Stronk
Level G
So here's a strange problem I'm having with my desktop with the following specs:
Ryzen 3600XT
Vega 64 Liquid (Power limited to 250W)
2x16GB 3000 Corsair LPX
1x8GB 3200 Corsair LPX
(All 3 sticks running at 3266Mhz @ 1.38V)
MSI B450M Mortar (Non-Max)
Corsair CX550
Now, I'm having shutdowns under random loads. I was using Manjaro till a few days ago, and kept getting segmentation faults or just hard shutdowns while compiling a dependency, namely rocfft (via yay). It either showed a segfault, or just shutdown, at any percentage between 7-10. Temps were mot a problem as they never crossed 80C.
To narrow down the problem, I first ran RAM at stock speed (2133) to see if that was the issue, but still ran into the problems. My CPU was also overclocked so I thought could be it , although highly unlikely. So I did a cmos reset and yet again the same segfault and hard shutdowns, and that left mobo as the only culprit. But then, I'm not sure why, I locked CPU clocks to 3Ghz, kept RAM at regular 3266 speed and voila - it compiled without a hitch!
So I concluded then that the CPU is faulty, and almost filed an RMA request with AMD but then below happened today.
I switched to Ubuntu 20.04 today, and trued running a benchmark which puts medium-high compute (NOT 3D graphics) load on the GPU and occasional single thread 100% load on the CPU, and I could never finish the benchmark because it just kept shutting down during the benchmark (ai-benchmark 0.1.2 in case this is relevant), and this time even downclocking the CPU doesn't seem to help. I ruled out GPU by even power limiting it to 150W, but still kept getting hard shutdowns. The CPU never went above 40W of power consumption, and the GPU never crossed 60C in Temps or the set power limit.
And the weird part about these shutdowns is that the computer wouldn't immediately switch on (by pressing power button) once it shut down, but instead I had to switch the mains off and on (either at plug point or on the PSU itself) to get it to power on, ie, I couldn't immediately switch it on once it shutdown without power-cycling.
I've tried running GPU-only stress tests and benchmarks and the computer didn't shutdown, have run CPU-only benchmarks and no shutdowns either so I'm really flummoxed as to why it is happening.
Any ideas? Does the weird shutdown behavior point to a PSU problem, or is the CPU still likely at fault? Any replies will be appreciated, thank you.
Ryzen 3600XT
Vega 64 Liquid (Power limited to 250W)
2x16GB 3000 Corsair LPX
1x8GB 3200 Corsair LPX
(All 3 sticks running at 3266Mhz @ 1.38V)
MSI B450M Mortar (Non-Max)
Corsair CX550
Now, I'm having shutdowns under random loads. I was using Manjaro till a few days ago, and kept getting segmentation faults or just hard shutdowns while compiling a dependency, namely rocfft (via yay). It either showed a segfault, or just shutdown, at any percentage between 7-10. Temps were mot a problem as they never crossed 80C.
To narrow down the problem, I first ran RAM at stock speed (2133) to see if that was the issue, but still ran into the problems. My CPU was also overclocked so I thought could be it , although highly unlikely. So I did a cmos reset and yet again the same segfault and hard shutdowns, and that left mobo as the only culprit. But then, I'm not sure why, I locked CPU clocks to 3Ghz, kept RAM at regular 3266 speed and voila - it compiled without a hitch!
So I concluded then that the CPU is faulty, and almost filed an RMA request with AMD but then below happened today.
I switched to Ubuntu 20.04 today, and trued running a benchmark which puts medium-high compute (NOT 3D graphics) load on the GPU and occasional single thread 100% load on the CPU, and I could never finish the benchmark because it just kept shutting down during the benchmark (ai-benchmark 0.1.2 in case this is relevant), and this time even downclocking the CPU doesn't seem to help. I ruled out GPU by even power limiting it to 150W, but still kept getting hard shutdowns. The CPU never went above 40W of power consumption, and the GPU never crossed 60C in Temps or the set power limit.
And the weird part about these shutdowns is that the computer wouldn't immediately switch on (by pressing power button) once it shut down, but instead I had to switch the mains off and on (either at plug point or on the PSU itself) to get it to power on, ie, I couldn't immediately switch it on once it shutdown without power-cycling.
I've tried running GPU-only stress tests and benchmarks and the computer didn't shutdown, have run CPU-only benchmarks and no shutdowns either so I'm really flummoxed as to why it is happening.
Any ideas? Does the weird shutdown behavior point to a PSU problem, or is the CPU still likely at fault? Any replies will be appreciated, thank you.
Last edited:
