Deciphering AHT Feedback (Mac Pro 3,1 Early 2008)

Question

Level 1

0 points

Deciphering AHT Feedback (Mac Pro 3,1 Early 2008)

Been having some issues with a recently acquired Mac Pro (Early 2008) Quad Core 2.8gHz, 32gb RAM, running a fresh install of El Capitan on a Samsung Evo 850 500GB SSD - occasional pauses and hangs, plus a few unusual problems:

won't read a specific external WD hard drive, for example, that functions on every other Mac i've connected it to
I have 32gb of RAM (admittedly not proper 800mhz or 'Apple Approved') of which only 16gb shows when all are plugged in, but 24gb if I remove two of the chips.

Everything I've read points to potential RAM problems so I've installed and run the AHT. I can post more detailed information if needed, but for the moment the only feedback i'm getting (through various configurations to try and lock down any specific faulty DIMMS or riser slots/cards) is as follows:

NO TROUBLE FOUND

POST: DIMM4/DIMM RISER A, DIMM3/DIMM RISER A

What does this 'POST:' code/designation imply? Any other suggestions would be amazing.

Mac Pro, iOS 9.2.1

Posted on Feb 7, 2016 4:12 PM

Reply

Answer 1

Best reply

Grant Bennet-Alder

Level 10

127,195 points

Feb 7, 2016 7:06 PM in response to TrentonTWS

Power On Self Test.

On this Mac, it runs with error correction Hardware used aggressively. ANY error, correctible or not, during those few seconds and the slot is declared Empty and will not be used by the system. Unless it has some excuse, that DIMM is BAD.

Reply

Answer 2

TrentonTWS Author

Level 1

0 points

Feb 7, 2016 7:16 PM in response to Grant Bennet-Alder

ah that makes sense. I'll run a few more tests with that in mind - but is it possible that it's the slot on the riser, rather than the specific DIMM? I have a feeling I've already tried the configuration with different DIMMs that have otherwise tested fine so i'm inclined to believe it's the slot if that's possible?

Reply

Answer 3

Feb 7, 2016 7:21 PM in response to TrentonTWS

but is it possible that it's the slot on the riser, rather than the specific DIMM?

Possible, yes.

Likely, No.

RAM errors are vexingly difficult to catch in the act. And if you look away for a moment, they stop failing.

But the error correction Hardware in Macs with Xeon processor watch each and every read from memory. If an error occurs in the system, it is corrected in one extended memory cycle, and processing continues. But the fact that a correction has occurred is collected by a background process, and tabulated. This can be read out in this STATIC report in System Profiler (about this Mac > ... > Report > Memory

graphic from anandtech.com

Reply

Answer 4

TrentonTWS Author

Level 1

0 points

Feb 7, 2016 8:32 PM in response to Grant Bennet-Alder

Ok, so I removed the 'offending' DIMMs from the riser, to which I was greeted with DIMM 1 and 2 on the same riser also returning a failed POST. Upon removing those, no error with an empty riser. Put 2 (tested and working) DIMMs into slots 1 and 2 and switched risers A & B, got the POST error again but this time from Riser B.

Does that sound like the riser card?

Reply

Answer 5

Feb 7, 2016 8:48 PM in response to TrentonTWS

I never say never.

Replacement riser-cards are available if you are accumulating evidence that one is failing or has failed.

There are so many thousands more transistors that run at high temperatures in the FBDIMMs, that they are the most likely failure.

Reply