Project: 2662 (Run 1, Clone 52, Gen 28) – 5500000 steps

[Phantom]

Seems like I got another bogus WU... Normally, Project 2662 WUs for me have only 250000 steps -- but for some reason, my system thinks this WU has 5500000 steps! (22x increase!)

This WU was running on a MacPro (2 x 2.66GHz Intel Core 2 Duo with 8GB memory) and was projected to complete around Sept. 14th so I temporarily stopped it... Way too late for the due date of Aug. 30th.

What do you recommend I do with this WU?

Code: Select all

[13:55:52] - Connecting to assignment server
[13:55:52] Connecting to http://assign.stanford.edu:8080/
[13:55:52] Posted data.
[13:55:52] Initial: 40AB; - Successful: assigned to (171.64.65.56).
[13:55:52] + News From Folding@Home: Welcome to Folding@Home
[13:55:52] Loaded queue successfully.
[13:55:52] Connecting to http://171.64.65.56:8080/
[13:55:58] Posted data.
[13:55:58] Initial: 0000; - Receiving payload (expected size: 5004103)
[13:56:02] - Downloaded at ~1221 kB/s
[13:56:02] - Averaged speed for that direction ~1093 kB/s
[13:56:02] + Received work.
[13:56:02] Trying to send all finished work units
[13:56:02] + No unsent completed units remaining.
[13:56:02] + Closed connections
[13:56:02] 
[13:56:02] + Processing work unit
[13:56:02] Core required: FahCore_a2.exe
[13:56:02] Core found.
[13:56:02] Working on Unit 01 [August 27 13:56:02]
[13:56:02] + Working ...
[13:56:02] - Calling './mpiexec -np 4 -host 127.0.0.1 ./FahCore_a2.exe -dir work/ -suffix 01 -checkpoint 15 -forceasm -verbose -lifeline 10889 -version 602'

[13:56:02] 
[13:56:02] *------------------------------*
[13:56:02] Folding@Home Gromacs SMP Core
[13:56:02] Version 2.01 (Wed Jul 16 08:26:53 PDT 2008)
[13:56:02] 
[13:56:02] Preparing to commence simulation
[13:56:02] - Ensuring status. Please wait.
[13:56:12] - Assembly optimizations manually forced on.
[13:56:12] - Not checking prior termination.
[13:56:13] - Expanded 5003591 -> 24742709 (decompressed 494.4 percent)
[13:56:13] Called DecompressByteArray: compressed_data_size=5003591 data_size=24742709, decompressed_data_size=24742709 diff=0
[13:56:14] - Digital signature verified
[13:56:14] 
[13:56:14] Project: 2662 (Run 1, Clone 52, Gen 28)
[13:56:14] 
[13:56:14] Assembly optimizations on if available.
[13:56:14] Entering M.D.
[13:56:21] Node 2 initialized
[15:35:47] - Autosending finished units...
[15:35:47] Trying to send all finished work units
[15:35:47] + No unsent completed units remaining.
[15:35:47] - Autosend completed
[18:19:42] Completed 55000 out of 5500000 steps  (1%)

[7im]

5500000 total steps, but it jumps up at 55000 steps each percentage completed. That's the same as counting to 100 by 1s, except the WU is covering a log more ground in each jump.

It's not a big deal. Keep folding. If the ETA doesn't level out to a more acceptable level after a few frames, and you're sure it won't finish in time, then you can delete it and move on.

[Phantom]

5500000 total steps, but it jumps up at 55000 steps each percentage completed. That's the same as counting to 100 by 1s, except the WU is covering a log more ground in each jump.

It's not a big deal. Keep folding. If the ETA doesn't level out to a more acceptable level after a few frames, and you're sure it won't finish in time, then you can delete it and move on.

For some reason unknown to me, the WU was created with too many steps for the deadline assigned. Yes, the WU seems to process in 1% increments; however, due to it's abnormally large size, it will not complete in time. I have seen (and reported) a prior Project 2662 with 4000001 steps (16x the normal size) and was told by kasson that there was something wrong with that WU. I am merely reporting the presence of another bad WU and have stopped burning CPU cycles on it since there is no possible way to complete the 5500000 total steps on my hardware before the deadline of Aug. 30th.

I will put the CPU cycles to use on another WU immediately.

Thanks.

[kasson]

Yes--unfortunately the best solution is to delete this one and move on. We're working on a solution, but it will take a little more time on our end.

[kasson]

We've stopped this one from re-assigning and are in the process of implementing a solution to prevent this problem from recurring. There shouldn't be any *new* WU's created in projects 2662 or 2668 with this problem, but there may be some existing WU's that are too long (either issued or waiting on our servers). We still need to track these down.