| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: Help Conquer Cancer
Thread: 6.56 not resuming properly? |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 18
|
|
| Author |
|
|
cristipurdel
Senior Cruncher Joined: Dec 13, 2008 Post Count: 158 Status: Offline Project Badges: 
|
I have a 6390M card, around 256 GFLOPS, and it processes a WU in around 24 minutes.
----------------------------------------The problem is that BOINC starts running the app after I have been idle for 2 minutes, and it does not continue where it left off, but starts from the beginning. For example, if I am away less than 24 minutes, in 3 hours, no WU is going to get completed because every time it will start allover. I am using 7.0.28 x64 with Cat 12.8. I also see the same behavior on a A8 3500M and on a E-360 6310M laptops, all running 7.0.28 x64 Anyone had the same issue or tried to replicate this? |
||
|
|
Falconet
Master Cruncher Portugal Joined: Mar 9, 2009 Post Count: 3295 Status: Offline Project Badges: 
|
Do you have LAIM(Leave application in memory) active?
----------------------------------------AFAIK, the GPU app has no checkpoints thus it goes back to the beginning when stopped 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 |
||
|
|
cristipurdel
Senior Cruncher Joined: Dec 13, 2008 Post Count: 158 Status: Offline Project Badges:
|
Do you have LAIM(Leave application in memory) active? AFAIK, the GPU app has no checkpoints thus it goes back to the beginning when stopped Leave application in memory options is unchecked Is there no bypass solution for the checkpoints? I did not experience any problems with the POEM app which was taking 4-5 hours on my card. ---------------------------------------- [Edit 2 times, last edit by cristipurdel at Oct 14, 2012 11:25:45 AM] |
||
|
|
Falconet
Master Cruncher Portugal Joined: Mar 9, 2009 Post Count: 3295 Status: Offline Project Badges:
|
Check it. This way the app will stay in memory so even if it is stopped it will resume from the last step. However, if you shutdown the computer, the app will restart from the beginning.
----------------------------------------EDIT: The POEM app may have checkpointing enabled. 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 [Edit 1 times, last edit by Falconet at Oct 14, 2012 11:28:17 AM] |
||
|
|
Crystal Pellet
Veteran Cruncher Joined: May 21, 2008 Post Count: 1322 Status: Offline Project Badges: 
|
Leave application in memory options is unchecked Is there no bypass solution for the checkpoints? I have the same experience. A resumed HCC GPU task always starts from the beginning no matter 'Leave Application In Memory' is checked or not. |
||
|
|
Falconet
Master Cruncher Portugal Joined: Mar 9, 2009 Post Count: 3295 Status: Offline Project Badges:
|
Leave application in memory options is unchecked Is there no bypass solution for the checkpoints? I have the same experience. A resumed HCC GPU task always starts from the beginning no matter 'Leave Application In Memory' is checked or not. In my computer it worked(although I had only one WU since the card produces invalid results) 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 |
||
|
|
Ingleside
Veteran Cruncher Norway Joined: Nov 19, 2005 Post Count: 974 Status: Offline Project Badges: 
|
Check it. This way the app will stay in memory so even if it is stopped it will resume from the last step. GPU-apps is always removed from memory, except during benchmarking.  "I make so many mistakes. But then just think of all the mistakes I don't make, although I might." |
||
|
|
cristipurdel
Senior Cruncher Joined: Dec 13, 2008 Post Count: 158 Status: Offline Project Badges:
|
Leave application in memory options is unchecked Is there no bypass solution for the checkpoints? I have the same experience. A resumed HCC GPU task always starts from the beginning no matter 'Leave Application In Memory' is checked or not. It is not working also in my case. So, there is no solution at all? |
||
|
|
JacobKlein
Cruncher Joined: Aug 21, 2007 Post Count: 21 Status: Offline |
My understanding is that "Leave application in memory" will not leave a GPU application in GPU memory.
----------------------------------------My understanding also is that GPU tasks can checkpoint (my GPUGrid.Net tasks and Poem@Home tasks do!), but currently the WCG HCC tasks do not appear to checkpoint at all (mine appear to always restart from the beginning, when resumed). I hope they implement checkpointing -- I too hate seeing processing time unnecessarily wasted. [Edit 1 times, last edit by JacobKlein at Oct 17, 2012 2:02:35 AM] |
||
|
|
Ingleside
Veteran Cruncher Norway Joined: Nov 19, 2005 Post Count: 974 Status: Offline Project Badges:
|
I hope they implement checkpointing -- I too hate seeing processing time unnecessarily wasted. With fast cards finishing a task in under 2 minutes, it's a waste of resources trying to get any kind of checkpointing. "I make so many mistakes. But then just think of all the mistakes I don't make, although I might." |
||
|
|
|