max_concurrent with caution! |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
That was the point I was making ... the pointlessness of having app_config, when the supply chain is just -not- working for, but working against any usage at device level. Loss to WCG, the idle thread picked up some Rosetta. I'll be darned to make amends in my app_config for WCG, when WCG does not supply an even feed of work across the day... "If there is no work for your selected projects..." completely pointless when all projects have already been selected!
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
That's lots of 'pointless' in one post, to include the pointlessness of asking for a solution. ;P
|
||
|
|
Rickjb
Veteran Cruncher Australia Joined: Sep 17, 2006 Post Count: 666 Status: Offline Project Badges:
|
@Sekerob, re. app_config
----------------------------------------App_config & max_concurrent work OK for CEP2 because you can restrict the number of CEP2 WUs in your work cache. In fact, max_concurrent may have been implemented by the BOINC programmers specifically for CEP2 - you're more likely to know that than I am. However, consider the situation where the number of WUs in your cache for a particular sub-project (MCM, say) is not restricted and you restrict the no. of cores permitted to crunch it (MCM). If your allocated (MCM) cores can't crunch fast enough to keep up with the rate of receiving (MCM) WUs, the number of (MCM) WUs in your cache will grow until there's no room for any other work for your other cores to crunch, and cores will sit idle as you have found. You could alleviate the problem by periodically aborting your excess (MCM) WUs to make room for other work for the other cores. I get over the problem by using VMs, but I have to monitor their work caches in case the supply of WUs for the restricted range of sub-projects in each VM runs dry. PS: I tried VirtualBox, but you get what you pay for (it's free). It's a slug that gobbles 10-20% of your crunching output. I use a different VM manager that's very efficient. [Edit 2 times, last edit by Rickjb at Jul 29, 2015 9:17:45 AM] |
||
|
|
KLiK
Master Cruncher Croatia Joined: Nov 13, 2006 Post Count: 3108 Status: Offline Project Badges: 
|
So tell me [again], why do we have app_config?... To see computing time go to waste? With the BOINC governance program, maybe time to rip out the [more control what IS processing] in the next version! so that u do only one CEP2, 'cause of the consequences running 8x CEP2 on HDD... ;) or, like me, to run 3/4 of CPU cores on any project...so I have a little more diversity in results...  |
||
|
|
deltavee
Ace Cruncher Texas Hill Country Joined: Nov 17, 2004 Post Count: 4892 Status: Offline Project Badges: 
|
So tell me [again], why do we have app_config?... To see computing time go to waste? With the BOINC governance program, maybe time to rip out the [more control what IS processing] in the next version! So we can use the <fraction_done_exact/> statement. |
||
|
|
Papa3
Senior Cruncher Joined: Apr 23, 2006 Post Count: 360 Status: Offline Project Badges:
|
PS: I tried VirtualBox, but you get what you pay for (it's free). It's a slug that gobbles 10-20% of your crunching output. I use a different VM manager that's very efficient. What is the URL of this "very efficient" VM manager you so highly recommend? |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
So tell me [again], why do we have app_config?... To see computing time go to waste? With the BOINC governance program, maybe time to rip out the [more control what IS processing] in the next version! So we can use the <fraction_done_exact/> statement. Could have been put into the cc_config.xml as who does not want fraction done exact, and there was already app_info.xml, and that -does- control what you got and which by how much resources was being computed. WCG techs wanted non of that ;? [Edit 1 times, last edit by Former Member at Jul 29, 2015 9:04:34 PM] |
||
|
|
Falconet
Master Cruncher Portugal Joined: Mar 9, 2009 Post Count: 3297 Status: Offline Project Badges: 
|
PS: I tried VirtualBox, but you get what you pay for (it's free). It's a slug that gobbles 10-20% of your crunching output. I use a different VM manager that's very efficient. What is the URL of this "very efficient" VM manager you so highly recommend?Aside from Virtualbox, I can only think of VMware. But there are probably others.  - AMD Ryzen 5 1600AF 6C/12T 3.2 GHz - 85W - AMD Ryzen 5 2500U 4C/8T 2.0 GHz - 28W - AMD Ryzen 7 7730U 8C/16T 3.0 GHz |
||
|
|
Papa3
Senior Cruncher Joined: Apr 23, 2006 Post Count: 360 Status: Offline Project Badges:
|
deltavee - may I ask how much total FA@H run time you currently have? (It must be an awesomely high number!!)
|
||
|
|
|