PC shutting off under intense load

[Pud of Mud]

Specs:

• AMD Ryzen 7 5800X w/ be-quiet! cooler
• 32 (4x8 3200mhz) GB CORSAIR RAM
• 1 TB CT1000P3SSD8 (Among other storage drives)
• AMD Radeon RX 6600 XT
• CORSAIR RM 750x 80+ Gold
• 3 Noctua case fans
• Windows 10 64-bit
• GIGABYTE B550M DS3H BIOS version F15
• No overclocks, modifications, etc.

 

Hey ltt forums! first time posting here; please tell me if I'm doing anything wrong on the platform and I'll attempt to fix it.

 

The issue is my PC randomly shuts off during intense load. Not consistently, either; without warning, the monitors shut to black and shortly after the USB peripherals disconnect and shut off. Its very abrupt and only has occurred during Minecraft Java with intense shaders and during Microsoft: Flight SimulatorGPU-Z Sensor Log.txt. In an attempt to debug the issue, I've recorded a log using CPU-Z during the crash which is attached below as a .txt file (Crash occurred 2024-03-10 17:23:31 within the log). From what I've seen online, it's either a power supply issue or a heating issue; however, I've found neither should be the issue from what I've seen so far. Any solutions?

 

[Robchil]

Welcome

 

No solutions, only considerations... 

 

since what you are explaining sounds like it just cuts when the PSU can't deliver any more.. 

 

not that it should deliver any less than earlier. BUT... PSU's deteriorate over time, have you tried another PSU? 

 

How long has this been happening? 

 

[Godlygamer23]

Does the issue occur in any other applications, such as Superposition or GravityMark?

[Pud of Mud]

8 hours ago, Robchil said:

Welcome

 

No solutions, only considerations... 

 

since what you are explaining sounds like it just cuts when the PSU can't deliver any more.. 

 

not that it should deliver any less than earlier. BUT... PSU's deteriorate over time, have you tried another PSU? 

 

How long has this been happening? 

 

It has been happening for a few months now; starting around January. Along with this, the PSU isn't that old (~1.5 years) and shouldn't be deteriorated at that age, right?

 

8 hours ago, Godlygamer23 said:

Does the issue occur in any other applications, such as Superposition or GravityMark?

It hasn't occurred in other apps, even during intense VR it hasn't happened. Even when I try and recreate the shutdown through stress-testing maps on Minecraft it doesn't happen; it's seemingly random during play.

[Robchil]

5 hours ago, Pud of Mud said:

It has been happening for a few months now; starting around January. Along with this, the PSU isn't that old (~1.5 years) and shouldn't be deteriorated at that age, right?

 

It hasn't occurred in other apps, even during intense VR it hasn't happened. Even when I try and recreate the shutdown through stress-testing maps on Minecraft it doesn't happen; it's seemingly random during play.

i looked at your log... your GPU hotspot is reaching 114C before the log dies... that's not normal..   as far as i've googled data on the RX 6600 XT, it seems the GPU thermal limit is at 115C... i would assume it hit it's thermal limit and shut down your system. 

 

VR btw.. isn't realy as demanding as other high res games as VR resolution is lower.. 

 

 

[Pud of Mud]

32 minutes ago, Robchil said:

i looked at your log... your GPU hotspot is reaching 114C before the log dies... that's not normal..   as far as i've googled data on the RX 6600 XT, it seems the GPU thermal limit is at 115C... i would assume it hit it's thermal limit and shut down your system. 

 

VR btw.. isn't realy as demanding as other high res games as VR resolution is lower..

I've continued to log data and kept testing and found that the hotspot as reached 115 within the new tests but it doesn't shutdown. I've used DDU and reinstalled drivers as well as increasing the fan curve within the drivers. So far with my stresstesting I've found no crashes sense, but due to the random nature of this issue I can't say it's fixed yet. I also question using the 'hotspot' rather than the regular GPU temp as when I use the driver-based statistics its only reporting the same number as the regular GPU temp. I really don't know though.

