I'm theorizing here, but if you're a novice PC builder and you buy a bunch of parts and put them together and the system doesn't boot, you're probably going to think you fucked up. You don't have an easy way of testing parts individually. Maybe you replace the motherboard first, and of course that doesn't work, so then you order some new RAM, and so on. If you spend long enough troubleshooting then it's too late to return the CPU.
I don't think I'm exactly a 'novice' system builder*, but when my son's computer died a couple of years back we bought replacement RAM, Motherboard and Power Supply before realising it was the CPU that had died (didn't have spare compatible parts to try out).
At least we've got spare parts if the other components die now...
In my 25-odd years of building and using PCs (and the odd few servers) I'd not had a CPU die on me, so it was the last thing I suspected. Wasn't until after we'd worked it out that we looked it up and found that it was somewhat a systemic problem with these specific CPUs after ~1 year of use. I was determined it couldn't have been the CPU - until it couldn't have been anything else.
*Although in hindsight I may have acted like one.
Extra story: 20-something years ago I had a slightly misaligned heatsink on AMD CPU which caused the machine to crash after a few minutes of gaming. I took the heatsink all the way off and decided to put my knuckle on the die and boot up the machine. It took what felt like a millisecond for the die to feel 'nuclear' hot and the machine, thankfully, auto-switched off. I've had a LOT of respect for the job heatsinks do since then.
Once the heatsink was properly seated, the machine booted up no worries, and that CPU kept going until it was system upgrade time. Bulletproof!
I took the heatsink all the way off and decided to put my knuckle on the die and boot up the machine. It took what felt like a millisecond for the die to feel 'nuclear' hot and the machine, thankfully, auto-switched off.
You're lucky you didn't get a severe burn on your hand or kill the CPU --- early AMD CPUs, notably from the late 90s / early 2000s, were infamous for catastrophically overheating without a heatsink. There's a video from Tom's Hardware showing that, after which AMD started adding more thermal protection.
(There are various assertions floating around that that video was fake, but given that no one seems to have made a "myth-busting" video in the 20+ years since it was released, despite the clear incentives for YouTubers to do so, and instead others have shown how hot those CPUs can get, I suspect it's real. Now that those CPUs and other hardware from that era are actually somewhat collectible, the likelihood of someone trying to repeat that video is even lower.)
I did think I had killed the CPU at the time, but being an impoverished student at the time I had no choice but to see if it was OK - and thankfully it was.
The heat was quick and intense enough that I took my finger off very quickly - it may have actually overheated and shut down faster than I took my finger (knuckle of my pinky) off, which might have saved me from a more lasting burn.
Thank you for sharing this story! I've long wondered what would happen without a heatsink, but have been too chicken to try it myself. It's a bit of a curse to learn best through destructive means :-)
