| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: FightAIDS@Home Phase 2
Thread: 24 Hour Deadline too Short |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 112
|
|
| Author |
|
|
NixChix
Veteran Cruncher United States Joined: Apr 29, 2007 Post Count: 1187 Status: Offline Project Badges: 
|
How do I set which projects I want a specific workstation to work on? Do I need to create two profiles... One that is exclusively FAH2 and another profile that works on other sub-projects other than FAH2? My issue is the rotation of sub-projects since BOINC seems to only work on one subproject. If the CPU is not maxed out, then it will add more duplicate sub-projects. I too don't like the one day deadline. Rather than 1 day workbuffer, I resorted to having 0 day buffer and let it download tasks as needed. I have good network connection so not adding tasks beforehand should give me more time for FAH2? That is almost what I do sunk818, except that I also have HSTB along with FAHB. This profile is only set on my newest computer, which is the only one that can handle the 1 day deadline. It seems to keep full. Seeing no jobs queued is unnerving, but I have only very occasionally running at less than full capacity.Cheers   |
||
|
|
Aurum
Master Cruncher The Great Basin Joined: Dec 24, 2017 Post Count: 2391 Status: Offline Project Badges: 
|
So they want to ramp up FAH2. Will they make the deadline reasonable???
----------------------------------------  ...KRI please cancel all shadow-banning |
||
|
|
Ajihad
Cruncher Joined: Apr 29, 2008 Post Count: 8 Status: Offline Project Badges: 
|
Opted out of the project after I realized that many of the tasks even aren't completed on time when I have my BOINC manager run 24/7. Sorry folks, but unless you change the deadlines for that project, I'll rather contribute my computing power to other important causes :(
|
||
|
|
Macroman
Advanced Cruncher Joined: Jun 4, 2005 Post Count: 112 Status: Offline Project Badges: 
|
I want to add my voice to those saying that the 24 hour deadline is too short. In my view this is a ploy by FAH2 to grab computing power away from other projects.
I wish that the BOINC application supported an option to allow the user profile to select how many slots are available to a project similar to what was done for the Clean Energy project. In this way users could continue to support FAH2 without turning off the other projects to the degree they choose to. |
||
|
|
ca05065
Senior Cruncher Joined: Dec 4, 2007 Post Count: 328 Status: Offline Project Badges: 
|
I believe there is already a limit of FAHB sent to a PC of one per thread.
It is possible to restrict the number of FAHB running using the app_config.xml file in the folder:C:\ProgramData\BOINC\projects\www.worldcommunitygrid.org. Here is my app_config.xml file: <app_config> <app> <name>cep2</name> <max_concurrent>2</max_concurrent> </app> <app> <name>hst1</name> <max_concurrent>8</max_concurrent> <fraction_done_exact/> </app> <app> <name>zika</name> <max_concurrent>8</max_concurrent> <fraction_done_exact/> </app> <app> <name>oet1</name> <max_concurrent>8</max_concurrent> <fraction_done_exact/> </app> <app> <name>mcm1</name> <max_concurrent>8</max_concurrent> <fraction_done_exact/> </app> <app> <name>fahv</name> <max_concurrent>8</max_concurrent> <fraction_done_exact/> </app> <app> <name>scc1</name> <max_concurrent>8</max_concurrent> <fraction_done_exact/> </app> <app> <name>mip1</name> <max_concurrent>8</max_concurrent> <fraction_done_exact/> </app> <app> <name>fahb</name> <max_concurrent>6</max_concurrent> <fraction_done_exact/> </app> </app_config> It then needs experimentation of the max_concurrent parameter and work buffer length in BOINC to achieve a satisfactory return period for non FAHB work units. With the above file I used a 0.5 day work buffer to achieve a 2 day return period for non FAHB work units. |
||
|
|
X1900AIW
Cruncher Joined: Feb 3, 2008 Post Count: 25 Status: Offline |
Challenging ... short deadline, +20% higher energy consumption (AVX?), no cache of workunits.
----------------------------------------Switched to FAH2 last night, 8C/16T with SMT, CPU time ~1 hour, some with 5 hours, new ones with 20% progress rate per hour. Hope there is constantly flow of new workunits. So far I can´t complain about too short deadline. [Edit 1 times, last edit by X1900AIW at Jun 7, 2018 7:00:36 AM] |
||
|
|
TheBizii
Advanced Cruncher Slovenia Joined: Jan 4, 2016 Post Count: 51 Status: Offline Project Badges: 
|
I have realized that FAH2_2000... WUs are taking around 4 hours on my computer, while FAH_1000... WUs take nearly an hour. While 1 day is just enough for my working units to get processed, I'd still recommend researchers increasing the deadline in order to avoid replicated WUs. I noticed my computer receives a lot of replicated WUs.
----------------------------------------[Edit 1 times, last edit by DeliciousShoesHD at Jun 7, 2018 11:47:10 AM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
It's not really a big thing if volunteering without a buffer on 24/7, but if running part time and deselect FAH2, then have the "If there is no work available for the project(s) I have selected above, please send me work from another project.", you still get them, too frequently. Even selecting all other projects and leave FAH2 off, and still there's log entries saying there was no work for all 7 and this fast lane project fills in. How is that possible, all 7 others no work? Have unticked the "if there is no work". Idling a thread is less waste then unnecessary duplicate processing after overnighting with the computer off.
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
This is an explanation fron Sgt.Joe at the beginning of this thread
----------------------------------------I think the answer to the 24 hour deadline is that the work units have such a large number of parts. For instance this one: FAH2_ 001464_ avx17629-0_ 000008_ 000077_ 028 _ 0-- The "028" means it is the 28th part of this work unit. I seem to recall somewhere that these work units may have more than 100 parts. If that were the case, rather than taking 100 days to complete the entire work unit it could take up to 1000 days ( about 2.75 years) to complete the sequence for that work unit if the deadline were 10 days. I would think the scientists would like their results of the completed work units in a timelier fashion. I believe there is some backend work about putting all the parts of the work unit together to get the final result. This implies that there could be some kind of storage constraint on holding the parts of the work unit until its completion. If I am off base on this explanation, someone with more information should feel free to correct me. Cheers [Edit 1 times, last edit by Former Member at Jun 11, 2018 11:29:06 AM] |
||
|
|
Aurum
Master Cruncher The Great Basin Joined: Dec 24, 2017 Post Count: 2391 Status: Offline Project Badges:
|
So the logic is, the more donors that blacklist our short fuse the more work we get done.
----------------------------------------...KRI please cancel all shadow-banning |
||
|
|
|