Games intermittently freezing

[Rigo]

G'day folks,
Self built rig, I rebuilt to correct some minor errors.
All new hardware, everything passed diagnostics and stress/burning tests.
Only potential issue was that the SSD was at 21% free space left, DISKGENIUS says it's good.
I've updated everything to the latest including the BIOS.
High CPU/GPU usage still showing when monitoriring with motherboard utility.
The customer said he had clean installed Win11 multiple times and with only one game installed and still getting freeze ups.
PSU 850W
Ryzen 7 9800X3D - up to 95°C
GPU Palit Gamerock GeForce RTX 5080 16GB GDDR7 - up to 56°C
Mobo Gigabyte B650 EAGLE AX Rev 1.1
RAM 32GB DDR5

 

[sapphirescales]

The customer said he had clean installed Win11 multiple times

But did he really, or did he just use the built in utility to do a fresh install of Windows? If there's an issue with corruption on the drive, that won't fix the issue. I just had a client in with a gaming laptop that would blue screen every 5 minutes. He did a reset using the built in utility multiple times and it didn't fix his issue. I ran my diagnostics and couldn't find anything wrong with it, so I did a fresh install using the Windows 11 flash drive and that fixed it.

Have you tried running OCCT? That's the most recent diagnostic software I've added to my collection and it's pretty good for diagnosing gaming systems/custom builds.

 

[lan101]

Maybe look into underclocking the amd processor. I did that a while back with one and that seemed to fix intermitten random rebooting etc.

It would only do it of course when gaming never with just browsing the web or basic things. Maybe even underclocking the video card a bit too. I found multiple youtube videos at the time explaining it and it seemed to work.

 

[Rigo]

But did he really, or did he just use the built in utility to do a fresh install of Windows? If there's an issue with corruption on the drive, that won't fix the issue. I just had a client in with a gaming laptop that would blue screen every 5 minutes. He did a reset using the built in utility multiple times and it didn't fix his issue. I ran my diagnostics and couldn't find anything wrong with it, so I did a fresh install using the Windows 11 flash drive and that fixed it.

Have you tried running OCCT? That's the most recent diagnostic software I've added to my collection and it's pretty good for diagnosing gaming systems/custom builds.

I've used Furmark for both the GPU - CPU burners.
I'll give OCCT next time thx

 

[Rigo]

Maybe look into underclocking the amd processor. I did that a while back with one and that seemed to fix intermitten random rebooting etc.

It would only do it of course when gaming never with just browsing the web or basic things. Maybe even underclocking the video card a bit too. I found multiple youtube videos at the time explaining it and it seemed to work.

This is just like the game stuttering with high CPU and GPU usage.
Don't think that he'd be happy with underclocking such an expensive rig

 

[timeshifter]

I'd suggest paying close attention to video drivers based on my personal experience. I built myself a new PC a year ago and was ready to send it all back. It was a nightmare of days and days of troubleshooting. In the end it was the current NVIDIA driver not getting along with Fortnite. There were a community of users with similar issues and they had settled on a prior GPU driver. I followed them and the system has be rock solid ever since.

https://www.technibble.com/forums/threads/new-pc-build-crashing.91732/

 

[lan101]

This is just like the game stuttering with high CPU and GPU usage.
Don't think that he'd be happy with underclocking such an expensive rig

Yeah I wouldn't want to but 95 degrees celcius is pretty up there. Hopefully it's a video driver issue instead.
This is one of the main reasons I don't like working on anyone's gaming machine. Lots of time and effort usually and many of them are tight wads and don't want to spend anything for it. At least that's been my experience over the years. Also with parts being crazily outrageous in price right now even further detours me from it lol.

Also many comments on those youtube videos were people saying how everything was terrible until they did the underclocking and then everything worked great so I would definitely try it if nothing else works.

Like @sapphirescales has pointed out before...right area with the right customers there can be great money in these gaming rigs...unfortunately for me I'm not in an area like that lol.

 

[DRPCNZ]

Wouldn't hurt to do some research and check the hardware is actually compatible with the game. Some games don't like some CPUs, or graphics cards. Time to let the client know its going to cost money to sort wether you find a fix or not

 

[GTP]

Yeah I wouldn't want to but 95 degrees celcius is pretty up there.

