Diagnose my MacPro!

I've got a circa 2013 MacPro (purchased January 2014). It's still my daily driver at work and still a great machine. Except when it isn't. I've had to hold down the power button to reboot it several times over the past few weeks when I get the "hypnowheel" mouse and the rest of the UI is unresponsive.

What's the current state of my diagnosis?

- This Mac has the low-end D300 video card, which wasn't covered by a recall that Apple did for other graphics cards. I never had these problems in the past.

- When it's in the "frozen" state, the ssh daemon is still up. I can log in, run stuff, see all my files.

- I was concerned that I might be looking at disk failure issues with my JBOD disk array. According to Disk Utility, all my SMART things are "verified" and I see nothing in the logs to suggest failures. Of course, modern disks go out of their way to retry and retry.

- I noticed that my Time Machine stripe wasn't connected this most recent time. This used to be a big and annoying problem early in the computer's life, but as Apple upgraded things, the stability of my disk mounting (via Thunderbolt 2) seems to have stabilized.

- My CrashPlan daemon, which I run on this machine, seems to be crashing. I'm guessing this is related to the terminal decline of their service. I doubt it's related to the wedging of the computer.

- Both recent freezes have been related to using Chrome. Scrolling, looking around, boom.

- The most recent freeze, I tried a thing I got from the Internet: unplugging and replugging my monitor. Black screen.

- I just did the PRAM/SMC reset. For grins.

Based on all of this, my current operating theory is that my video card is dying and/or recent Chrome upgrades are pushing on it in ways that tickle bugs in the driver or firmware. Supposedly, there's a sequel on the way, but it's not here yet. The iMac Pro isn't particularly attractive to me because I have a nice monitor already and would just want to swap out the computer part. The tentative path forward is to just suck it up and deal with weekly reboots. Not very attractive.

My brand new mac wedged sometimes when Chrome would hit any page that used WebGL.

https://apple.stackexchange.com/question...ser-on-mac-os-x

See if that stops the wedging as it relates to the Chrome web browser.

Updates: Since doing the PRAM / SMC reset, I'm now seeing graphics glitches when scrolling. Sometimes, but far more often than beforehand. I've had these issues even with Chrome not running. At one point, while coding in IntelliJ, my screen blanked out for two seconds, and now it's back again.

My Chrome browser doesn't seem to have any way to disable WebGL. I do have the "WebGL Draft Extensions" disabled, which seems to be the default.

Very disturbing. I fear that my computer is crapping out as I'm using it. I really, really wanted to wait until a MacPro sequel came out. I might be forced to act sooner than that.

I'm not sure about macs, but if this were a PC, I would immediately go to "bad VRAM" here.

And look at how much fun it is to replace the graphics cards. And that's assuming I can source the right cards.

Apple seems to have some hardware diagnostic tests you can run, google for that. Perhaps it can confirm/deny the idea that it's the video card(s).

Interesting. My Mac is new enough that you simply power up and hold down the 'D' key. Unfortunately, the report, after six minutes of waiting, is "No issues found." ("Reference Code: ADP000", which maps again to "no issues found.")

Sigh.

I assume you've ruled out cooling? That is: you gave it a good burst from an air-duster, right?

My PC started throttling down. Eventually I opened the case (to fit an NVMe drive) and discovered that the radiator for the water cooling had enough dust caked on it that if I'd combed and spun it, I could have made a reasonable sized sweater.

Hmm. Cooling. That's something worth investigating. This computer has one monster fan for the whole thing, currently spinning very slowly, but then I've recently cleaned enough dust bunnies from behind my monitor that there's every reason to include dust as a possible cause of the problem. Question to the gallery: how far into the disassembly process would you go? Getting to the point where you've got access to the internal heat sink from top and bottom is non-trivial.

Meanwhile, after running the built-in diagnostics, my computer's been much better behaved. No video glitching. We'll see if this holds up. I'm at a loss for understanding *why* a series of diagnostic tests could have positive side-effects on my computer, but... perhaps they did.

Install iStats Menus, that will give you detailed temperature readouts from sensors all over the mainboard, gpus etc

Then you can see if anything is running hot.