95 hours for a WU? They must be 'aving a laff!

[Snapshot]

Finished my third WU this morning and that core was rewarded with another 2126 that will take 95 hours to complete! I've killed that service and will install Linux and the SMP client as soon as the other core finishes - a 93 hour 2124 with 'only' 54 hours left. I'm just grateful one of the other completed WUs was replaced by another 1499 screaming along at almost 700PPD but what will I get when that one finishes?

Jonathan

 

[Snapshot]

[11:18:47] Project: 2126 (Run 86, Clone 71, Gen 3)
[11:18:47]
[11:18:47] Assembly optimizations on if available.
[11:18:47] Entering M.D.
[11:18:53] Protein: p2126_villin_sigd3
[11:18:53]
[11:18:54] Writing local files
[11:18:54] Using table 3
[11:18:54] Extra SSE boost OK.
[11:18:54] Writing local files
[11:18:54] Completed 0 out of 50000000 steps (0)
[11:49:47] Timered checkpoint triggered.
[12:16:14] Writing local files
[12:16:14] Completed 500000 out of 50000000 steps (1)
[12:19:51] Service stop request received.
[12:19:51] ***** Got a SIGTERM signal (2)
[12:19:51] Killing all core threads

Looks okay to me and is roughly the same for all the others I'm running, fast and slow. CPU is an i820 so not too down on power.

 

[Snapshot]

No you're right, they are horrible. 38 mins/frame for 147 ppd on a 3ghz opty (possibly a bit low due to gaming).

You have to live with it really and fold whatever they give you. "Fold for the science". When people start doing things to try and get new wu's it messes up the whole system.

You're right, of course. We should remember why we're really doing this although I suspect many are just folding for the competitive element. I'll be leaving most of the processes running but I'll stop the i820 when it finishes its current run and try Linux on it. That PC is called Test and is set up so I can swap HDDs easily.

Jonathan