💾 Archived View for hyperborea.org › log › 2021-05-30-overload.gmi captured on 2021-12-03 at 14:04:38. Gemini links have been rewritten to link to archived content
➡️ Next capture (2022-01-08)
-=-=-=-=-=-=-
I finally got around to trying out No Man’s Sky a few weeks ago. I started on a super-hot planet, where you need to find shelter and/or resources to recharge your suit’s hazard protection system to keep cool. Got killed a few times trying to figure out what I was doing. And after about 20 minutes, my computer spontaneously shut itself down.
I waited a few minutes to let it cool down, then tried again. Managed to figure out a bit more of what I needed to do in the game, and then the same thing happened.
Fine, I tuned the graphics detail downward. This time it stayed up, I was able to make a little progress fixing the ship, and actually save the game without it crashing.
But over the last few weeks, the computer started doing it again, even at the lower settings — and in other games that I’ve been playing for longer without any trouble, like Minecraft Dungeons.
Clearly something was broken. So I had to figure out what it was.
OK, this isn’t a super-powerful gaming rig. It’s got plenty of RAM and a decent enough graphics card (which I upgraded last summer, in a case of lucky timing, just before the chip shortage caused prices to skyrocket), but the CPU and motherboard are about five years old. And they weren’t cutting edge at the time. Even so, the system usually runs really smoothly, so I haven’t felt a need to upgrade the processor. And the games ran just fine, right up to the moment the computer shut off.
So: Insufficient power supply? Maybe. But it wasn’t shutting down randomly, and it wasn’t shutting down immediately (like when I installed a new component and suddenly it would power down when I tried to burn a CD), it was shutting down after sustained stress.
Most likely: Cooling. I installed Core Temp and ran it while I played some Minecraft Dungeons. As the fan sped up, the CPU temperature climbed up to around 70 C. I checked the max operating range: 61 C.
CPU World has tons of stats and specs on various processors.
So I opened up the case, blasted a year’s worth of dust bunnies out of it, hooked everything back up and ran the game again with the case open so I could check to see that all the fans were running. And so I could see whether the temperature was holding or not.
Nope. Well, the fans were running fine, but the temperature was still climbing into the danger zone.
OK, so I didn’t need to replace the case fan, or the CPU fan. Did I need a bigger heat sink? Not likely, since it was overheating on games that had been running fine before. I wasn’t seeing any signs of memory or processing errors, just the shutdowns, so the actual processor, motherboard and RAM were probably still okay. Which of course might change if I didn’t fix this!
But the thermal paste that conducts heat from the processor to the heat sink was also five years old. I could replace that easily and see if that did the trick.
At first I was worried that I’d have to order it online and wait for shipping, since with Fry’s gone, the only electronics stores around here are Best Buy and Apple (not likely to help here), and I haven’t even looked to see what low-level components Best Buy even carries in years. But it turns out they do carry thermal paste, and since I was already on the website, I just placed a store pickup order and went to pick it up when they opened. (I think it was the first time I’d actually walked into a Best Buy since before the pandemic hit. Mostly I’ve been doing curbside pickup or ordering shipping when I need electronics.)
So I removed the heatsink, which took more force than I expected since it was still sticky. Then I wiped the old paste off of both the heat sink and the CPU using a microfiber cloth left over from a phone screen protector and some rubbing alcohol, then waited for it to dry. Once it did, I squeezed a pea-sized lump of paste onto the middle of the processor and reattached the heatsink.
With the case open and everything hooked up again, I was able to run Minecraft Dungeons without breaking 50 C on the CPU even when surrounded by zillions of mobs. The fan didn’t even speed up! So I closed it up again, and so far things seem to be working fine again. I’ve run a couple more Dungeons levels (including a daily challenge with the Fiery Forge, in which at one point you need to overload the cores!) without the processor breaking a sweat, so I think I can stop monitoring the temperature now.
And No Man’s Sky is running fine now too!
But I think I’ll keep the lower graphics settings for now. Just in case.
—Kelson Vibber, 2021-05-30
Previous: 💉💉 Double-Shot
Next: 🍎 What's up with MacOS Updates?
Cross-posted at K-Squared Ramblings