Seeing 95 - 98 °C on a Ryzen 7 9800X3D under load is normal. AMD designed these CPUs to run safely at these temperatures - the Tj Max is ~95 °C, which is the threshold where the CPU will start throttling if needed.

The chip’s Precision Boost pushes performance right up to this thermal limit, so occasional spikes into the high 90s are expected under heavy workloads. If your cooling is adequate, there’s no risk of damage, crashes, or shutdowns just from these temperatures.

Mine regularly runs at 98.xx Deg C running heavy CPU/GPU intensive Boinc tasks like LatinSquares, NumberFields@Home, NFS@Home and SRBase. It doesn't shutdown, reboot or crash.

Even with the AC on my Ryzen 7 9800X3D runs at 90+ without issues.

 

[britechguy]

@GTP

Tmax even allows for brief spikes above that temperature. It's much like speed rating on tires, where the sustained allowable speed is not a limit that you can't exceed (and probably never do, unlike CPU temps) for a few moments and then settle right back down into "fully within limits" again.

I have grown bone-tired of all the screaming about "running hot" when the temperatures being reported are most often well below Tmax. I always loved what one engineer once said to me in regard to approaching upper or lower limits (of anything): A millimeter is as safe a distance as mile. You can run perfectly fine at Tmax because that is the upper end of "within normal limits." That's what it's always been meant to represent, not some temperature that you have to fear will end your computer if you're running at it or very near to it below it for long periods of time. Not that most of us generally do when "coasting along" doing non-processor-intensive activities, but when you're in the middle of those and the machine is at that degree of CPU usage for extended periods, being near Tmax is normal (with decent, typical air cooling).

 

[GTP]

I tossed up between air cooling and hydra cooling and settled for a 360mm liquid cooling solution.
The AIO pump is set in the BIOS to "ramp up" as needed. The AIO fans are at 100% (set to that in BIOS) and the only other cooling fan (aside from the 3 in my 4060Ti) is a small case fan at the rear set to 900rpm max.

The AIO pump rarely maxes out even with CPU intensive loads (described above), staying at around 1000~1300 rpm. (Max is 2300 rpm.)

Even with the CPU at 98 deg C I can play any game or do any other work with ease.
As I type this it's running at 97.4 deg C crunching for Amicable Numbers. I'm also watching a Freetube vid (movie) and chatting on Signal.

 

[sapphirescales]

Yeah I wouldn't want to but 95 degrees celcius is pretty up there.

I don't like it either but these modern Ryzen CPUs are DESIGNED to hit 95c. They will literally boost until they reach that temperature. You can keep them below that temp, but it requires some pretty serious cooling (a 360mm liquid cooler). I usually manually adjust the power limits in the BIOS so they can't boost that hard. For a 9800X3D I recommend setting the power limit to 120w. If you don't do this, it will boost until it reaches 160w or 95c, whichever comes first. That being said, since they're meant to hit 95c, I doubt that's the problem with this computer.

@Rigo Make sure you run the "Power" test in OCCT. This will run everything on the computer at full power. If you've got a nasty pain in the butt to diagnose issue like degraded or flaky VRMs on the motherboard, this will reveal that problem.

 

[YeOldeStonecat]

What make/model power supply?
Some graphic cards (such as the one here)....really desire quality solid power at the 12v rail....and a power supply that can keep up with "spikes" of thirst from the graphic card.

 

[lan101]

I don't like it either but these modern Ryzen CPUs are DESIGNED to hit 95c. They will literally boost until they reach that temperature. You can keep them below that temp, but it requires some pretty serious cooling (a 360mm liquid cooler). I usually manually adjust the power limits in the BIOS so they can't boost that hard. For a 9800X3D I recommend setting the power limit to 120w. If you don't do this, it will boost until it reaches 160w or 95c, whichever comes first. That being said, since they're meant to hit 95c, I doubt that's the problem with this computer.

@Rigo Make sure you run the "Power" test in OCCT. This will run everything on the computer at full power. If you've got a nasty pain in the butt to diagnose issue like degraded or flaky VRMs on the motherboard, this will reveal that problem.

