I've had two of these beta tasks so far, one is still running, and the other has had an error and failed.

Furthermore, the currently running task is taking an exceptionally long time to finish on a GTX 670. It has been running for about 19 hours and is only 32% done, with an estimated 27 hours left. I was wondering if this is normal, although I presume not. Also, GPU load isn't that low, it is roughly 77% for the GPU working on this task.


Edit:The task that I previously mentioned that was still running, reverted to 0% progress when I restarted my computer. I restarted because I believed the driver (314.07) crashed while I was gone. Sorry, but I had to abort that WU, it was simply taking too long.

Link to failed WU:

http://www.gpugrid.net/workunit.php?wuid=4201831

Link to aborted WU mentioned in the edit:

http://www.gpugrid.net/workunit.php?wuid=4201826

Same here. I got this one.

http://http://www.gpugrid.net/workunit.php?wuid=4201816

I had to restart my PC at 77% and now it is reverted to 0%.
Seems that checkpointing doesnt work correctly.

What should i do? Abort it? This unit looks buggy!

I also appear to have a possibly-buggy task.

I'm working on ACEMD beta version v6.49 (cuda42) task: replica49-NOELIA_TRYPtrial-0-1-RND1465_3 here http://www.gpugrid.net/workunit.php?wuid=4201857 on Windows 8 x64 using GeForce 660 TI, and it says it's been working for about 33 hours on it, yet, if I suspend BOINC and then resume BOINC, the progress % goes back to 0. Then, when it restarts, remaining time increases while it is running.

I think something is wrong with the task or application, and I'm considering aborting it (since it seems like it's not going to finish normally).

Regards,
Jacob

I aborted it.
It did not appear that it would checkpoint nor complete in a timely fashion during the time slots that my machine is available.

You might want to look into that task, to see what is causing its problems.