MemTest86 fails to run on iMac Pro ...

[mossygreen]

I followed the directions for creating a bootable USB image, disabled the Apple boot device restrictions, and can boot from the MemTest86 image. I get a text screen that echos lines that relate to starting up (e.g., reading CPU MSR info, etc.). The last line I see says something about looking at multiprocessor support.

Then, the screen blanks ... there is a pause of several seconds ... and the machine reboots. I think I should see the top-level splash screen shown in the documents, but I don't.

Is this configuration expected to work, or is the graphics setup of the iMac Pro too different?

Does it sound like something that could be cause by a bad USB stick image? That drive is a few years old so perhaps there is bit rot somewhere ...

Any help/ideas appreciated.

[Richard (PassMark)]

Can you post the debug log from MemTest86?

https://www.memtest86.com/tech_debug-logs.html

[mossygreen]

I had no idea there was a log ... but it's where the docs pointed. There are two of them - I've attached both. In both instances the end result is the same but the tail of the log is slightly different.

[keith]

Thanks for the logs.

Looks like this model Mac has a UEFI firmware bug preventing multi-processing.

Can you try adding the following to the blacklist.cfg file under EFI\BOOT\ of the USB drive:

"Mac-7BA5B2D9E42DDD94",ALL,EXACT,RESTRICT_MP

[JAustin]

Can you try adding the following to the blacklist.cfg file under EFI\BOOT\ of the USB drive:

"Mac-7BA5B2D9E42DDD94",ALL,EXACT,RESTRICT_MP

I came here to report this same problem, saw this response, followed this suggestion, and Memtest86 10.4 Free runs now.

I have 64 GB to test and only one CPU is active, so the test has a long time left to run, but is running…

[mossygreen]

I got busy with other things so I just got around to trying the suggested workaround. I can also confirm that this fixes the restart on startup issue and that MemTest86 appears to run fine. With one pass looking like it would take roughly 4-5 hours, I interrupted the test until I can just let it run overnight.

I will note that this is really only a workaround - some memory errors won't be caught with a single core politely doing it's thing. However, since all I did was upgrade the memory on the system I am only looking for a gross test that the DIMMs are okay.

Thanks!

[David (PassMark)]

I will note that this is really only a workaround

Correct.

In an ideal world we could report the firmware bug to Apple and they would fix it.
Unfornately this is not an ideal world & Apple don't care.