[vox-tech] Persistant hardware problem kicking my butt

[Peter Jay Salzman]

I'm emotionally drained by a persistant hardware problem, and when I become
emotionally involved with a problem, I tend to not think clearly.  So I'd
like to bounce something off the list.  Any thoughts greatly appreciated.


Relevent Hardware:
   Abit NF7 (NForce 2 chipset)
   Athlon XP 2200
   2 DDR-400 DIMMs.  512MB each.
   Mad Dog something-or-other DVD/CD writer.  Writes DVD9.  USB 2.0.
   

1. I was burning a bunch of DVDs under Win2k.  Nero's verification
   consistantly failed.

2. I switched to Linux and used k3b.  Verification of burned image
   failed as well.

3. Certain programs under Linux were segfaulting out of the blue.  They
   never segfaulted before (Quake III and Unreal Tournament).  I didn't
   detect any GPF under Windows, but who knows what really goes on?

4. Ran memtest86 three times.  All three times, memtest86 segfaulted.
   It indicated that DIMM 0 was the problem.

5. Removed DIMM 1.  Reran memtest86.  It segfaulted.

6. Removed DIMM 0.  Placed what used to be DIMM 1 into slot 0.  Reran
   memtest86 for about 8 hours.  No errors.

7. Ran Quake III / Unreal Tourn. in Linux for a while.  No segfaults.



At this point, I thought the problem was fixed.  I was wrong.



8. Burned a CD with k3b.  It was a collection of text and binary files.
   Here were a collection of md5sums of what I burned:

      p at lucifer$ md5sum /dvd/*
      ec17179683b1ec7c584371d52d482dc7  2004_suprnova_april_fools.jpg
      2a35dd12706e6a109aa5d51a5b7414d6  ballmerwindows.wmv
      1969446ec71b76c01a852ffabcba7b83  Call Of The Wild.txt
      76319f20a91df19621a2c6ed0144803f  common_passwords.txt
      7fe852ab7b3609657aa3c7e633ba7c02  Imagine (John Lennon).mp3
      611759e7fa9455f2cdc574da4eb875bf  Leonardo.txt
      83681d4c6c34e5c9f924c3b0ea732038  Makefile
      bda631d8fac719f1efb741a61f2f4902  Metamorphosis.txt
      0f3a64e11edc97f32e089d07e8bee7bc  polygons.c
      cacbcb3cf986d1cf658c8b8fff6f6ce6  rip.pl
      8a72cd8e4e59263f4813892c3ab3b91f  roth.zip
      e8bc0afb5ab15a10ab233ceab86936bf  sid.bmp
      3136b05433e0c0e6877699bcd4e603dc  The Complete Ninjas Handbook.pdf
      1a715e61acf7a1bae7f87af03687f11b  tomsrtbt-2.0.103.tar.bz2
      b879804716164d24f92519f33774e9f4  try.c
      bb4416e6db326f9f30e4c73953cbcadf  vmlinuz-2.6.9
      dc6921500007d1b305cc9b583352ef2c  War And Peace.txt
      f9436029f9e7bfd3800860e768a1febd  write_errors.txt
      a553f7a13fb2a0f02c8ff53ae7c378f9  Zarathustra.txt
      5f363e0e58a95f06cbe9bbc662c5dfb6  zero

   Roth.zip is an old DOS game in a zip archive.  The md5sum is different
   from the md5sum of the hard disk copy:

      p at lucifer$ md5sum /dvd/roth.zip
      8a72cd8e4e59263f4813892c3ab3b91f  /dvd/roth.zip

      p at lucifer$ md5sum ~/Desktop/files/roth.zip
      db144e6264cfd84d6276b9488e57a279  /home/p/Desktop/files/roth.zip

   I umount, remount the DVD.  The md5sum changes but is still wrong.
   
      p at lucifer$ umount /dvd/; mount /dvd/; md5sum /dvd/roth.zip
      a1fc154848aa932a7cf44e723ebc6082  /dvd/roth.zip

      p at lucifer$ md5sum /dvd/roth.zip
      a1fc154848aa932a7cf44e723ebc6082  /dvd/roth.zip

      p at lucifer$ umount /dvd/; mount /dvd/; md5sum /dvd/roth.zip
      169e7a2327a054f7859066f91b485f45  /dvd/roth.zip

      p at lucifer$ umount /dvd/; mount /dvd/; md5sum /dvd/roth.zip
      51d5d96a4108285080f16261916f6e74  /dvd/roth.zip

   I finally get the correct md5sum.

      p at lucifer$ umount /dvd/; mount /dvd/; md5sum /dvd/roth.zip
      db144e6264cfd84d6276b9488e57a279  /dvd/roth.zip

   When I mount the burned DVD on a drive connected to a different system,
   the md5sum appears to be consistantly correct.


9. Sometimes, when the machine boots up, POST freezes right after it says
   "checking RAM" but just before it starts to check the RAM.  This is a
   new development.  I can make this problem go away only by shutting off
   the power supply (not powering down.  actually shutting the PS off).


10. I still detect no segfaults that were present before I removed DIMM 0.


11. I've been running memtest86 for a couple of hours now without errors.



I think that's all the data I have.  It's imperative that I have a reliable
fast machine right now to run physics simulations.  That's crucial.  To me,
the available clues seem contradictory.  Somethings points to RAM.  Other
things point to something resident on the mother board.  Still other things
point to the DVD drive.

Help?

Pete

-- 
Save Star Trek Enterprise from extinction: http://www.saveenterprise.com

GPG Fingerprint: B9F1 6CF3 47C4 7CD8 D33E  70A9 A3B9 1945 67EA 951D