*ugh*

I just lost (again) another ~500h CPU time, as I witnessed one of my machines suddenly showing its progress at 0.x % again...
Model : http://climateapps2.oucs.ox.ac.uk/cpdnboinc/result.php?resultid=812546

This wasn\'t the first time either, quite a number of Models took a 2nd start for no reason, causing enourmous amounts of CPU time to go down to the dumps :(

(BOINC V5.2.13)

Well, what can I say. That\'s it, I\'ve now set all Systems (that still do have Models \"attempting\" to finish) to deplete work; after they\'re done I\'ll unattach from CPDN.

Losing a bit here and there due to Software Bugs is not problematic, but CPDN has so far wasted no less than ten thousands of hours of CPU time on me.
I simply can\'t afford that to happen, in these numbers we begin talking about serious amounts of money keeping those machines in operation, which I won\'t accept seeing wasted to poor Software.

---

Scientific Network : 44800 MHz - 77824 MB - 1970 GB

Overclocked machines?

With all that trouble, no backups?

You\'ve done a lot of work for the project; thanks for that. Sorry to lose you.

---

"We have met the enemy and he is us." -- Pogo
Greetings from coastal Washington state, the scenic US Pacific Northwest.

I hate to leave like that as well, since I always liked CPDN alot. Everybody around here has always been very friendly and helpful.

Backups on 24 machines \"on the fly\" are impossible without interfering with the running BOINC Clients (which reportedly can hate access to their Directory while running) are impossible I\'m afraid. And technically, it shouldn\'t be needed (except to protect against Harddrive failure or alike)

None of the machines are overclocked, all on stable performance settings and work 100% fine on all other Projects. With all the reports about CPDN bugs (and recent BBC CPDN Bugs as well), I\'m afraid it\'s in the Software with no solution I could go for to fix it.

If it wasn\'t for the tremendous amounts of lost CPU time, I would ignore it but with the inherent Model runtimes, it\'s just too much to justify :(

PS.
I\'ll periodically check the Progress around CPDN Client development and sure will be back, as soon as the Problems are solved. It\'s too good of a project to let go completely...

---

Scientific Network : 44800 MHz - 77824 MB - 1970 GB

Something like that just happened to me.... I had just reattached, downloaded a BBC model v. 5.08, had got over 1% and suddenly this morning it went back to 0.64%. Pretty disappointed, I must say.... ;( No overclocking whatsoever, running on an HT machine. When I finish this WU (still 2848 hours to go!) unless it fails first, I think I will detach again..... as much as I like CPDN... :(

---

<br>***********************************<br>
WinXP Home, P4 3.2GHz HT, 512Mb RAM. Running SETI, CPDN, Predictor, LHC, Einstein, Orbit, Lattice and Folding

I think someone wrote an automated backup for the BBC model (RRodney on the BBC boards), you could try that?

http://bbc.cpdn.org/forum_thread.php?id=2748

---

I'm a volunteer and my views are my own.
News and Announcements and FAQ