 

Also, my VR headsets resolution is much higher than my primary monitor; as I game on 1080p. So I'd assume running VR on ultra settings for things such as flight simulator would be more demanding than non-VR flight simulator.

[Robchil]

I would disassemble the card and repaste it, and check if the pads need renewal too. A so called GPU service. 

 

It is less than 3 years but plenty use can cause dry paste too. 

 

unless you OC'ed it hotspot should stay below 80C with fresh paste and pads that fits. 

 

[QuantumSingularity]

99.9999999% certain you have poor contact with the die on AT LEAST one of the corners. Average GPU temp is normal at 71°C, but that 115°C hot spot is not. At least not when your average temp is so low. When you disassemble the GPU you will probably be able to see the issue with your own eyes. Repasting a GPU btw is a bit different than CPU. You need to spread the paste evenly across the die, every single millimeter. Don't wait on just the mounting pressure from the cooler. 

[BoneS71]

I will make a new thread for my issue

Edited March 11 by BoneS71
I'm new and didn't mean to hijack a thread

[Robchil]

2 minutes ago, BoneS71 said:

Critical    3/11/2024 10:55:44 AM    Kernel-Power    41    (63)

It just means your system shut down due to power was cut. 

you can try having your system on and flipping the power switch on the PSU, or just janking the powercord out of it, should be the same error msg from that. 

 

i would think your system sent a cut power signal to your PSU when the GPU max thermal limit was reached, but since you say you tried it several times and it reached 115C and didn't shut down it's hard to say.. service your GPU anyway since it's not good for it to reach thermal limit. 

 

 

[QuantumSingularity]

3 minutes ago, BoneS71 said:

My System hardware. is as follows:

Windows 11 PRO

ASUS ROG Strix B550-XE Gaming WiFi AMD AM4 ATX Motherboard

AMD Ryzen 9 5900X - Ryzen 9 5000 Series Vermeer (Zen 3) 12-Core 3.7 GHz Socket AM4 105W None Integrated Graphics Desktop Processor

G.SKILL TRIDENT Z ROYAL SERIES 32GB (2x 16GB) PC4-28800 DDR4-3600MHz

Asus ROG Thor 1200W Platinum, 80+ Platinum Power Supply

ASUS ROG Strix LC Radeon RX 6900 XT TOP OC 16GB GDDR6 Graphics Card

ASUS ROG Ryujin II 360 ARGB All-in-one Liquid CPU Cooler 360mm Radiator

Corsair iCUE 5000T RGB 

I Adjusted the TPM from the default to TPM I and my system stayed up longer that before I updated the BIOS to the lates for my MB. I can be watching reviews of LTT, JayzTwoCents or Gamers Nexus and al of a sudden I have now volume. may Task Manager has frozen and I have to push the reset button.

 

I all of my 33 years of buulding computers I have never ran into this issue or given this code:

 

Critical    3/11/2024 10:55:44 AM    Kernel-Power    41    (63)

 

I have no idea where to go from here either,

 

Tim

The error in the log is normal when you restart w/o properly shutting down. With the formentioned issues though, i would start with memtest, since it's very likely to be a memory error.

[Robchil]

8 minutes ago, BoneS71 said:

My System hardware. is as follows:

Windows 11 PRO

ASUS ROG Strix B550-XE Gaming WiFi AMD AM4 ATX Motherboard

AMD Ryzen 9 5900X - Ryzen 9 5000 Series Vermeer (Zen 3) 12-Core 3.7 GHz Socket AM4 105W None Integrated Graphics Desktop Processor

G.SKILL TRIDENT Z ROYAL SERIES 32GB (2x 16GB) PC4-28800 DDR4-3600MHz

Asus ROG Thor 1200W Platinum, 80+ Platinum Power Supply

ASUS ROG Strix LC Radeon RX 6900 XT TOP OC 16GB GDDR6 Graphics Card

ASUS ROG Ryujin II 360 ARGB All-in-one Liquid CPU Cooler 360mm Radiator

Corsair iCUE 5000T RGB 

I Adjusted the TPM from the default to TPM I and my system stayed up longer that before I updated the BIOS to the lates for my MB. I can be watching reviews of LTT, JayzTwoCents or Gamers Nexus and all of a sudden I have now volume. My Task Manager has frozen and I have to push the reset button.

