• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Bad Fermi Work Units?

J

jebo_4jc

[H]ard|DCer of the Month - April 2011
2FA
Joined
Apr 8, 2005
Messages
14,572
I've been folding 6800s just fine for the entire time they've been in general distribution. Now, both of my GTX470s went down, and I'm afraid what's going to happen to my GTX460 once it finishes its unit.

The WU in question:

6800 (Run 1418, Clone 1, Gen 21)
6800 (Run 4473, Clone 0, Gen 31)

Anybody else? Or is my system somehow fubar'ed.
 
Me but different WU.

6800 (Run 4623, 0, 20)

My MSI GTS450 has be bailing out (nans) on this 6800 for the last day. Been running great for last few weeks. No problems. So I reinstalled drivers, using stock clocks. It makes about 10 checkpoints then pukes. I haven't had the time to dig into it, hopefully later tonight.
 
Hm. Mine may have been a fluke on my end. I rebooted and haven't seen any problems since.
 
jebo_4jc said:
Hm. Mine may have been a fluke on my end. I rebooted and haven't seen any problems since.
Click to expand...
It could just have been a string of bad WUs or the same corrupt units being resent to your clients over and over. The trick to stopping the resending of the same WUs is to change the machine ID. Sometimes, there's nothing you can do except reboot the machine. This may have been your situation. It has occurred to my GPU clients in the past on several occasions where nothing but a reboot stopped the crashes. It is a very strange issue and I know of others who had the same experience.
 
update: I may have a bad GTS450 card. Pulled the card from the Win7 rig and in put it into a naked rig running WinXP SP3 32bit and it's still NAN'ing. Stock clocks and nothing else running, temps in the 40C range.

Wonder if it's the power supplies?? The Win7 box has a 600W Corsair and the WinXP box has an antec 430W. It ran fine on the Win7 box for a couple weeks then started acting up. :confused:


Code: 
                     Folding@Home Client Version 6.30r2

                          http://folding.stanford.edu

###############################################################################
###############################################################################

Launch directory: C:\Documents and Settings\CrafyOtter\Application Data\Folding@home-gpu


