Ryzen 5900x (stock)
MSI Mortar B550m
Corsair Vengeance Pro 32GB DDR4 3600Mhz
Galax RTX 3060Ti Ex (stock)
CM Gold 650w PSU
NZXT x53 240MM AIO
I have this problem with my system whenever I play games or use a demanding application I get a BSOD at stock settings with only XMP on for the RAM. But when I clock the CPU down to 3700Mhz at 1.2v I am able to game fine without any crashes. The temps at stock settings are between 75-85c and of course its lower at 3.7Ghz between 70-80c.
VSOC voltage to 1.1 v
and share hwinfo outputs normal and while running prime 95
check will you be able to replicate it on your machine
I already tried doing that, it does make the system perform better but the issue still remains.
Also I could only get a -10 and not a -30 like he gets in the video. Maybe AMD sent him a gold sample for review.
Have you checked whether any thermal throttling is triggered in stock setting, where you get BSOD?
any warning set based on temperature in BIOS ?
Was having a similar problem, where system would shutdown or just freeze randomly while doing heavy CPU based tasks. Weirdly, clocking my 3600XT down to 3ghz fixed it. Eventually, PSU turned out to be the issue
check the dedicated 8 pin connector for cpu power whether it is loose from psu to motherboard socket
I think if it was. System wouldn’t turn on would it?
Oh. That’s what I am thinking too, is 650w gold enough for power?
Should be enough, but PSU quality doesn’t depend on 80+ certification. Gigabyte’s 80+ gold PSUs have proven to be quite bad in terms of reliability and general use. What is the exact model of your PSU, and how old is it?
I bought new in Dec 2020, built the whole system actually. CM 650W MasterWatt Modular PSU.
650W is far enough for this build…check whether the cpu temps exceed > 90 on issue, also check for MOS temps !! is your AIO configured in SW or BIOS ?
AIO is configured in Nzxt cam and also in bios. How do I check the MOS temps?
That’s seems to be more than adequate actually, and shouldn’t be causing any issues. Yea would be a good idea to check for Mos and VRM temps, you can see them in hwinfo under the motherboard group. Keep hwinfo on while gaming, and see what Temps the motherboard temp readings peak at while doing tasks which crash your pc. Also in case you haven’t already tried, try using ram with and without xmp, and consider running windows’ built in memory checker tool
What is the BSOD error code?
Did you check the error code in windows event viewer? If you haven’t, search “event viewer” in start, go to windows log>system & you should be seeing some critical errors. Check this after reboot whenever your system crashes, it might hold some clues.
Also, what cooler are you using, are you sure about your thermal compound application? Is this a new system & the issues have been present since day 1 or is it a couple of months old & you have recently started seeing this issue? I am sure you are already using the latest chipset drivers.
Apart from checking all the above stuff, you should also stress one part of the system at once, like GPU though heaven benchmark & CPU through something like cinebench. It might be possible that it crashes only in one scenario which might help in identifying the problem. Do keep in eye on the temps as well through something like HW monitor. There could be multiple things at play here among the following & we should try to identify the issue in this order:
Update the chipset drivers as well as reset BIOS to optimized settings.
Reconnect all the connectors & possibly try to run the tests with everything on an open bench with just the stock drive & no extra stuff if possible.
Test 1: Stress CPU using cinebench, see if you are able to finish this test or not & also keep an eye on the max temps(VRM, Tdie, clock speeds etc.)
Test 2: If Test 1 fails, try running the same test with your stable settings.
Test 3: Stress just the GPU using heaven or something which is very light on the CPU but stress the GPU to 100%. See if it is stable or not, the max hot spot temperatures also the memory temps if your GPU has that sensor.
Test 4: If Test 3 fails, try running the same test with frames limited to 60Hz & see if that passes.
Based upon this result of these tests, it could be an issue with:
CPU temperature, will be observed in Test 1 & 2.
GPU temperature, will be observed in Test 3 & Test 4. At first glance we might think that it is a CPU related issue as lowering CPU voltages fixes the issue but it is possible that lowering CPU voltages is making you CPU bound in games & your GPU utilization is not full. If there was an issue with GPU temps, it might not shop up in this scenario.
If all tests are passing separately, it might be an issue with your power supply. While the wattage of the power supply is more than enough, although I am not sure as it also depends on what else have you hooked up in your system, these parts draw high current & are prone to show up even in slightest of imperfections in the PSU.
I’ll try to get this done on the weekend, will keep you posted. Thanks for the help.