 

I all of my 33 years of buulding computers I have never ran into this issue or given this code:

 

Critical    3/11/2024 10:55:44 AM    Kernel-Power    41    (63)

 

I have no idea where to go from here either,

 

Tim

make your own thread.. don't hijack. 

[BoneS71]

14 minutes ago, Robchil said:

make your own thread.. don't hijack. 

Sorry I didn't mean to. I didn't know how to start my thread. I'm new to this sort of help forum. I'll delete it and try again

[Robchil]

1 minute ago, BoneS71 said:

Sorry I didn't mean to. I didn't know how to start my thread. I'm new to this sort of help forum. I'll delete it and try again

It's a big red button saying "start a new topic" at the top of the threads

 

[BoneS71]

I thought I just deleted this thread of mine

 

[BoneS71]

I found it. Thank youI think I need to stop using such a big TV and so close to me.LOL

[Pud of Mud]

1 hour ago, Robchil said:

I would disassemble the card and repaste it, and check if the pads need renewal too. A so called GPU service. 

 

It is less than 3 years but plenty use can cause dry paste too. 

 

unless you OC'ed it hotspot should stay below 80C with fresh paste and pads that fits. 

 

1 hour ago, QuantumSingularity said:

99.9999999% certain you have poor contact with the die on AT LEAST one of the corners. Average GPU temp is normal at 71°C, but that 115°C hot spot is not. At least not when your average temp is so low. When you disassemble the GPU you will probably be able to see the issue with your own eyes. Repasting a GPU btw is a bit different than CPU. You need to spread the paste evenly across the die, every single millimeter. Don't wait on just the mounting pressure from the cooler. 

Thank you both! I'll try doing a GPU service. I've got some thermal paste lying around for a CPU, that should work; yes?

[Godlygamer23]

23 minutes ago, Pud of Mud said:

Thank you both! I'll try doing a GPU service. I've got some thermal paste lying around for a CPU, that should work; yes?

Thermal paste is only thermal paste - there is no 'CPU' or 'GPU' thermal paste. Depends on the application on where you would use it sometimes, but in most situations, it's likely fine whatever you have. 

 

Can you run Superposition and log the whole benchmark run? I wanna see how the GPU behaves, and certainly post here if it crashes. 

[Pud of Mud]

47 minutes ago, Godlygamer23 said:

Thermal paste is only thermal paste - there is no 'CPU' or 'GPU' thermal paste. Depends on the application on where you would use it sometimes, but in most situations, it's likely fine whatever you have. 

 

Can you run Superposition and log the whole benchmark run? I wanna see how the GPU behaves, and certainly post here if it crashes. 

I've applied new thermal paste to the GPU and reinserted it into my system, than ran Superposition and logged it. GPU-Z Sensor Log - Superposition benchmark.txt

 

I didn't see the hotspot go above 101 and the average seemed lower than before (I'm still unsure what a normal hotspot temp is). I also ran the same world/scenario from the crash on Minecraft from the original log post-new thermal past and logged it.GPU-Z Sensor Log - mienrcaft after.txt

 

Funnily enough, I haven't had a crash sense posting here in the first place!

 

[Robchil]

17 minutes ago, Pud of Mud said:

I've applied new thermal paste to the GPU and reinserted it into my system, than ran Superposition and logged it. GPU-Z Sensor Log - Superposition benchmark.txt

 

I didn't see the hotspot go above 101 and the average seemed lower than before (I'm still unsure what a normal hotspot temp is). I also ran the same world/scenario from the crash on Minecraft from the original log post-new thermal past and logged it.GPU-Z Sensor Log - mienrcaft after.txt

 

Funnily enough, I haven't had a crash sense posting here in the first place!

 

it hasn't passed 106C i see... that sounds better, i still think it is a  bit high tho. where your pads still soft? should be, you'll notice if they are dried out. 

after the first heat cycle i would tighten the cooler on the GPU again, and test again. 

