Project: 10079 (Run 82, Clone 0, Gen 19)

[Hammerhead]

Project: 10079 (Run 82, Clone 0, Gen 19) was taking an excessive amount of CPU time. About 2 hours per % step. Thought it would finish just before the final deadline but aborted near the end and then the core deleted it on restart because it was past final deadline. The core was running near 100% CPU all the time so it wasn't that it wasn't getting enough CU time.

Code: Select all

[07:01:40] *------------------------------*
[07:01:40] Folding@Home Gromacs GB Core
[07:01:40] Version 2.27 (Dec. 15, 2010)
[07:01:40]
[07:01:40] Preparing to commence simulation
[07:01:40] - Assembly optimizations manually forced on.
[07:01:40] - Not checking prior termination.
[07:01:40] - Expanded 39313 -> 201448 (decompressed 512.4 percent)
[07:01:40] Called DecompressByteArray: compressed_data_size=39313 data_size=2014
48, decompressed_data_size=201448 diff=0
[07:01:40] - Digital signature verified
[07:01:40]
[07:01:40] Project: 10079 (Run 82, Clone 0, Gen 19)
[07:01:40]
[07:01:40] Assembly optimizations on if available.
[07:01:40] Entering M.D.
[07:01:46] Mapping NT from 1 to 1
[07:01:46] Completed 0 out of 10000000 steps  (0%)
[08:45:58] - Autosending finished units... [May 27 08:45:58 UTC]
[08:45:58] Trying to send all finished work units
[08:45:58] + No unsent completed units remaining.
[08:45:58] - Autosend completed
[09:11:26] Completed 100000 out of 10000000 steps  (1%)
[11:23:48] Completed 200000 out of 10000000 steps  (2%)
[13:28:49] Completed 300000 out of 10000000 steps  (3%)
.
.
.
[10:20:57] Completed 9100000 out of 10000000 steps  (91%)
[10:20:57] Unit 5's deadline (June 4 07:01) has passed.
[10:20:57] Going to interrupt core and move on to next unit...
[10:21:01] CoreStatus = 3 (3)
[10:21:01] Client-core communications error: ERROR 0x3
[10:21:01] This is a sign of more serious problems, shutting down.

Next WU is for the same project and it seems to be running OK.

Code: Select all

[10:26:02] Project: 10079 (Run 312, Clone 0, Gen 24)
[10:26:02]
[10:26:02] Assembly optimizations on if available.
[10:26:02] Entering M.D.
[10:26:08] Mapping NT from 1 to 1
[10:26:09] Completed 0 out of 10000000 steps  (0%)
[10:54:08] Completed 100000 out of 10000000 steps  (1%)
[11:22:01] Completed 200000 out of 10000000 steps  (2%)

[bruce]

This is an interesting situation which I do not understand.

In your case, you were unable to complete the WU by the final deadline (8.0 day) so it was discarded. After the Preferred Deadline (4.0 days) it was reassigned. This may have happened to several people but the records that I can see simply have no information about that sort of thing.

The secondary effect, though, is that more than one WU is distributed until somebody actually completes it. In this case, 4 people were able to complete it for credit.

Days taken to complete WU: 2.19
Days taken to complete WU: 2.54
Days taken to complete WU: 2.64
Days taken to complete WU: 4.67

I really curious why there's such a wide variation in completion times. Is it just hardware?

[Hammerhead]

I don't think it's hardware which in this case is a Dell Dimension 9100 w/ dual core Pentium D @ 3 GHZ. I kept checking with Task Manager and the a4 core was always getting near 100% CPU. Again, the current WU is for project 10079 (Run 312, Clone 0, Gen 24) as well and time for % steps is about 27 minutes, not 2 hours. I didn't reboot the PC between WUs so I can't think of anything. I always start the client with -forceasm, don't know if that matters. Maybe just file it as a fluke.

[John_Weatherman]

A combination of hardware and how long the machines were on, as these are single core WUs, I would suspect.

[Hammerhead]

I keep that PC on 24/7 so it wasn't shut down once during the entire 8 day run.