Continuing Quadro K2000M Issues on Thinkpad W530

TL;DR: broken driver, slow hardware, crappy support, driver and hardware creating thermal issues which have potentially bricked my laptop on two occasions now. read on for a more detailed report.

I’m still having issues with this laptop. It’s been three years since I bought it, and it has not once worked correctly under linux.

I am now suspecting that the nvidia driver has also caused some temperature issues, that have caused my laptop to become unstable, and eventually die. I’ve already had the motherboard AND cpu replaced once, and I might need to do it again soon after attempting to get the nvidia GPU working under linux again a few weeks ago. Ever since then, my laptop has started locking up again, and things have started crashing.

I’ve done some burn-in tests under windows, and the results are somewhat better, no lockups, but the gpu drivers crash and reset during GPGPU testing. Temps for the CPU hit over 100c which I find high, but that is within the safe zone (Tcase is 105c, and shutdown temp is 130c), and for the GPU it was over 80c iirc. That doesn’t seem to be fatally high.

On a number of occasions and linux nvidia driver versions I have seen the kernel auto-shutdown as it hit a thermal limit. Usually this coincides with a black screen after grub.

Today I have tried two newer drivers, the latest long lived (352.63) and the latest stable (358.16) release. Neither of them initialize the gpu/screen correctly. Luckily they don’t cause my laptop to do an emergency shutdown, but they render my laptop completely useless. I am left with nothing but the grub “loading…” text and then nothing. The HDD light flashes for a while, so the system is booting, but it likely stops when it gets to trying to startup X. I have no proof of that, but it would seem a reasonable assumption given CTRL+ALT+DEL still works to reboot no matter how long I wait.

I am really quite disappointed with the drivers and support in general. These issues have been reported by multiple people, over YEARS and it’s still not been fixed or even reasonably dealt with. It was acknowledged as “an issue” once, then blamed on the kernel and ignored. The kernel got the “fix” for the “problem”, but it still seems to be broken. Go figure.

I configured this laptop with nvidia graphics because it is supposed to be the best supported discreet graphics under linux. Maybe that is true for Gforce desktop graphics.

I’ll go more in depth and grab bug reports again if they are asked for, but only if I can get some guarantees this will actually get dealt with. 3 years with no support is way too long. It’s absolutely ridiculous. I’ve posted a few different reports that got NO response or any kind of acknowledgement. Why would I bother to go to any trouble if it’s not going to even be looked at?