A riddle for the really smart IT/Windows/PC folks

This is the weirdest one I've ever had happen.

The computer is a Lenovo Erazer X315.

This is a graphics issue. The motherboard has an AMD R7 GPU on it, which I've disabled in the BIOS.

The computer also came with with a PCI-E GPU card, with an AMD R9-255.

Last night, and again tonight, both displays blanked out briefly, then came back on their own. When I looked in Device Manager, the PCI-E card was identified as an AMD R7, not an R9.

I told Win10 to update the driver. It went on the net, found a new driver, and installed it. Poof, Device Manager identifies it as an R9-255 again!

Twice. Same problem, same solution. I've seen plenty of GPU crap out in my day, but I've never seen one fail in such a manner that Windows mis-identifies the chip and downgrades the driver... only to fix the problem by telling Windows to update the driver.

I suspect the graphics card is on it's way out. Not quite 2 years old, but nothing lasts anymore.

What say the hive mind?

edit: In thinking about it, both reversions happened when I was doing something reasonably graphics intense. The first was watching YouTube videos full-screen, the second while running a gaming client called Firestorm. I'm thinking that the card might be overheating and deliberately "underclocking" to a lesser level, which windows is seeing as a lesser GPU.

In a misleading bit of marketing, the OEM only R9-255 uses the same GPU as the R7-250X, so a generic driver would see it as an R7.

I think you've got it figured out, the computer is misidentifying the card and changing drivers on you. This is a driver bug (Edit: Or possibly a firmware bug) which has to be fixed by the manufacturer. As a workaround, I have 2 ideas:

• Try running the R7 driver and see how it goes. If there are no problems or loss of performance, stick with it for now.
• Look in the card's driver configuration and see if there are GPU clock control options, try disabling automatic clock control, lock it to the highest non-OC speed if possible. This will increase power consumption and waste heat production though.

So this thought is almost entirely from left field, but it's a thought.

I have a Lenovo laptop, about 6 months into ownership the screen started getting yellow, then black spots, then no output. Acting like it had been dropped or smacked with something hard.

Turns out, much like the ps3, the built in nvidia graphics card would overheat to the point it was melting the connection to the board.

Still under warranty, Lenovo fixed it no questions asked.

Not sure if it would apply in the least, as yours does sound like more of a driver problem, but it is something to keep in mind if driver changes don't work.

Filth.

The answer is "filth".

I took the whole box out back and blasted the snot out of it with canned air.

Downloaded some GPU-beating stress tests and ran it as hard as I could. Got the GPU temperature up to 78C. No crashes, no dropped drivers.

Don't be a moron like me. Clean your computer now and again!

First time I've heard of dirt causing a problem other than overheating or mechanical jamming!

In reply to GameboyRMH:

Well, yes. I think the card was overheating and clocked itself down into a protection mode, causing Windows to think it was a different GPU.

Nope, that wasn't it. Just happened again. Did a full uninstall of all the AMD software/drivers, downloaded a fresh one, and installed it. Let's see what happens next.

EvanR wrote: Nope, that wasn't it. Just happened again. Did a full uninstall of all the AMD software/drivers, downloaded a fresh one, and installed it. Let's see what happens next.

Yep this. You need to remove the drivers completely and install only the one you want.

I ran in to a kind of similar issue with a driver for a plotter.