Yeah until I looked it up I didn't realize that specific one was designed to run that hot lol. I believe that's what I did with another amd ryzen 7 series was adjust the limits etc. Once I did that all the random crashing in games went away. I'm a little out of touch apparently on modern higher end hardware lol.

 

[xrobwx71]

This is just like the game stuttering with high CPU and GPU usage.
Don't think that he'd be happy with underclocking such an expensive rig

What game?
PSU 850W make/model?
Do they have stable internet?
Maybe this will give you some insight. https://www.sysnative.com/forums/th...d-dpc-isr-latency-posting-instructions.23367/

 

[sapphirescales]

Yeah until I looked it up I didn't realize that specific one was designed to run that hot lol. I believe that's what I did with another amd ryzen 7 series was adjust the limits etc. Once I did that all the random crashing in games went away. I'm a little out of touch apparently on modern higher end hardware lol.

Yeah the first time I installed one of these processors and it hit 95c instantly, I literally thought "ah, sh*t, I must have forgotten to remove the protective peel from the CPU cooler. How long have I been doing this again? I'm such an idiot!" then I checked and of course I hadn't forgotten to remove the peel, so I tightened the cooler thinking I just didn't have enough mounting pressure. Then I replaced the fan with one that ran much faster (3,000rpm vs. the 1,000rpm of the one I had in there) and it went down to like 93c while sounding like a jet engine LOL.

 

[britechguy]

The fact of the matter is that it's essential to take the time to look up Tmax for any given processor where you (or your client) have overheating concerns. That number has been going up and up for some time. I think we may very well be reaching our limits.

Gemini: [What is the current highest Tmax, in degrees Celsius, for any currently produced CPU in 2026?]
Pertinent snippet:
----------------------
As of early 2026, the current highest official maximum operating temperature (1$T_{junction}$ Max or TjMax) for a mass-produced consumer CPU is 110°C.2

This thermal limit is primarily found in Intel's high-performance mobile (H-series) processors, specifically the Core Ultra 200H series (Arrow Lake-H) and the Core Ultra 100H series (Meteor Lake).

While 100°C was the industry standard for over a decade, recent architectures have pushed thermal envelopes higher to maintain boost clocks in thin-and-light or high-density laptop chassis.
-----------------------

It has been a very, very long time since anything below 90 degrees Celsius was considered to be "hot." Even now, many processors are designed to run above that temperature.

 

[lan101]

Yeah the first time I installed one of these processors and it hit 95c instantly, I literally thought "ah, sh*t, I must have forgotten to remove the protective peel from the CPU cooler. How long have I been doing this again? I'm such an idiot!" then I checked and of course I hadn't forgotten to remove the peel, so I tightened the cooler thinking I just didn't have enough mounting pressure. Then I replaced the fan with one that ran much faster (3,000rpm vs. the 1,000rpm of the one I had in there) and it went down to like 93c while sounding like a jet engine LOL.

Yeah lol. Crazy how much things have changed over the years. I remember when going to like 70c or higher was considered like insanely too hot lol.

 

[sapphirescales]

Yeah lol. Crazy how much things have changed over the years. I remember when going to like 70c or higher was considered like insanely too hot lol.

Well 70c or under is still where you want to be when it comes to normal processors (65w or less) even on a stock cooler. If you take that 7600X and put a 65w power limit on it, it will run at less than 70c under full load using the AMD stock cooler. Though I don't recommend running the stock cooler on a 7600X because if the BIOS ever loses the power limit settings it will thermal throttle like crazy but a cheap $20 tower cooler will keep it around 60c and if the BIOS ever loses the settings it will keep it at 95c like "normal." I try to avoid using processors that draw more than 65w because I'm looking for longevity over maximum performance, and for 99% of people, a 65w CPU provides more than enough performance.

 

[HCHTech]

It has been a very, very long time since anything below 90 degrees Celsius was considered to be "hot." Even now, many processors are designed to run above that temperature.

Agreed - temps like this still make me wince, though. I think I'd look in to a robust AIO water cooling setup, and like @YeOldeStonecat said, a different power supply - if only as a test. We've seen quite a few gaming rigs over the last couple of years since taking over another firms residential customers, and we've got at least 2 fancier 1000-watt PSUs for testing.