EARLY_UNIT_END... anyone else?

[DaIceMan]

My main SMP folding machine has had EARLY_UNIT_END on the last 2 units. 1 at 17% I think and this one (log below) at 54%. Anyone else having trouble with this, or know what I can do to fix it? I had restarted the computer early in this unit (8 or 9%) to do the windows updates, but it started back up and and appeared to be working fine. I just noticed that I had a unit for 165 points and saw the early.

Thanks in advance.

[12:30:29] Writing local files
[12:30:29] Completed 540000 out of 1000000 steps (54 percent)
[12:36:16] Warning: long 1-4 interactions
[12:36:16] Gromacs cannot continue further.
[12:36:16] Going to send back what have done.
[12:36:16] logfile size: 347753
[12:36:16] - Writing 348289 bytes of core data to disk...
[12:36:16] ... Done.
[12:36:16] - Failed to delete work/wudata_02.arc
[12:36:16] - Failed to delete work/wudata_02.dyn
[12:36:16] - Failed to delete work/wudata_02.bed
[12:36:16] - Failed to delete work/wudata_02.sas
[12:36:16] - Failed to delete work/wudata_02.goe
[12:36:16] Warning: check for stray files
[12:38:16]
[12:38:16] Folding@home Core Shutdown: EARLY_UNIT_END
[12:38:16]
[12:38:16] Folding@home Core Shutdown: EARLY_UNIT_END
[12:38:20] CoreStatus = 7B (123)
[12:38:20] Client-core communications error: ERROR 0x7b
[12:38:20] Deleting current work unit & continuing...
[12:40:25] - Preparing to get new work unit...
[12:40:25] + Attempting to get work packet
[12:40:25] - Connecting to assignment server
[12:40:25] - Successful: assigned to (171.64.65.64).
[12:40:25] + News From Folding@Home: Welcome to Folding@Home
[12:40:25] Loaded queue successfully.
[12:40:26] - Error: Attempt #1 to get work failed, and no other work to do.
Waiting before retry.
[12:40:44] + Attempting to get work packet
[12:40:44] - Connecting to assignment server
[12:40:45] - Successful: assigned to (171.64.65.64).
[12:40:45] + News From Folding@Home: Welcome to Folding@Home
[12:40:45] Loaded queue successfully.
[12:40:56] + Closed connections

[dicecca112]

well if the WU was only 166 points, then its not a SMP

Could be unstable overclock or other machine instability, or lose of internet connectivity. Is the forceasm flag on?

[vbironchef]

Sometimes I notice when I restart computer more than 1 or 2 times during a smp client folding the program becomes a real stinker. Takes a very long time to complete, if it does at all. When all else doesn't work I delete program and start all


over again. If I don't do that I run into problems that you are having. It usually takes 24 hours for the program to pick up speed. That's just my personal experience. Happy folding! victor/natasha


P.S. I noticed that your smp is folding a 1 million wu. That might be the problem. I have a q6700 that can sometimes handle that. Sometimes it does what is happening to you.

[dicecca112]

Sometimes I notice when I restart computer more than 1 or 2 times during a smp client folding the program becomes a real stinker. Takes a very long time to complete, if it does at all. When all else doesn't work I delete program and start all


over again. If I don't do that I run into problems that you are having. It usually takes 24 hours for the program to pick up speed. That's just my personal experience. Happy folding! victor/natasha


P.S. I noticed that your smp is folding a 1 million wu. That might be the problem. I have a q6700 that can sometimes handle that. Sometimes it does what is happening to you.

add the flag forceasm, and then you won't have that restart issue. What happens is that its a PITA to get the SMp client to shutdown right. It kills the the SSE optimizations. That is why it goes slow.

[skier]

i have the same problem with my (sig) computer except it usually doesnt finish 30% if it even gets past 10 , it is using the text-only console, and all the WUs it has been trying have been 2,500,000 steps or more @ about 30mins every % and i have tried reinstalling, and not allowing >5mb units, it still wont finish.

-so it could be an unstable OC on something?

[dicecca112]

yes something is unstable

[skier]

is it most likely the CPU? because sometimes, it locks up (so i have to hardware restart) my computer if i dont have speedfan up on the screen(i know, its strange, but that the only way it would keep going)

[DaIceMan]

pretty sure it's not the overclock... it's tested stable for everything I've thrown at it. Just seemed strange that it happened to 2 units in a row. I've had it happen to 1 WU, but that's usually because I had to stop the program and reboot. On the 2nd one that failed, I didn't do anything to provoke it.

[dicecca112]

pretty sure it's not the overclock... it's tested stable for everything I've thrown at it. Just seemed strange that it happened to 2 units in a row. I've had it happen to 1 WU, but that's usually because I had to stop the program and reboot. On the 2nd one that failed, I didn't do anything to provoke it.

What have you done to test the overclock

[DaIceMan]

Orthos for 2 hours+, OCCT 30 minutes, 27 days straight folding 24/7 SMP... and nothing has shown any sign of faltering.

[dicecca112]

Your not stable nor can you call it stable if orthos was only run for 2 hours and OCCT was only run for 30 minutes. If you can run those for 8-12hrs then your stable

[Bwall]

I had 3-4 WU's fail last week also on both of the machines that were folding. I thought it was the overclock so maybe it's not.

[Darkstar]

This is an unknown error from the fahlog of SMP clients. Because 0x7b is not defined in the F@h client or SMP fahcore, it is believed to be a Windows or SMPD/MPICH error code. Known causes are unstable systems from too much overclocking, changing network settings while the client is running, or stopping restarting the client. See the List of Known Issues.

from the wiki

[DaIceMan]

Your not stable nor can you call it stable if orthos was only run for 2 hours and OCCT was only run for 30 minutes. If you can run those for 8-12hrs then your stable

24 hours OCCT stable... so I further contend, I don't think it's the overclock.

[dicecca112]

touche. What about the internet connection

[DaIceMan]

Perfectly fine now. Had some serious issues with it before, but raised enough heck that I finally got action. I'm only aware of 2 disconnects in the last 4-6 weeks.

[DaIceMan]

just to follow up.... 20 hours Ortho stable as well. No other problems since those 2 units. Must have been a glitch in the matrix.

[vbironchef]

I just had a 1 mil wu. It crashed and burned as well. It restarted and I doubt it will make it through the whole wu. Oh well.

Happy turkey day everyone!