Sidebar Sidebar
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

memtestosx failed... help decode my log!

B

Brendon Bauer

macrumors 6502
Original poster
May 14, 2007
344
0
Good 'ol USofA
I'm no guru when it comes to running things via command line. I ran memtestosx in single user mode (I'm running a fresh install of Leopard by the way) using the command /Applications/memtest/memtest all 3 -L. The test ran fine (I have 4GB of ram) until it got to around 57 or 64 something... that's when it read "FAILED! bla bla bla." Then after a while later it completely locked up and I had to hard reboot. I'd love to share my log but apparently I don't have privileges to access /private/var/root/ in OS X so I can't get to the dang log. Somebody please help :). I'd love to get to the log and share it with you to find out why in the heck the test failed. I sure hope I don't have bad ram!

Edit: log added below
 
Brendon Bauer said:
I'm no guru when it comes to running things via command line. I ran memtestosx in single user mode (I'm running a fresh install of Leopard by the way) using the command /Applications/memtest/memtest all 3 -L. The test ran fine (I have 4GB of ram) until it got to around 57 or 64 something... that's when it read "FAILED! bla bla bla." Then after a while later it completely locked up and I had to hard reboot. I'd love to share my log but apparently I don't have privileges to access /private/var/root/ in OS X so I can't get to the dang log. Somebody please help :). I'd love to get to the log and share it with you to find out why in the heck the test failed. I sure hope I don't have bad ram!
Click to expand...

You could try to login as root, open directory Utility in Applications/Utilities and then click on the lock, enter password and in the edit menu choose Enable Root User.
 
This is what I've tried:

I went back into single-user mode. I did a "cd /private/var/root/" which was fine. I did an "ls" and it did not list anything with .log at the end or anything remotely looking like memtest. I was told in the memtest directions that it would save a log to /private/var/root as memtest.log. Does it not save a log if it crashes? That seems silly, and counterproductive.

Now I don't know how to explain what happened... :/ Should I not worry about it if my computer is running fine?
 
Ok, so went back into single user mode to try another run at memtestosx. I set it to one pass this time, and had it save the log into the memtest folder so I could access it. This time it came up with lots of crap, so I quit the test before it was going to freeze. I typed "reboot" and it rebooted but immediately kernal panicked... Then I restarted it manually and now I'm here. Here's my log. Is my ram bad? Could I possibly have not pushed it in far enough? What should I do... send it back and replace my original ram?

Code: 
Memtest version 4.21 (64-bit)
Copyright (C) 2004 Charles Cazabon
Copyright (C) 2004, 2005, 2006, 2007 Tony Scaminaci (Macintosh port)
Licensed under the GNU General Public License version 2 only

Log file written to /Applications/memtest/memtest.log

Mac OS X 10.5.1 (9B18) running in single user mode
Memory Page Size: 4096
System has 2 Intel processor(s) with SSE
Requested memory: 3744MB (3926577152 bytes)
Available memory: 3744MB (3926577152 bytes)
NOTE: Memory request is too large, reducing to acceptable value...
Allocated memory: 3639MB (3816632960 bytes) at local address 0x0000000101000000
Attempting memory lock... locked successfully
Splitting allocated memory into 2 comparison buffers...
Buffer A: 1819MB (1908316480 bytes) starts at local address 0x0000000101000000
Buffer B: 1819MB (1908316480 bytes) starts at local address 0x0000000172be9940

Running 1 test sequence... (CTRL-C to quit)

Test sequence 1 of 1:

Running tests on entire 3639MB region...
  Stuck Address       :    ok
  Linear PRN          :    ok
Running comparison tests using 1819MB buffers...
  Random Value        :    

FAILURE! Data mismatch at local BUFA address 0x0000000119e55ac8, BUFB address 0x000000018ba3f408
BUFA Data: 0x7ef9c4aef37b7701, BUFB Data: 0x7ef9c4aef37b7777

  Compare XOR         :    

FAILURE! Data mismatch at local BUFA address 0x0000000119e55ac8, BUFB address 0x000000018ba3f408
BUFA Data: 0x9506cf6744088641, BUFB Data: 0x9506cf6744088637

  Compare SUB         :    

FAILURE! Data mismatch at local BUFA address 0x0000000119e55ac8, BUFB address 0x000000018ba3f408
BUFA Data: 0x8948d30ac81d4e70, BUFB Data: 0x8948d30ac81d4e66

  Compare MUL         :    

FAILURE! Data mismatch at local BUFA address 0x0000000119e55ac8, BUFB address 0x000000018ba3f408
BUFA Data: 0xc080c57e9c23d250, BUFB Data: 0x2ac60cdee0dd3192

  Compare DIV         :    

FAILURE! Data mismatch at local BUFA address 0x0000000113c038c8, BUFB address 0x00000001857ed208
BUFA Data: 0x0000000000000001, BUFB Data: 0x0000000000000000

  Compare OR          :    ok
  Compare AND         :    ok
  Sequential Increment:    ok
  Solid Bits          :    

FAILURE! Data mismatch at local BUFA address 0x0000000143004ec8, BUFB address 0x00000001b4bee808
BUFA Data: 0x0000000000000046, BUFB Data: 0x0000000000000000
 
I went back and ran the test again but instead told it to use only 3200mb for testing. The test ran fine! Does that mean that part of that last 800mb of my ram is bad, or that it was just trying to use too much memory to test?

