Arghhhh.......lost another -bigadv :(

zero2dash said:
S....at this point I could care less about going from 3.6 > 3.8 or 3.9, both of my cpu's are duds, they won't do 4.0 even with 1.4 Vcc and Vtt, and the TPF difference for the extra 200-300mhz (3.6 > 3.8 or 3.9) is practically nothing, just a 100 PPD or so.
Click to expand...
Is that with A3 or -bigadv?
 
APOLLO said:
Is that with A3 or -bigadv?
Click to expand...

-bigadv

Basically used this FF post and scrolled through to see i7 920 TPF with those clocks. Largely anything 3.6-3.9 is pretty similar; 4+ the gap widens a little more.

Code: 
Username           Avg Time  Completion       Bonus       PPD  Proj# CPU Model  Ghz   CPU Core FAH GPU RAM OS                TGhz
tomc001              0:33:15   55:25:00       32,510    25,080 2683  i7-920      3.60 1   4    8   0   3.9 Suse 11.2         14.40
foo_fight            0:33:16   55:26:40       32,490    25,060 2681  i7-920      3.70 1   4    8   0   6   2.6.24-24         14.80
foo_fight            0:34:58   58:16:40       31,070    23,260 2681  i7-920      3.70 1   4    8   0   6   2.6.24-24         14.80
leexgx               0:34:40   57:46:40       31,310    23,560 2681  i7-920      3.80 1   4    8   0   5   Frednotfred VM    15.20
JackOfAll            0:35:20   58:53:20       30,770    22,890 2681  i7-920      3.80 1   4    8   0   6   2.6.29.6          15.20
leexgx               0:37:00   61:40:00       29,490    21,360 2681  i7-920      3.80 1   4    8   4   5   Frednotfred VM    15.20
Spongebob25          0:36:52   61:26:40       29,590    21,480 2681  i7-920      3.90 1   4    8   0   6   2.6.28-14         15.60
linuxfah             0:29:45   49:35:00       35,820    29,630 2683  i7-920      4.00 1   4    8   0   6   2.6.31.2          16.00
Bigstan              0:32:34   54:16:40       33,110    25,870 2681  i7-920      4.00 1   4    8   2   6   Frednotfred VM    16.00
Spongebob25          0:33:17   55:28:20       32,480    25,040 2683  i7-920      4.00 1   4    8   0   6   2.6.28-15         16.00
Spongebob25          0:29:54   49:50:00       35,670    29,410 2683  i7-920      4.06 1   4    8   0   6   2.6.28-15         16.24
linuxfah             0:29:39   49:25:00       35,920    29,780 2681  i7-920      4.15 1   4    8   0   6   2.6.31            16.60
DriveEuro            0:33:57   56:35:00       31,910    24,310 2681  i7-920      4.15 1   4    8   1   6   2.6.30            16.60
Bigstan              0:29:25   49:01:40       36,170    30,140 2681  i7-920      4.20 1   4    8   0   6   Frednotfred VM    16.80
DonMarkoni           0:29:32   49:13:20       36,040    29,960 2683  i7-920      4.20 1   4    8   0   3.9 2.6.31.5          16.80
DonMarkoni           0:32:19   53:51:40       33,340    26,180 2683  i7-920      4.20 1   4    8   2   3.9 2.6.31.5          16.80
Tasmac               0:28:37   47:41:40       37,020    31,410 2683  i7-920      4.30 1   4    8   0   3   2.6.28-10         17.20
Fe[R]gity            0:27:12   45:20:00       38,630    33,900 2683  i7-920      4.45 1   4    8   0   6   2.6.33-020633

Obviously difference in os accounts for some point differences as well but I take from that list that really 3.6-3.9 is pretty close in TPF/PPD. A hundred here or there, not much more.
Actually after my main finally did a couple frames, I'm getting a 36:59 TPF on it and a 32:32 TPF on my dedicated...due to the 2 260's I'm sure (on the main). I may shoot that back up to my stable 3.90 on it after all and just see how the heat goes (with summer coming). I guess what I've learned from this 'experiment' is that on a dedicated, 3.6>3.9 doesn't have much difference but factoring in gpu clients it's a totally different story. :eek: (I did get the main stable at 3.9; the dedicated boxen has a real picky cpu that craves more juice.)
 
Last edited:
zero2dash said:
-bigadv

Basically used this FF post and scrolled through to see i7 920 TPF with those clocks. Largely anything 3.6-3.9 is pretty similar; 4+ the gap widens a little more.

Obviously difference in os accounts for some point differences as well but I take from that list that really 3.6-3.9 is pretty close in TPF/PPD. A hundred here or there, not much more.
Click to expand...
That's interesting. On my S771 machines, 300MHz translates to about 2000 PPD but that's going from 2.6GHz to 2.9GHz, so the percentage of frequency increase is greater.
 