[11:21:48] - Ask before connecting: No
[11:21:48] - User name: CraftyOtter (Team 33)
[11:21:48] - User ID: 25A1CB312705E97D
[11:21:48] - Machine ID: 2
[11:21:48] 
[11:21:48] Gpu type=3 species=0.
[11:21:48] Loaded queue successfully.
[11:21:48] Initialization complete
[11:21:48] - Preparing to get new work unit...
[11:21:48] Cleaning up work directory
[11:21:48] + Attempting to get work packet
[11:21:48] Gpu type=3 species=0.
[11:21:48] - Connecting to assignment server
[11:21:49] - Successful: assigned to (171.64.65.64).
[11:21:49] + News From Folding@Home: Welcome to Folding@Home
[11:21:49] Loaded queue successfully.
[11:21:49] Gpu type=3 species=0.
[11:21:51] + Closed connections
[11:21:51] 
[11:21:51] + Processing work unit
[11:21:51] Core required: FahCore_15.exe
[11:21:51] Core found.
[11:21:51] Working on queue slot 07 [December 1 11:21:51 UTC]
[11:21:51] + Working ...
[11:21:51] 
[11:21:51] *------------------------------*
[11:21:51] Folding@Home GPU Core
[11:21:51] Version 2.15 (Tue Nov 16 09:05:18 PST 2010)
[11:21:51] 
[11:21:51] Build host: SimbiosNvdWin7
[11:21:51] Board Type: NVIDIA/CUDA
[11:21:51] Core      : x=15
[11:21:51]  Window's signal control handler registered.
[11:21:51] Preparing to commence simulation
[11:21:51] - Looking at optimizations...
[11:21:51] DeleteFrameFiles: successfully deleted file=work/wudata_07.ckp
[11:21:51] - Created dyn
[11:21:51] - Files status OK
[11:21:51] sizeof(CORE_PACKET_HDR) = 512 file=<>
[11:21:51] - Expanded 43132 -> 169787 (decompressed 393.6 percent)
[11:21:51] Called DecompressByteArray: compressed_data_size=43132 data_size=169787, decompressed_data_size=169787 diff=0
[11:21:51] - Digital signature verified
[11:21:51] 
[11:21:51] Project: 6800 (Run 2244, Clone 1, Gen 3)
[11:21:51] 
[11:21:51] Assembly optimizations on if available.
[11:21:51] Entering M.D.
[11:21:54] Tpr hash work/wudata_07.tpr:  2231559225 4183444657 1177324001 1137849801 162260058
[11:21:55] Working on PEPTIDE (1-42)
[11:21:55] Client config found, loading data.
[11:21:55] Starting GUI Server
[11:22:03] Setting checkpoint frequency: 500000
[11:22:03] Setting checkpoint frequency: 500000
[11:24:15] Completed    500001 out of 50000016 steps (1%).
[11:26:27] Completed   1000001 out of 50000016 steps (2%).
[11:28:38] Completed   1500001 out of 50000016 steps (3%).
[11:30:50] Completed   2000001 out of 50000016 steps (4%).
[11:33:02] Completed   2500001 out of 50000016 steps (5%).
<snip snip>
[12:27:54] Completed  15000005 out of 50000016 steps (30%).
[12:30:06] Completed  15500005 out of 50000016 steps (31%).
[12:32:17] Completed  16000005 out of 50000016 steps (32%).
[12:34:29] Completed  16500005 out of 50000016 steps (33%).
[12:36:40] mdrun_gpu returned 52
[12:36:40] NANs detected on GPU
[12:36:40] 
[12:36:40] Folding@home Core Shutdown: UNSTABLE_MACHINE
[12:36:43] CoreStatus = 7A (122)
[12:36:43] Sending work to server
 
CraftyOtter said:
Wonder if it's the power supplies?? The Win7 box has a 600W Corsair and the WinXP box has an antec 430W. It ran fine on the Win7 box for a couple weeks then started acting up. :confused:
Click to expand...
Which Antec? If it's a Truepower then it could be problematic, but if it's an Earthwatts then it should be fine.
 
Personally I wouldn't touch any PSU other than a PC Power and Cooling. But that is just me.

The one at 430 watts would worry me some. I would be looking at the 12 line amperage. Not sure what is recommended but should be over 30A somewhere. If it is dual or quad rail I won't use it. Again thats just me.
 
rick5127 said:
Personally I wouldn't touch any PSU other than a PC Power and Cooling. But that is just me.

The one at 430 watts would worry me some. I would be looking at the 12 line amperage. Not sure what is recommended but should be over 30A somewhere. If it is dual or quad rail I won't use it. Again thats just me.
Click to expand...
You should be aware that PCP&C's current offerings are a big step down from their older models, and there are much better choices available on the market.
 
Really... no I didnt know that. I know they got bought out but was hoping they didn't change their quality. I just ordered two more of their 750 Silencers. Hope I didn't mess up.
 
rick5127 said:
Really... no I didnt know that. I know they got bought out but was hoping they didn't change their quality. I just ordered two more of their 750 Silencers. Hope I didn't mess up.
Click to expand...
Well, they aren't bad per se, but they are overpriced, don't perform exceptionally, and aren't modular.
 
Zero82z said:
Which Antec? If it's a Truepower then it could be problematic, but if it's an Earthwatts then it should be fine.
Click to expand...


Major delay in response, sorry. It's an earthwatts 500.

It must have been a problematic WU. Didn't change a thing and eventually it moved onto a different series 6800 and is folding fine for a few days now.

I had manually deleted queue/unitinfo and the contents of work/ several times but kept downloading the same exact project 6800 (Run 4623, 0, 20) and then choking on it 25+ times ( according to Tracker) :confused:

/shrugs/
 
You must log in or register to reply here.
Back
Top