System experienced unexpected shutdown following CPU upgrade.
System experienced unexpected shutdown following CPU upgrade.
A little information first and foremost. I recently scored a buddy discount on an RTX 3090 which has run well in my system for a week or so. To match the specs a little more closely I also decided to upgrade my 5600X to a 5900X. The CPU showed up, I quickly swapped it into my system and ran 3DMark to see how it compared. As expected it ran great! ... Right up until it didn't. Under instances of sudden load (so far as I can determine) my PC will suddenly reset. Sometimes with a BSOD, sometimes without; instead just throwing my straight to my BIOS splash screen. So I began the usual troubleshooting. Update Windows. Update Drivers. Update BIOS. Disable XMP. Remove any OC. Check HDD and SSD health with CrystalDiskInfo. Try to narrow down exact cause of the crash. The initial crash came when swapping audio devices in Discord. I assumed it was an audio driver freaking out. I restarted and was fine for another 5-10 minutes. Then it crashes again. I uninstalled the driver to my audio interface and reinstalled the latest version, as well as reinstalling Nvidia Broadcast. I was stable for a while so I loaded up Ready or Not to play with a buddy. We played through most of a round and I crashed when opening a door. Based on the previous crashes while doing Discord things, I assume it's still the audio driver. I remove Nvidia Broadcast. We then decide to play Hearthstone where I was stable for around an hour. "OK that was the issue" I thought and decided to boot up Halo Infinite. Crash . Restart and start Halo Infinite without joining Discord. Crash . Ok clearly there's something else going on. I check temperatures, they look fine. I unplug everything except Mouse, Keyboard, Ethernet and my Monitor. Crash I then went down the rabbit hole of troubleshooting each of the items listed and the system remained seemingly stable. I booted up and played a bit match in BF2042 to stress the system and it ran fine. I rebooted Infinite with zero changes, and (of course) crash . I kept fighting it for a while longer before giving up disgusted at about 3am. System Specs! Windows 10 Version 10.0.19045 MSI MAG X570 Tomahawk BIOS version 7C84v1B (08/30/2022) Ryzen 9 5900X (latest drivers) Noctua NHD15 Chromax RTX 3090 - Driver 527.26 32GB Corsair Vengeance RGB Pro 3600mhz (XMP DISABLED for testing) Corsair RMx 850w PSU (2021) Before going to bed I checked Windows Event Viewer to try and get an idea of what was happening when it crashed. 99% of the crashes had something along the lines of: Any ideas? I'm at a loss.
I noticed a comparable setup with 5900x and 3090 chips. My HX850 caused OCP issues and forced a reboot. This seems typical for 3000-series GPUs because of sudden voltage spikes. I upgraded my power supply to the HX1200 and the problem was fixed.
I also thought it might be connected to power issues. When setting up the 3090, I realized I was testing the limits of what was safe with an 850-watt PSU. That night, I quickly picked up a 1000-watt model at Best Buy just in case others agreed it was the probable cause.
It doesn’t seem like a stronger power supply resolved the issue. I changed my setup to an EVGA 1000 G+ and it worked fine, even in Halo Infinite—though it crashed after about half an hour in the same way before. This time I noticed a BSOD with an error message: “Edit: I added some logs from the Reliability Monitor.”
The G+ isn't a strong choice and seems to be a lower version of the RM850 you originally had.
It was what Best Buy had available, which isn’t my top pick—though I thought it would work. I’m a bit puzzled because this issue still seems relevant if it were the PSU. My understanding was that the problem stemmed from sudden spikes in GPU load surpassing the PSU’s capacity, roughly doubling its rated power for the card. That would mean the system could handle around 700 watts for brief periods. The maximum draw numbers I found for a 5900x model are about 200 watts. This leaves plenty of room for unexpected surges. Do I possibly misunderstand somewhere? If so, and if a better PSU is needed, I might have to wait a week or more before testing. I’m also confused by your comment about liquid coolers. While they do run cooler than air, the main reason I prefer air cooling is its simplicity and fewer failure points. The fans are rated for around 82.5CFM, which fits your specs. This doesn’t account for the fact that the NH-D15 is a well-regarded cooler known for quality, performance, and reliability. Are you noticing temperature issues rather than power consumption? My max observed temperature is about 76°C—well below what would cause a problem. I’m willing to consider this if it’s about cooling efficiency.
It’s worth considering power as a possible cause. That’s reassuring! The only other option mentioned is an HX1500i. I don’t require much power but could stop using it if needed. There’s another thread suggesting turning off Core Performing Boost, which seems to point toward a CPU issue. Does that seem plausible?