X1000 acting up (not quite solved)

[abalaban]

Hi,

Bill, a few weeks ago Steven reworked the wiki URLs (there is no "index.php" part anymore, instead there is a "wiki"). So the new correct URL is
http://wiki.amigaos.net/wiki/AmigaOne_X1000

(BTW in such case, one can use the search box at the top of the page )

[jaokim]

I suggest you grab memtester off Depot & run it overnight to test your memory. We had one other user with a flakey system & memtester found one bad memory stick, he replaced it & fixed. Memtester takes a long time to check memory, so run it overnight, you only need 1 pass to check it/them.

The X1000 has been behaving on and off, but I finally got around to run Memtester. I however decided to abort it, which ended the program nicely, but then immediately my screen turned black, and all I could do was to turn the X1000 off.

So it looks like it is the memory...

[LyleHaze]

I finally got around to run Memtester. I however decided to abort it, which ended the program nicely, but then immediately my screen turned black, and all I could do was to turn the X1000 off.

So it looks like it is the memory...

I wouldn't bet on it.
If a program was going to test ALL memory, it would have to overwrite the RAM used by the OS to do so.

I'm not the author of memtester, but I wouldn't assume mem is bad because memtester aborted badly.

[jaokim]

Well, I don't really know what to try. I have to do something. I do feel it seems unlikely that the memory should be faulty, but given that the log indicates something memory related, and it just frooze after running memtester. It's almost as if the memory fails when in use...?

the log :

[HELO][DRAM]SDRAM: ECC off, Non-ECC DIMM used on channel 0.
[RELO][HELO][DRAM]SDRAM: ECC off, Non-ECC DIMM used on channel 0.
[RELO][HELO][DRAM]SDRAM: ECC off, Non-ECC DIMM used on channel 0.
...

[LyleHaze]

In your previous post, you said:
" I however decided to abort it, which ended the program nicely, but then immediately my screen turned black,"

But in your last post you said:
"and it just frooze after running memtester."

That's another situation entirely.

My comment was trying to point out that terminating a memory test early MIGHT have been the cause of the machine "going black", and then requiring a reboot.
That does NOT mean that your memory is faulty, nor does it mean that your memory is good.
All it means is that the memory test program might not exit gracefully if aborted.

If you'd like to test your memory, please do!
After the memory tester has run a complete cycle, you'll have more reliable information about your memory than you can get by aborting the program before it's finished.

I hope that makes a bit more sense.

[jaokim]

I hear what you say, and can agree.
I do however have no other clue than a faulty memory. Given that the black screen freeze is what happened even when I didn't run memtester, it seems likely that memtester rather than being directly responsible, triggered the error by doing its "memory stuff".
It can of course be a plethora of stuff that's wrong, but the memory seems in the given context likely, and it is a relatively inexpensive thing to check.

[Christopher Follett]

I hear what you say, and can agree.
I do however have no other clue than a faulty memory. Given that the black screen freeze is what happened even when I didn't run memtester, it seems likely that memtester rather than being directly responsible, triggered the error by doing its "memory stuff".
It can of course be a plethora of stuff that's wrong, but the memory seems in the given context likely, and it is a relatively inexpensive thing to check.

Are you talking about CFE Mem tester?

The bit of log you posted is the memory starting up. It always does this at every boot.
Its always the first few lines of log.

[jaokim]

The memtester is the program linked above. It's started from CLI, so I don't think it is the CFE MemTester.

The log would be okay if it wasn't repeating itself, but now it either fails to give any output at all, or just repeats these lines over and over:

Code: Select all

the log : 
[HELO][DRAM]SDRAM: ECC off, Non-ECC DIMM used on channel 0.
[RELO][HELO][DRAM]SDRAM: ECC off, Non-ECC DIMM used on channel 0.
[RELO][HELO][DRAM]SDRAM: ECC off, Non-ECC DIMM used on channel 0.
...

[LyleHaze]

I think it's a good time to open a ticket with AmigaKit.

I am curious about something though..
While it's doing it's constant reset thing, what is the states of each
of the eight LEDs on the Nemo board?

The book says there are six, but there's really eight

Also, assuming you are in an X1000 case, they are numbered
(left to right) 8, 7, 6, 5, 4, 3, 2, 1

I expect 8 and 7 to be very dimly lit (almost off)
These are driven by Xena and dimly lit is correct if Xena is in reset mode.
The other six should be (from 6 to 1).. Off ON ON ON Off Off, if I recall
correctly. It might be interesting to know if any of these are changing
when it reboots itself.

Good Luck!

[Christopher Follett]

The memtester is the program linked above. It's started from CLI, so I don't think it is the CFE MemTester.

The log would be okay if it wasn't repeating itself, but now it either fails to give any output at all, or just repeats these lines over and over:

Code: Select all

the log : 
[HELO][DRAM]SDRAM: ECC off, Non-ECC DIMM used on channel 0.
[RELO][HELO][DRAM]SDRAM: ECC off, Non-ECC DIMM used on channel 0.
[RELO][HELO][DRAM]SDRAM: ECC off, Non-ECC DIMM used on channel 0.
...

How much ram do you have?
Try reseating RAM.