My dual-G5 Mac was running for several hours on only one core because of a task that got stuck, gs_13_1196035221_6455_0. At the time I spotted it the CPU time was showing as “---” despite having been “Running” for quite a while. Activity Monitor showed astronomy 1.07 using between 60% and 80% of a CPU, considerably less than normal. According to the logs (which I can provide on request) this task started at the same time as another, which finished in the expected 3.5 minutes, but continued to occupy its core for much of the next eight hours (it was paused and resumed a couple of times, apparently cycling normally with other projects). When I suspended the task, the astronomy process continued to show about 7% CPU usage. Resuming it didn’t nudge it out of its rut, but when I quit & relaunched BOINC the task started and finished normally.

My dual-G5 Mac was running for several hours on only one core because of a task that got stuck, gs_13_1196035221_6455_0. At the time I spotted it the CPU time was showing as “---” despite having been “Running” for quite a while. Activity Monitor showed astronomy 1.07 using between 60% and 80% of a CPU, considerably less than normal. According to the logs (which I can provide on request) this task started at the same time as another, which finished in the expected 3.5 minutes, but continued to occupy its core for much of the next eight hours (it was paused and resumed a couple of times, apparently cycling normally with other projects). When I suspended the task, the astronomy process continued to show about 7% CPU usage. Resuming it didn’t nudge it out of its rut, but when I quit & relaunched BOINC the task started and finished normally.

Thats an odd one... there really isn't any spot the code could get into an infinite loop AFAIK. I'll take a look into it.

gs_13_1196035221_6455_0

Thats an odd one... there really isn't any spot the code could get into an infinite loop AFAIK. I'll take a look into it.

Thanks. It occurs to me that the problem may be related to the app’s interaction with BOINC, as the blank CPU time suggests that the computation never got properly under way. I’ve seen other projects’ apps stall similarly, just requiring a fresh launch of BOINC to resume, but usually with some time—sometimes a great deal!—showing on the clock.

I just noticed the task was still “Ready to report”, so I did an Update in case there’s anything of interest in the result’s output file: it should be available now.