Several early terminations

Discussion in 'Distributed Computing' started by lssmit02, Mar 31, 2005.

  1. lssmit02 macrumors 6502

    Joined:
    Mar 25, 2004
    #1
    Several work units have terminated early for me since I moved folding to a flash drive. This may be coincidental, of course. I was wondering if others had seen similar results?

    Here is an example what the log says:

    [12:09:59] Folding@Home Gromacs Core
    [12:09:59] Version 1.80 (March 16, 2005)
    [12:09:59]
    [12:09:59] Preparing to commence simulation
    [12:09:59] - Looking at optimizations...
    [12:09:59] - Created dyn
    [12:09:59] - Files status OK
    [12:09:59] - Expanded 135940 -> 650585 (decompressed 478.5 percent)
    [12:09:59] - Starting from initial work packet
    [12:09:59]
    [12:09:59] Project: 1602 (Run 4, Clone 79, Gen 9)
    [12:09:59]
    [12:09:59] Assembly optimizations on if available.
    [12:09:59] Entering M.D.
    [12:10:05] Protein: p1602_NTL4MUreGS
    [12:10:05]
    [12:10:05] Writing local files
    [12:10:05] Testing CPU type...
    [12:10:05] Done testing.
    [12:10:05] Extra AltiVec boost OK.
    [12:10:06] Writing local files
    [12:10:06] Completed 0 out of 1000000 steps (0)
    [12:11:54] Quit 101 - Fatal error:
    [12:11:54] Step 2467, time 4.934 (ps) LINCS WARNING
    [12:11:54] relative constraint deviation after LINCS:
    [12:11:54] max 0.001621 (between atoms 768 and 770) rms 0.000074
    [12:11:54]
    [12:11:54] Simulation instability has been encountered. The run has entered a
    [12:11:54] state from which no further progress can be made.
    [12:11:54] This may be the correct result of the simulation, however if you
    [12:11:54] often see other project units terminating early like this
    [12:11:54] too, you may wish to check the stability of your computer (issues
    [12:11:54] such as high temperature, overclocking, etc.).
    [12:11:54] Going to send back what have done.
    [12:11:54] logfile size: 7884
    [12:11:54] - Writing 8562 bytes of core data to disk...
    [12:11:54] ... Done.
    [12:11:58]
    [12:11:58] Folding@home Core Shutdown: EARLY_UNIT_END
    [12:12:07] CoreStatus = 72 (114)
    [12:12:07] Sending work to server

    Any info would be appreciated
     
  2. daveL macrumors 68020

    daveL

    Joined:
    Jun 18, 2003
    Location:
    Montana
    #2
    I have 7 cpus folding and none have terminated early any time in recent memory, although they all fold to a hard drive, like most people that fold. Maybe some else that uses a flash drive to fold will chime in?
     
  3. bousozoku Moderator emeritus

    Joined:
    Jun 25, 2002
    Location:
    Gone but not forgotten.
    #3
    Many times, it's a physical constraint and the work unit itself needs to be re-configured somehow to allow the simulation to be calculated properly.
     
  4. daveL macrumors 68020

    daveL

    Joined:
    Jun 18, 2003
    Location:
    Montana
    #4
    Does this happen less frequently with tinker WUs? That's about all I've gotten for a while now, and I haven't had any terminate early. Anyway ...
     
  5. lssmit02 thread starter macrumors 6502

    Joined:
    Mar 25, 2004
    #5
    These have all been gromacs. I'll look into the physical constraints of the flash drive - it's formatted as PC-DOS (as all flash drives are to start, I think), and had close to 100 megs of space when I installed folding.

    Perhaps I should reformat the flash drive to HFS, and remove all other files from the drive.
     
  6. daveL macrumors 68020

    daveL

    Joined:
    Jun 18, 2003
    Location:
    Montana
    #6
    I would think a native Mac filesystem on the Flash drive would be more reliable, but I haven't done any testing myself. It certainly can't hurt.
     
  7. bousozoku Moderator emeritus

    Joined:
    Jun 25, 2002
    Location:
    Gone but not forgotten.
    #7
    As far as I remember over the past 3 years, I've never had it happen with a Tinker WU. In fact, I had mentioned in another thread that I believed the Double Gromacs core was created just to tackle these matters of precision.

    By the way, it has nothing to do with the flash drive.
     
  8. daveL macrumors 68020

    daveL

    Joined:
    Jun 18, 2003
    Location:
    Montana
    #8
    Thanks for the info. BTW, since the flash drive is a 3rd party USB device, using a MS file system format, data corruption is not out of the realm of possibility, although I'm not saying it's likely.
     
  9. bousozoku Moderator emeritus

    Joined:
    Jun 25, 2002
    Location:
    Gone but not forgotten.
    #9
    It would be highly unlikely since the data in question would be in real memory, not on external storage.
     
  10. stubeeef macrumors 68030

    stubeeef

    Joined:
    Aug 10, 2004
    #10
    I have had 2 early, near the end, terminations (gromacs) that got 10 pts and 18 tps, and the most recent 0 pts. Went 100% got the wu but 0 fricken pts.!

    all within the last week, and never had any before. I have lost a good 150pts on these and I am ticked!

    It is happening on the dual 450 running the dual processor program.
     
  11. lssmit02 thread starter macrumors 6502

    Joined:
    Mar 25, 2004
    #11
    Well, just to be sure, I reformatted the flash drive, and only installed the F@H folders. Everything's be fine since then. I suspect that the early terminations were caused by running out of room on the flash drive, causing the client to be unable to save the data. However, I can't verify this, because I thought there was enough space before the reformatting of the drive. Nevertheless, things have been fine since I reformatted.

    Most irritating thing: constant flashing from the USB drive. I'll have to cover that over with electrical tape :rolleyes:
     
  12. lssmit02 thread starter macrumors 6502

    Joined:
    Mar 25, 2004
    #12
    Isn't it irritating :mad:

    What error message did you get?
     
  13. stubeeef macrumors 68030

    stubeeef

    Joined:
    Aug 10, 2004
    #13
    THIS
     

    Attached Files:

Share This Page