Are your computer getting all messed up?
Are your computer getting all messed up?
Thanks for all your tips so far! Here is a weird update: whenever I turn the intake fans to full speed (not the liquid cooler or GPU fans), crashes happen less often. My setup has a big 360mm AIO liquid cooler on top for exhaust and three smaller 120mm fans in front pulling air in. The case is an NZXT H6 Flow. I crashed about two times this past week, once while idle and another ten minutes later. It's still hard to figure out why because there are no error logs or BSOD messages (except for Kernel 41 showing up in Event Viewer). Here are the temps when my programs open: CPU 37°C, GPU 50°C, RAM DIMM1 37°C, RAM DIMM3 35°C, SSD M.2 48°C, and both SSDs at 25°C and 27°C. Motherboard is 24°C. I changed all my fan settings to start fans up sooner than before, plus ordered new Noctua fans to swap out the front intake and another one in the back for exhaust so I can test if that works. I don't think this problem is just about overheating (especially since I mainly crash when idle and not during stress tests), but if fixing it stops it, then loud fan noise isn't a big deal to me. I'd rather have stability than performance at this point 😅
Okay! So after installing new fans and tuning them so that they're 50% at idle, the issue has completely stopped. I've been running my computer for weeks with absolutely no issues at idle, medium, high, or variable loads. I am also running my RAM at 6000 MT/s. What worked for me is aggressive fan tuning (you can hear them at idle), disabling any type of CPU/RAM enhancements in BIOS (such as PBO and BCO), updating/installing all drivers to their newest versions (even beta versions or ones I didn't need like the WIFI driver). I also did a complete reinstall of Windows, AMD chipset drivers, and Nvidia drivers. Here's what the build looks like if anyone was curious.
My CPU has never gotten hotter than 60 degrees Celsius with my water cooler, but it feels like my RAM is having trouble staying cool. I'm going to try different fan settings soon and see if it crashes when the computer runs slower. I also turned off memory context restore so that every time I turn on the PC, my RAM timings reset from scratch. That takes about 45 seconds to finish.
Because it is cold, I think we are getting not enough power. The same thing you say: there is no point in keeping things that hot when it hurts the battery and makes it less reliable.
I can be sure the computer will work fine if you tweak the fan speeds. Yesterday, when my fans ran at normal settings, the machine crashed in just one hour. Now, the RAM voltages are higher (specifically VDDCR_SOC is 1.235V instead of the old 1.03V), so I am not using very low power anymore. I don't usually run something that uses way too much CPU power, so I never get it to temperatures above 60C. Even though some benchmarks pushed the temperature up to 90C, it didn't crash then either.
Final update: I think things are fixed now! My PC is running smoothly at 5200 MHz. Before, I would get occasional crashes whenever I tried to go faster than that speed. I looked up my RAM on AMD's website and found out the top speed they support for my memory is also 5200 Mhz. Now, my fans run at a normal speed instead of crashing. The little light on the DRAM was actually telling me it was just about the CPU or RAM, not anything else. Changing to another motherboard really helped too, because that one MSI board had failed twice already.