This result won\'t finish. It\'s been around 94% for days. You can see that the last trickle noted in the database page is from 4 March, and up until then it had been trickling about once every 20hours or so. Since then, there have been trickle messages appearing in the messages tab of the BOINC manager on that machine, but no further recorded trickles in the database or credits awarded. The percentage complete is changing, but I\'ve noticed it going down as well as up, even when I haven\'t stopped BOINC. At the moment I have suspended all other projects on that host to try to avoid interruptions to the Climate Prediction application, but that has not helped.

What has happened to my result? How can I get it to finish? It seems a shame to abort it and waste all that effort.

Neil

---

There have been some faulty units in the past that go into a loop. This is what sounds like happened to yours. This is not a problem with your computer.

Sometimes a full stopping of BOINC, reboot the computer clean, and the restart will fix it. Many times it does not.

If it does not, abort it. The work you did is GOOD, and the team is happy that it\'s made it as far as it has. A lot of information is already there for them in the work you did. It is NOT a waste.

So, give the reboot trick a try. Make sure you do the full BOINC shutdown first, as this sometimes makes the difference.

Good luck, and hope this doesn\'t deter you from doing another unit.

---

In the README about avoiding crashes (link to it through my signature), item #4 is about looping models.

A few members have got a model through the loop by transferring a backup taken before the looping started from an AMD machine to an Intel, or vice-versa. If you can\'t try that, just abort it.

Best of luck with your next model.

---

Cpdn news