Code: 
[02:19:40] Completed 142500 out of 250000 steps (57%) 
[02:58:04] Completed 145000 out of 250000 steps (58%) 
[03:03:24] - Autosending finished units... [May 12 03:03:24 UTC] 
[03:03:24] Trying to send all finished work units 
[03:03:24] + No unsent completed units remaining. 
[03:03:24] - Autosend completed 
[03:36:23] Completed 147500 out of 250000 steps (59%) 
[03:55:32] 
[03:55:32] Folding@home Core Shutdown: INTERRUPTED 
[03:55:40] CoreStatus = FF (255) 
[03:55:40] Sending work to server 
[03:55:40] Project: 2681 (Run 3, Clone 2, Gen 100) 
[03:55:40] - Error: Could not get length of results file work/wuresults_02.dat 
[03:55:40] - Error: Could not read unit 02 file. Removing from queue. 
[03:55:40] Trying to send all finished work units 
[03:55:40] + No unsent completed units remaining. 
[03:55:40] - Preparing to get new work unit... 
[03:55:40] Cleaning up work directory 
[03:55:45] + Attempting to get work packet 
[03:55:45] Passkey found 
[03:55:45] - Will indicate memory of 4458 MB 
[03:55:45] - Connecting to assignment server 
[03:55:45] Connecting to http://assign.stanford.edu:8080/ 
[03:55:46] - Successful: assigned to (171.67.108.22).

Anyone know why this would happen? It just stopped the bigadv and got a new one starting from 0% again. The VM didn't do anything weird in it's video output, just wrote this in the log file. The computer wasn't shutdown or anything and the VM was just running minimized and I didn't touch it. The CPU has been stable for about 2 months 24/7 with no other errors so how can I prevent this?
 
You really can't. That is the nature of bigadv work units. Some of them just don't work right all the way through. They are more of a beta-type of thing.
 
I've had a few of those 'CoreStatus = FF' crash and flushes; nothing you can really do. :(
 
Well at least it's not my fault. It's the CC so it would be nice to be producing points. Should have done A3's. Missing a whole bigadv kind of suck cause it's over in 3 days.
 
Worst I've ever had crash and flush was like 96%...I think they heard me on the moon dropping f-bombs about it. :D
 
zero2dash said:
Worst I've ever had crash and flush was like 96%...I think they heard me on the moon dropping f-bombs about it. :D
Click to expand...

Mine crashing after 100% a couple pages back in this thread is my worse...don't think you could get much worse that that one. The same machine has folded and uploaded one or two bigadv since then without issue. It just happens.
 
I have had similar problems recently, I will shutdown the VMWare using a virtual machine version that includes a backup even... and it seems that lately I have a good chance of losing the Work Unit. For the past 2 months I could shut a WU down 5 - 6 times if needed with no problems at all, and now just recently I have lost 3 WUs in the past 2 weeks (one of them at 96%). I may go back to the Suspend/Resume method instead of "shutdown -h now" even though I have the backup. My rig is also my gaming machine, hence the need to shut it down often.

Here was the latest failure from tonight at 21%:
[01:54:55] - Expanded 30331003 -> 159726549 (decompressed 101.8 percent)
[01:54:56] Called DecompressByteArray: compressed_data_size=30331003 data_size=159726549, decompressed_data_size=159726549 diff=0
[01:54:57] - Digital signature verified
[01:54:57]
[01:54:57] Project: 2682 (Run 7, Clone 4, Gen 16)
[01:54:57]
[01:54:57] Entering M.D.
[01:55:03] Using Gromacs checkpoints
[01:55:22] fcCheckPointResume: file hashes different -- aborting.
[01:55:22] s different -- aborting.
[01:55:30] CoreStatus = FF (255)
[01:55:30] Sending work to server
[01:55:30] Project: 2682 (Run 7, Clone 4, Gen 16)
[01:55:30] - Error: Could not get length of results file work/wuresults_08.dat
[01:55:30] - Error: Could not read unit 08 file. Removing from queue.
 
I've never used anything but suspend/resume if I need to take the VMs offline. So far I have not lost a WU due to suspending a VM.
 
I now only use Suspend / Resume and haven't lost one yet.

Just curious though, for those of you in Native Linux if you need to shut down, what do you do?
 
I would assume control+c which you can also use in the VM.
I use control+c then power the VM down - no problems.
 
Well damn............I just ended up losing a 2684 bigadv at 92% on folding1(the Bloomfield).:mad:

I had yet another power outage today and this is the 3rd power outage i've had in 4 days. I live inside the city of Cincinnati, so this is obviously inexcusable. My systems were able to recover from the previous ones as i've been around to restart the client after they rebooted, but this time it was lost. I had to reinstall the client as I kept getting a core client error and thus lost 2 more bigadv's in a row. Reinstalling it did the trick, as I suspect something must have really gotten corrupted.

Luckily folding2 was able to recover with its bigadv at 96%, and thus was able to turn it in.

Both systems are working on 2684's again :(
 
You must log in or register to reply here.
Back
Top