Message boards : Number crunching : Wu going backwards?
Author | Message |
---|---|
pieface Send message Joined: 20 Sep 05 Posts: 17 Credit: 797,661 RAC: 0 |
I have had a WU (25451) using rosetta 4.77 runing by itself on my P-IV 2 Ghz machine (CPID 1496 - Win XP pro, Boinc 4.45) for about 16-1/2 hours now, and the pct complete is stuck at 1.00 while the estimated time to finish keeps getting bigger (now at around 1618 hrs). Do some of these units loop-de-loop ? |
Divide Overflow Send message Joined: 17 Sep 05 Posts: 82 Credit: 921,382 RAC: 0 |
That certainly sounds like odd behavior for a 4.77 WU. On a 2 Ghz P4 you should be completing WU's in a couple of hours. Perhaps it's time to abort that unit and try another. |
pieface Send message Joined: 20 Sep 05 Posts: 17 Credit: 797,661 RAC: 0 |
Odd... I suspended the wu, shut down boinc manager then brought it back up and unsuspended the unit. It went back to zero cpu time and then stalled at one pct for a bit, but now it seems to be back on track going up in increments of 8.33 pct. |
Polian Send message Joined: 21 Sep 05 Posts: 152 Credit: 10,141,266 RAC: 0 |
I noticed the same thing with my linux box on this result. The only thing that clued me in was the fact that my CPU utilization was down to 75% (4 CPU's, 4 instances running = 100% normally). A ctrl-C on the client followed by restarting it rolled back 2 hours of CPU time on that result and it finished normally. |
[B^S] thierry@home Send message Joined: 17 Sep 05 Posts: 182 Credit: 281,902 RAC: 0 |
If you cut BOINC before the WU is at 8.33%, it will start again from zero because 8.33 is the first checkpoint. |
Message boards :
Number crunching :
Wu going backwards?
©2024 University of Washington
https://www.bakerlab.org