 

there are thermal paste and liquid metal..  i see some use liquid metal on GPU's, but anyone fixing GPU's would not recommend that, as it's easier to contain liquid metal under the ICS(metal cap) on a CPU than on open GPU core.. that even have resistors ON the core so you should mask it if you do.. anyway.

 

Liquid metal is as conductive as solder, since it sort of is like a solder, and even dissolve solder... so if you spill it you need to get it off realy fast, and if you get it under the core or chips, it should be delivered to someone with the skill to remove the chips to clean it and reball and mount them back on. 

 

I've seen some thermal paste conductive too, but most are not. 

 

[QuantumSingularity]

Temps are a bit better, but 35°C delta between average and hot spot is A LOT. With the 6600Xt you should be looking at 65-70°C average and about 80°C hotspot.  Are you sure you spread the paste evenly across the whole die? Don't be afraid if a lil bit squishes or falls from the side, it's electrically non-conductive. What paste did you use? Arctic's MX-4 and MX-6 do a perfect job for GPUs if you don't wanna blow your cash on the more expensive stuff. Also check if you pads are put correctly and aren't forming a bulge. When screwing down the cooler back on the GPU did you use cross pattern or just went corner by corner for tightening the screws? Also what did you use to spread the thermal paste?

[Godlygamer23]

3 hours ago, QuantumSingularity said:

99.9999999% certain you have poor contact with the die on AT LEAST one of the corners. Average GPU temp is normal at 71°C, but that 115°C hot spot is not. At least not when your average temp is so low. When you disassemble the GPU you will probably be able to see the issue with your own eyes. Repasting a GPU btw is a bit different than CPU. You need to spread the paste evenly across the die, every single millimeter. Don't wait on just the mounting pressure from the cooler. 

Spreading thermal paste manually isn't required on a GPU any more than it's required on a CPU. 

[Pud of Mud]

15 minutes ago, QuantumSingularity said:

Temps are a bit better, but 35°C delta between average and hot spot is A LOT. With the 6600Xt you should be looking at 65-70°C average and about 80°C hotspot.  Are you sure you spread the paste evenly across the whole die? Don't be afraid if a lil bit squishes or falls from the side, it's electrically non-conductive. What paste did you use? Arctic's MX-4 and MX-6 do a perfect job for GPUs if you don't wanna blow your cash on the more expensive stuff. Also check if you pads are put correctly and aren't forming a bulge. When screwing down the cooler back on the GPU did you use cross pattern or just went corner by corner for tightening the screws? Also what did you use to spread the thermal paste?

I used Corsair's TM30 and applied it to the whole die using a crisscross pattern

22 minutes ago, Robchil said:

it hasn't passed 106C i see... that sounds better, i still think it is a  bit high tho. where your pads still soft? should be, you'll notice if they are dried out. 

after the first heat cycle i would tighten the cooler on the GPU again, and test again. 

 

there are thermal paste and liquid metal..  i see some use liquid metal on GPU's, but anyone fixing GPU's would not recommend that, as it's easier to contain liquid metal under the ICS(metal cap) on a CPU than on open GPU core.. that even have resistors ON the core so you should mask it if you do.. anyway.

 

Liquid metal is as conductive as solder, since it sort of is like a solder, and even dissolve solder... so if you spill it you need to get it off realy fast, and if you get it under the core or chips, it should be delivered to someone with the skill to remove the chips to clean it and reball and mount them back on. 

 

I've seen some thermal paste conductive too, but most are not.

I've looked at the pads and they were quite solid, with some squish to them. I've retitened and the temps are still 100 on the hotspot.

[Godlygamer23]

Just now, Pud of Mud said:

I used Corsair's TM30 and applied it to the whole die using a crisscross pattern

I've looked at the pads and they were quite solid, with some squish to them. I've retitened and the temps are still 100 on the hotspot.

This is about what I get on my 3070 while running Superposition. 

[Pud of Mud]

4 minutes ago, Godlygamer23 said:

This is about what I get on my 3070 while running Superposition. 

So I shouldn't be worried if it tops out on ~101 hotspot while doing something incredibly intense? Even though the GPU average temp is ~70?