OK. I'm putting this under System because I'm not 100% sure what the issue is, or what category it should go in.
Short version: When playing something graphics intensive (GPU near or at 100%) I occasionally have a problem where the computer goes from on and running the game to IMMEDIATELY off. No messages, no shut-down process. Just one moment I'm shooting zombies (7 Days to Die) or raiders (Fallout 4), and the next moment the computer is silent and off. At that point I count to five, and it reboots and acts like nothing is wrong.
Talked to somebody I know with some Geek Squad experience and he started talking PSU rails and GPU draw.
He's a smart guy. But he couldn't dumb it down to where I could figure out what the issue is, exactly, or what I might be able to do about it.
Is it something I have to change in how my GPU is wired?
Is it a switch in/on the PSU?
For that matter, is it a GPU or PSU hardware issue?
Or is it something else entirely?
My best guess, from what I'm seeing and what he's saying is that when the GPU gets an especially heavy workload, it's causing some sort of auto-fault shut off.
But is that a hardware shut off? Something that can be monitored (to verify) in Windows?
OK the hardware (built this in 2019, have run a couple thousand hours without any real known issues other than this)...
CPU: INTEL i7-8700K
GPU: Radeon RX Vega 64
PSU: Corsair HXI 750
MOBO: ASUS HERO X
RAM: CORSAIR 16GB (2x8GB)
CPU COOLER: CORSAIR H115
WINDOWS 10
I should note that I have Radeon Adrenaline running with GPU load and temperature as an overlay.
NOT running hot when this happens.
Temperature on GPU is consistently below 68C the way I have the temp curves set up.
So...what the heck is this?
I'd like to be able to verify the issue in some way, and fix it.
From what my friend said, it might have to do with the power cables from the GPU "sharing a rail"?
As far as I know the PSU is still set to factory settings.
(I do not have the data cable to the PSU connected. Ran out of internal USB ports and have not wanted to take things apart to add the Commander PRO yet.)
I guess the best answer would be "change the GPU connections from X to Y".
Second best would be "Make sure switch X on the PSU is in Y position".
After that I'm guessing we get to less fun choices like "You need a bigger PSU" (which I do not think is that case---GPU mostly runs without any issues at max/220W, and as I calculated when I built the computer, there SHOULD be about 100W of margin, even with full load and fans running).
Thank you for taking a moment to help me figure this out.
Short version: When playing something graphics intensive (GPU near or at 100%) I occasionally have a problem where the computer goes from on and running the game to IMMEDIATELY off. No messages, no shut-down process. Just one moment I'm shooting zombies (7 Days to Die) or raiders (Fallout 4), and the next moment the computer is silent and off. At that point I count to five, and it reboots and acts like nothing is wrong.
Talked to somebody I know with some Geek Squad experience and he started talking PSU rails and GPU draw.
He's a smart guy. But he couldn't dumb it down to where I could figure out what the issue is, exactly, or what I might be able to do about it.
Is it something I have to change in how my GPU is wired?
Is it a switch in/on the PSU?
For that matter, is it a GPU or PSU hardware issue?
Or is it something else entirely?
My best guess, from what I'm seeing and what he's saying is that when the GPU gets an especially heavy workload, it's causing some sort of auto-fault shut off.
But is that a hardware shut off? Something that can be monitored (to verify) in Windows?
OK the hardware (built this in 2019, have run a couple thousand hours without any real known issues other than this)...
CPU: INTEL i7-8700K
GPU: Radeon RX Vega 64
PSU: Corsair HXI 750
MOBO: ASUS HERO X
RAM: CORSAIR 16GB (2x8GB)
CPU COOLER: CORSAIR H115
WINDOWS 10
I should note that I have Radeon Adrenaline running with GPU load and temperature as an overlay.
NOT running hot when this happens.
Temperature on GPU is consistently below 68C the way I have the temp curves set up.
So...what the heck is this?
I'd like to be able to verify the issue in some way, and fix it.
From what my friend said, it might have to do with the power cables from the GPU "sharing a rail"?
As far as I know the PSU is still set to factory settings.
(I do not have the data cable to the PSU connected. Ran out of internal USB ports and have not wanted to take things apart to add the Commander PRO yet.)
I guess the best answer would be "change the GPU connections from X to Y".
Second best would be "Make sure switch X on the PSU is in Y position".
After that I'm guessing we get to less fun choices like "You need a bigger PSU" (which I do not think is that case---GPU mostly runs without any issues at max/220W, and as I calculated when I built the computer, there SHOULD be about 100W of margin, even with full load and fans running).
Thank you for taking a moment to help me figure this out.