New Win SMP beta being flaky

Soldato
Joined
19 Jan 2006
Posts
16,889
Anyone else having issues with the win smp new beta since 1st Feb??

Have been running the win smp for ages now no hassle what so ever, however since installing the new winsmp beta on 2nd feb it's be very very inconsistant.

Basically seems to stop working at the merest little thing interupting it. It occasionally used to stop when my wireless connection went down but it now seems to be stopping after 5/6 frames all the time. Overnight last night again it's just stopped.:mad::mad:

No new overclocks or anything like that - it's not erroring but just stopping

# SMP Client ##################################################################
###############################################################################

Folding@Home Client Version 5.91beta6

http://folding.stanford.edu

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

Launch directory: C:\Program Files\Folding@Home Windows SMP Client V1.01
Executable: C:\Program Files\Folding@Home Windows SMP Client V1.01\fah.exe


[22:18:18] - Ask before connecting: No
[22:18:18] - User name: booyaka (Team 10)
[22:18:18] - User ID: 12D0A86B52470D11
[22:18:18] - Machine ID: 1
[22:18:18]
[22:18:18] Loaded queue successfully.
[22:18:18]
[22:18:18] + Processing work unit
[22:18:18] Core required: FahCore_a1.exe
[22:18:18] Core found.
[22:18:18] Working on Unit 07 [February 7 22:18:18]
[22:18:18] + Working ...
[22:18:18]
[22:18:18] *------------------------------*
[22:18:18] Folding@Home Gromacs SMP Core
[22:18:18] Version 1.74 (March 10, 2007)
[22:18:18]
[22:18:18] Preparing to commence simulation
[22:18:18] - Ensuring status. Please wait.
[22:18:35] - Looking at optimizations...
[22:18:35] - Working with standard loops on this execution.
[22:18:35] Examination of work files indicates 8 consecutive improper terminations of core.
[22:18:41] - Expanded 2940977 -> 15163684 (decompressed 515.6 percent)
[22:18:41]
[22:18:41] Project: 2653 (Run 0, Clone 151, Gen 55)
[22:18:41]
[22:18:42] Entering M.D.
[22:18:49] Calling FAH init
[22:18:49] Read topology
[22:18:50] g local files
[22:18:50] Completed 50000 out of 500000 steps (10 percent)
[22:18:50] Extra SSE boost OK.
[22:18:50] les
[22:18:50] Completed 50000 out of 500000 steps (10 percent)
[22:18:51] Extra SSE boost OK.
[22:34:06] Writing local files
[22:34:06] Completed 55000 out of 500000 steps (11 percent)
[22:48:01] Writing local files
[22:48:01] Completed 60000 out of 500000 steps (12 percent)
[23:00:42] Writing local files
[23:00:42] Completed 65000 out of 500000 steps (13 percent)
[23:13:24] Writing local files
[23:13:24] Completed 70000 out of 500000 steps (14 percent)

Folding@Home Client Shutdown at user request.

Folding@Home Client Shutdown.

That's me having to shut it down this morning @ 7.43am cause i's just not running.
 
I'm running two WinSMPs on the same machine. Both the latest Beta. No problems since Sunday when I started them.

ms9cw
 
I switched 3 of my rigs over to the new beta before I left and have had no such problems with any of them :confused:
 
the new beta WinSMP died after two frames on my home rig too, thought it was my ram timings - but died again when slackened. Switched back to Linux as Finstall now links to new beta :)

Running fine at work with the looping batch file
 
weird - will have a gander tonight again but really lost out on about 3 days folding now.

Will try a full uninstall/delete of all current files/work and reinstall and see if that works - if not off to try the standard units (NOOOOOOOOOOOOOOOOOO!)
 
I had trouble with the SMP client, rebuilt the machine and dropped the cpu clock by 100MHz and has been running fine for 3 days so far

fair enough - not sure what's happening but i've got 2 x 2gb of new ram to install and i "might" :eek::eek: have a go at this linux thing;)
 
thanks will have a gander later but "scared" with just the word linux!!:eek:

So was I - but if you can follow instructions to the letter (every command is case senative) you'll be fine.

Most of my early problems (and I had loads) were due to typo's, and the current Ubuntu version and finstall have removed the other issues I struggled with.

Would i recomend it over windows - no.
but then again I would recomend vista either ;)
But it no more complex that OCing a pc. And with the guide SB wrote it just following a path (don't ask me what the commands mean - all I know is that it works and I trust this lot with the code I'm entering)

The only extras i'd add to the guide is you don't even need to burn the iso file to disk - just run it from the harddrive (saveing 10 mins on the install time) - in VMware aftercreating drive - click edit, select HHD and click 'from iso' on right hand side. confirm location and start VM
and you need to type 'sudo make install' as the last command to install fahmon in ubuntu.
 
Last edited:
cheers for that - i take it vmware is basically just like creating a virtual drive with linux on it??

As i write this at home again - completed about 4 frames on both smp clients and then for some unknown reason my wireless has dropped out and smp has stopped. Very strange as my wireless is rock solid stable and never drops connection until i start up the new beta and it starts losing signal/strength etc etc.

Will maybe go with some standard units in the mean time till i get a chance to sort out vmware/linux etc etc with a spare hour or so.

nightmare
 
Just had a client crash on me. To my knowledge, I've never had a client crash before (that wasn't due to me closing it). Log:

Code:
[15:31:19] *------------------------------*
[15:31:19] Folding@Home Gromacs SMP Core
[15:31:19] Version 1.74 (March 10, 2007)
[15:31:19] 
[15:31:19] Preparing to commence simulation
[15:31:19] - Ensuring status. Please wait.
[15:31:36] - Assembly optimizations manually forced on.
[15:31:36] - Not checking prior termination.
[15:31:42] - Expanded 2945702 -> 15229121 (decompressed 516.9 percent)
[15:31:42] - Starting from initial work packet
[15:31:42] 
[15:31:42] Project: 2653 (Run 31, Clone 112, Gen 55)
[15:31:42] 
[15:31:43] Assembly optimizations on if available.
[15:31:43] Entering M.D.
[15:31:50] Rejecting checkpoint
[15:31:51] Protein: Protein in POPC
[15:31:51] xtra SSE boost OK.
[15:31:51] 
[15:31:52] Extra SSE boost OK.
[15:31:53] Writing local files
[15:31:53] Completed 0 out of 500000 steps  (0 percent)
[15:41:52] Timered checkpoint triggered.
[15:48:36] Writing local files
[15:48:37] Completed 5000 out of 500000 steps  (1 percent)
[15:58:37] Timered checkpoint triggered.
[15:59:31] CoreStatus = 7B (123)
[15:59:31] Client-core communications error: ERROR 0x7b
[15:59:31] Deleting current work unit & continuing...
[15:59:45] Killing all core threads
[15:59:45] Killing SMP core threads
[15:59:45] Killing 2 cores
[15:59:45] Killing core 0
[15:59:45] Killing core 1

Folding@Home Client Shutdown at user request.
[15:59:45] ***** Got a SIGTERM signal (2)
[15:59:45] Killing all core threads
[15:59:45] Killing SMP core threads
[15:59:45] Killing 2 cores
[15:59:45] Killing core 0
[15:59:45] Killing core 1

Folding@Home Client Shutdown.

The folding forum says:
21) Early Unit End Error 0x7b may or may not be a bug, but it needs to be defined. And if data is being sent back to the servers, it is not acknowledged or credited. See also #11 above. We have reason to believe this is not a client/core error (there is no 0x7b error defined in the fah code), but is more likely caused by clients made unstable from overclocking or a network settings change (#14) (or each may cause it). Further investigation and testing is necessary.
But my Q6600 is running at stock, the temps are great and I've never had a problem with it before. The uni network is stable and nothing's changed on it.
 
I had the same problem running the old beta. Really didn't like having two WinSMP clients together. I recently had a LinuxSMP client shut down without warning (was fine when I restarted it).
 
Back
Top Bottom