Mac Pro 3,1 black screen then fans on full

Panic Reports are stored at:

/Library/Logs/DiagnosticReports

If you copy and paste that string into:

Finder > Go menu > Go to Folder

it will take you to where those reports are stored.

They are named with Date&Time and end in .panic

You can post the entire report here, by using the “additional text” Icon in the reply header (looks like a paper with writing).

Please don’t post more about 20 lines of any other types of reports — they are interminable, and any information useful for this purpose is on the first screenful.

There are three quick take-aways from any panic report.

1) The panic-reason,

2) the extensions present at the "scene of the crime", and

3) the BSD process in which the problem occurred.

One more item that is important is the names of any third-party Extensions you have added. They are shown FIRST in the extensions loaded section

Well, I don't want to jinx it, because this issue has been haunting me for YEARS, and I have to assume it could return at any moment. But that said, it's been a month since I moved my graphics card to slot 2, and I haven't had any crashes so far...

These two most recent posts are NOT similar to your previous ones.

They are segmentation faults -- references to areas of I/O or memory not mapped to any task.

They occurred in the File System Event Daemon, the always-running File system event processing task. You should consider doing a through Disk Repair, especially of your entire boot drive (each partition separately including the physical device.)

Yup, a Mac Pro 3,1.

Unaltered bios etc.

Crashing went on for a couple of years, steadily getting worse and worse. Swapping motherboards didn't reduce the frequency, but did reduce the severety. In hindsight, I'm not actually sure when the problem started - System never had issues with a GTX 480 and running Mountain Lion. Problem seemed to be introduced somewhere after upgrading the OS (several times), the card, and the drivers.

Both before and after, I was fiddling with different configurations of Nvidia and CUDA drivers. Used a website (http://www.macvidcards.com/drivers.html) to determine and download the best drivers based on my OS.

I'm running Yosemite (OSX 10.10.5). I could technically move the 3,1 up one more version, but I'm terrified it'll become (even) less stable.

I've never had the logs checked by an expert.

You hare showing an APFS Volume in El Capitan.

APFS was introduced in High Sierra, 10.13, two MacOS version later.

If the System Management Controller fails to turn the fans speeds DOWN over several seconds, they fail-safe to full speed. The fans going full speed are a side effect of some sort of serious malfunction that stopped processing -- they are not the cause of the malfunction.

My guess was that there might be some stored panic reports.

But If you are on the trail of bad power -- then by all means, press ahead with that first.

<< 45 minutes later and the exact same issue occurred. >>

Kernel panic for ANY reason will present very similar symptoms.

You should check the logs again and see if you have a stored panic report.

The 'approved method' for this debugging is to strip the machine down to the smallest working set of parts that can do anything, run it that way for a while, then slowly add aback one component at a time until it ides, then suspect the last-added part.

I'm joining in here, on the off-chance that sharing my similar nightmare will help contribute to a solution getting found.

MacPro 3,1, Yosemite, Nvidia GTX 680. I have a very similar issue, and nothing I've tried has ever fixed it. The only thing that even made a dent in it was replacing my motherboard, which changed it from a hard crash with spinning fans and auto-reboot to a softer crash that at least leaves logs.

For me it's a cascading failure. I'll run fine for a week or two, then crash. The next crash will come much sooner, maybe a day or two. Then after a couple hours, then after a minute or two, rendering the computer useless. Before the motherboard replacement, it would stop booting altogether. The only way to get the system up and running again was stripping out everything leaving just the boot drive, one stick of RAM, and putting back the stock graphics card it came with. System would boot, and then I could put everything back and it would run fine again for 1-6 weeks.

Since replacing the motherboard, it never runs smoothly for more than 2 weeks, but it also rarely stops booting. I can usually re-stabilize by resetting PRAM and SMC, restarting, booting in safe mode, restarting AGAIN, then doing another fresh restart. Why do I have to do that several times before the system stops crashing? No idea. But it works for a while. Then sooner or later, it always crashes again. Could be idle, could be in the middle of heavy work. There's no pattern.

Here's other things I've tried that have NOT solved the problem:

• 3 other Nvidia cards have been tried. Some made it worse, none made it better.
• Many different configurations of nVidia and CUDA drivers, including ones recommended for my exact build of Yosemite.
• Cloned my boot drive (SSD), and ran an OS repair. All it did was make my nVidia drivers invalid.
• Swapped in the cloned SSD and started running on that. Drive is slightly larger, and plugged into a different bay, just in case. No change, so it's definitely not a hardware issue with the boot drive.
• PSU has been tested and found good, though not replaced.

I'm desperate to find a fix, and just about out of ideas. Upgrade to El Capitan and pray...?

Hard to say. System has been stable for a couple of weeks, and I don't want to jinx it. If past experience is a guide, it could start crashing again any day.

That said, the most recent thing I tried came from stumbling onto a very old thread about a related error message. It suggested moving the graphics card to slot 2, as slot 1 in the 2008 MacPro was a bit flaky.

Did something that simple do the trick? I don't know yet. But maybe worth trying?

No new crashes yet, but that of course could change at any time ; )

I turn my Mac off every night, so that would statistically bring down the frequency of crashes, just by being on for fewer hours a week.

Interesting to hear that the slot change had an impact! This issue is rare enough that I suspect it's a confluence of multiple issues, and that might be why it's so hard to find a clear fix. Card slot seems to be PART of the problem, but not the whole deal.

I don't remember if it came up earlier - Are you running NVIDIA's web drivers, most specifically the correct ones (including CUDA) for your OS? Getting my system to its current state took a motherboard swap, slot shift for the graphics card, AND the correct drivers. Other things that may have helped included a fresh battery on the motherboard, a whole lot of cleaning and resetting of PRAM/NVRAM...