| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Retired Forums
Forum: Member-to-Member Support [Read Only]
Thread: Why does it keep on resetting itself?I |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 25
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I downloaded this a few hours ago, its been running for approx. 3 hours or so. It has gone up to 2%, then jumped back to 0% and reset CPU time to 0, then up to 4%, and the same happened again. I've counted it doing it three times in total.
Is this a normal thing for it to do or is there a problem? Thanks Russell |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I've seen the same thing on one of my three systems. I may reach as high as 2 or 3% before dropping back to zero. I have it running on two other systems fine, but this one appears to be a lost cause. I'm not restarting the program and I've seen this happening by just glancing at the screen saver every hour or two.
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I'm seeing the same thing - I have it installed on 4 systems and 3 are running fine. The 4th one keeps resetting.
The interesting thing is that this 4th system was able to finish the first task just fine and shipped the results during the night. Now the second task is getting to about 3% and resetting. I wonder if there is something wrong with this latest task? |
|||
|
||||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
lo peeps I have copied this from another post, It's from Rick Alther World Community Grid Application Developer. Hope this helps you understand what's going on with those units. I also had incidents as you have mentioned and this made sense to me.
 A "work unit" consists of various data including the protein sequence. The longer the sequence the "harder" it is to predict the fold. Longer sequence proteins take longer to process (sometimes much longer). Due to the nature of the work, it's impossible to predict how long any specific work unit may take. i.e. it's non-determinsitic. That's really the beauty of putting this on a grid: it's a difficult computation problem. Sometimes a work unit may not "converge" as we call it. This means it just hasn't found any way to fold this specific protein in a given amount of time and gives up. When it determines it can't converge it just stops working on it, sends it back (telling us it didn't converge) and you get a new work unit. However, it will try very, very hard before it gives up. This would explain why you saw it processing for probably 4 hours and still be at 0.0%...then suddenly you saw it go to 0.4% and the CPU time went down. Points-wise, you should still get credit for the effort though. Have some faith in that you really are helping to promote our understanding of the human proteome. If this was easy to do, we wouldn't have put it on the grid! ---------------------------------------- Rick Alther World Community Grid Application Developer |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
When it determines it can't converge it just stops working on it, sends it back (telling us it didn't converge) and you get a new work unit. However, it will try very, very hard before it gives up. Hello from Russia! I have exaclty the same problem What I dont understand in the quoted phrase is: Does it sends back the unit each time it has stopped working on it? (meaning that I just overlooked the send-get procedure and the unit I see now, at 0% instead of 2%, is new) If the unit is not sent back but instead is re-run, ie "tried very hard", then my question is: How many times the unit is resetted before the program gives up and gets another one? Overall, this statement is very vague. Best wishes, Artyom. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
The same thing has been happening to my system for the past 9 hours. The highest I've gotten is 7% before it resets, most of the time 2-3%. I hope this task gives up soon and I get another one we can see through to the end.
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
has it ever happened before to anyone? because it seems like its happening quite frequently all of the sudden.. though maybe its just more noticable now with all kinds of new members for it to happen to...
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hello again from Russia.
----------------------------------------I installed UD monitor to track and cache the units. And: when the unit has resetted from 2.5% to 0%, the UD monitor showed that WU number changed from WU 13088 to WU 15044 at the same time, firewall logs showed that a big chunk was downloaded to my pc (over 1 mbyte). at the time as was not doing anything on the net. Hence I conclude that these kinds of units are simply "ditched" and the new ones are dowloaded. Therefore, two questions arise: 1. Why the Agent program does not tell the user that it does but leaves him to guess and burrow through the logs? 2. Is the time spent on the "unlucky" WUs added to the user's stats, and so do his "points" get an increase too? P.S. Take into account that non-us and non-europe users might be paying for traffic per megabyte, as I am. Regards, Artyom, Ekaterinburg. [Edit 3 times, last edit by Former Member at Nov 21, 2004 5:12:32 AM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I can now confirm that the WUs are changing, as well. In the UD Monitor log I see that the WUs are being aborted and then new ones being downloaded and executed. Seems like we're hitting a batch of proteins where roseta can't determine the folds. At least I know it isn't with my system now.
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Therefore, two questions arise: 2. Is the time spent on the "unlucky" WUs added to the user's stats, and so do his "points" get an increase too? [color] If I'm right,no, I don't think you are credited for the unsucessful WU. At least I was'nt!! Wasted so many hours.. |
||
|
|
|