Code: 
Memtest version 4.21 (64-bit)
Copyright (C) 2004 Charles Cazabon
Copyright (C) 2004, 2005, 2006, 2007 Tony Scaminaci (Macintosh port)
Licensed under the GNU General Public License version 2 only

Log file written to /Applications/memtest/memtest.log

Mac OS X 10.5.1 (9B18) running in single user mode
Memory Page Size: 4096
System has 2 Intel processor(s) with SSE
Requested memory: 3200MB (3355443200 bytes)
Available memory: 3745MB (3927392256 bytes)
Allocated memory: 3200MB (3355443200 bytes) at local address 0x0000000101000000
Attempting memory lock... locked successfully
Splitting allocated memory into 2 comparison buffers...
Buffer A: 1600MB (1677721600 bytes) starts at local address 0x0000000101000000
Buffer B: 1600MB (1677721600 bytes) starts at local address 0x0000000165000000

Running 1 test sequence... (CTRL-C to quit)

Test sequence 1 of 1:

Running tests on entire 3200MB region...
  Stuck Address       :    ok
  Linear PRN          :    ok
Running comparison tests using 1600MB buffers...
  Random Value        :    ok
  Compare XOR         :    ok
  Compare SUB         :    ok
  Compare MUL         :    ok
  Compare DIV         :    ok
  Compare OR          :    ok
  Compare AND         :    ok
  Sequential Increment:    ok
  Solid Bits          :    ok
  Block Sequential    :    ok
  Checkerboard        :    ok
  Bit Spread          :    ok
  Bit Flip            :    ok
  Walking Ones        :    ok
  Walking Zeroes      :    ok

All tests passed!  Execution time: 3333 seconds.
 
It doesn't really matter where the Ram is defective, because it's still going to be defective. I would try to get an RMA of the RAM because it's only going to lead to problems down the road.
 
How do I know if it's the RAM being defective? It says in the directions for memtestosx that if it tries to test too much ram that the test can fail and lock up. Apparently the guy has tried to make it so it automatically uses the right amount but if it doesn't work, to let him know. Do you think it's telling me that the RAM is defective, or that the test just failed because of using too much of the ram?
 
Id keep adding on to the room more and more that memtest is allowed to use until you get to one of two points:

1 it fails and you are using up all your memory (eg. the memory is fine and its trying to use up more ram than u have)

2 it fails before you give memtest a number close to the amount of ram that you have installed

Start with 3400 ram and then 3600 3800 etc until you can figure out that u tested all the ram and it came back just fine :confused:
 
It's likely to be bad ram. Out of the amount of ram it has successfully allocated, it found an error at such and such location.

I use memtestosx on the a similar machine, with the same amount of ram and get no errors - so it is likely to be a problem with the ram.

I'd suggest taking out one SO-DIMM and running the test on each module separately. It's likely you'll be able to narrow down the bad so-dimm module.

A little troubleshooting is surely better than trying to figure out why you get seemingly random kernel panics later on.
 
Well, I do recall that between the first test (that failed) an the second test with less ram allocated (that passed with flying colors) I did open up the ram compartment and just try to push on the ram in farther. It didn't move... but maybe it did? Hmm... Just for kicks I am running another test with "all" to see what happens, and it's passed most everything now with "OK". It never made it past the first check when it miserably failed the first time I tested. Maybe the ram wasn't quite seated right? Hmm... If everything passes alright with "OK"s on "all" I have nothing to worry about right?
 
It should be fine, and it could have been that the Ram wasn't seated properly. I would let the test run for a little while (at least an hour) and see if you get any errors.
 
My first guess is the RAM is bad. But it is possible that other components are responsible including your memory controller, logic board and processor. I understand the operations it is trying to do but i don't know why its complaining about the differences in two partitions of your RAM. The test is splitting your RAM into two areas and performing basic arithmetic operations on I don't know what values and then comparing something. What I don't know.
 
I'm pretty sure the RAM is ok. I've run the test completely two more times, all with "OK"s so it passed all tests. Each test took over an hour. I don't think there could be anything wrong if it never says failed anymore...
 
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back