5800x temp problems. Help!

[Cfischer6797]

To start here is my system

Cpu - 5800x

Gpu - EVGA 3080 FTW

Case - Lian Li O-11 dynamic

CPU cooler - Lian li 360 AIO - top mounted

6 intake fans, 3 exhaust fans

 

So when I first built my computer at the start of last year, I had set everything up, and was monitoring temperatures closely, and the CPU was around 72-75c while running cinnebench. Which was great!

Fast forward to yesterday, while playing Hunt: Showdown I can hear my PC being a bit louder then i expected. I open up Ryzen master and I see my CPU sitting at 95c!!!!!

I freaked out shut down my computer, was looking at settings and everything was stock. Never OC'ed. and cleared my CMOS just to be safe. Adjusted fan curves and CPU pump curves to be a bit more aggressive. reapplied thermal paste and made sure that was all good. and just ran the game again...

Same thing...

So I undervolted it to 1.2v and have the speed set to 4.0 and while gaming its still getting to around 85c.

While sitting here typing this out with nothing else open, im sitting at 48c

is my cooler dying?

What am I missing here?

[GuiltySpark_]

I don't suggest manual undervolting with Zen3 but that's besides the point.

 

Is the pump RPM OK? 

[Cfischer6797]

Just now, GuiltySpark_ said:

I don't suggest manual undervolting with Zen3 but that's besides the point.

 

Is the pump RPM OK? 

Per the Bios its working just fine, but even in the BIOS its sitting at 40c,

[GuiltySpark_]

I would share a screenshot of Ryzen Master in Advanced Mode when the system is at idle. 

 

A CMOS reset couldn't hurt as well. 

 

The AIO may just be clogged, can you feel a temperature difference between the tubes? That method isn't really that great of a way to tell anything but its something. Might as well reseat it with new paste just to be sure.

[Zando_]

From a spot of googling, looks like high CPU usage for Cryengine games is normal, and Hunt is known for being very CPU intensive. Seeing the temps people are claiming, I assume it's using AVX instructions on top of that. 

1 minute ago, GuiltySpark_ said:

Is the pump RPM OK? 

^^^ Your pump should be going 100% at those temps. 

2 minutes ago, Cfischer6797 said:

Per the Bios its working just fine, but even in the BIOS its sitting at 40c,

Most AIOs have a utility you can run that lets you see and set fan and pump curves, see what it's running under load + how the coolant temps are.

 

AMD does say the max operating temp is 90C for the 5800X, so you definitely don't want to be running 95C regularly, I do doubt it'll kill the chip quickly though. 

[Cfischer6797]

Just now, Zando_ said:

From a spot of googling, looks like high CPU usage for Cryengine games is normal, and Hunt is known for being very CPU intensive. Seeing the temps people are claiming, I assume it's using AVX instructions on top of that. 

^^^ Your pump should be going 100% at those temps. 

Most AIOs have a utility you can run that lets you see and set fan and pump curves, see what it's running under load + how the coolant temps are.

 

AMD does say the max operating temp is 90C for the 5800X, so you definitely don't want to be running 95C regularly, I do doubt it'll kill the chip quickly though. 

When running cinnebench my temps shot up to 105c, and didnt even thermal throttle after only 15 seconds.

[Cfischer6797]

4 minutes ago, Zando_ said:

From a spot of googling, looks like high CPU usage for Cryengine games is normal, and Hunt is known for being very CPU intensive. Seeing the temps people are claiming, I assume it's using AVX instructions on top of that. 

^^^ Your pump should be going 100% at those temps. 

Most AIOs have a utility you can run that lets you see and set fan and pump curves, see what it's running under load + how the coolant temps are.

 

AMD does say the max operating temp is 90C for the 5800X, so you definitely don't want to be running 95C regularly, I do doubt it'll kill the chip quickly though. 

I already cleared CMOS before undervolting. and that didnt help. Thats why i ended up undervolting

[Cfischer6797]

7 minutes ago, GuiltySpark_ said:

I would share a screenshot of Ryzen Master in Advanced Mode when the system is at idle. 

 

A CMOS reset couldn't hurt as well. 

 

The AIO may just be clogged, can you feel a temperature difference between the tubes? That method isn't really that great of a way to tell anything but its something. Might as well reseat it with new paste just to be sure.

CMOS has already been cleared,

 

New paste has been changed out.

 

And I can feel a difference in the tubes temps

[Zando_]

8 minutes ago, Cfischer6797 said:

When running cinnebench my temps shot up to 105c, and didnt even thermal throttle after only 15 seconds.

If it's at all like a lotta intel boards, your board may disable a bunch of limits by default for better performance. 

From what AMD has said on the matter, these chips are designed to hop up to 90C when they need to (allows them to pursue a longer boost time). Getting higher than that is definitely not intended, you probably have an issue with your AIO. Either mounting (it's not pulling heat out of the CPU), or the cooler itself (the radiator isn't pulling heat out of the cooler). AIO failures are usually pump related, I don't think I've seen reports of them clogging or anything, just the pumps failing. Thus the insistence on checking the pump rpm under load, you can also take a standard screwdriver, place the point against the CPU block, and put the handle on the bone just in front of your ear. Same concept as bone conduction headphones, it'll let you hear/feel the vibration from the pump spinning. If you don't then the pump is dead. 

 

EDIT:

1 minute ago, Cfischer6797 said:

And I can feel a difference in the tubes temps

Yikes yeah that's a clog or pump failure, coolant temp should be basically the same throughout the whole loop assuming proper flow. 

[GuiltySpark_]

Something is off, I'd be curious if my pump is running at maximum potential. I use an EK 360 and it runs at full speed (2600RPM) at all times. I noticed you're also using a custom profile in Ryzen Master? You may want to revert all that to defaults using the reset button in the bottom left.

 

Idle with a -18 all core undervolt using the curve optimizer.

 

[Cfischer6797]

3 minutes ago, Zando_ said:

If it's at all like a lotta intel boards, your board may disable a bunch of limits by default for better performance. 

From what AMD has said on the matter, these chips are designed to hop up to 90C when they need to (allows them to pursue a longer boost time). Getting higher than that is definitely not intended, you probably have an issue with your AIO. Either mounting (it's not pulling heat out of the CPU), or the cooler itself (the radiator isn't pulling heat out of the cooler). AIO failures are usually pump related, I don't think I've seen reports of them clogging or anything, just the pumps failing. Thus the insistence on checking the pump rpm under load, you can also take a standard screwdriver, place the point against the CPU block, and put the handle on the bone just in front of your ear. Same concept as bone conduction headphones, it'll let you hear/feel the vibration from the pump spinning. If you don't then the pump is dead. 

 

EDIT:

Yikes yeah that's a clog or pump failure, coolant temp should be basically the same throughout the whole loop assuming proper flow. 

 Wouldn't one tube have hot water heading to the radiator from the block, and the other tube have cooled down water heading to the block?

[Cfischer6797]

5 minutes ago, GuiltySpark_ said:

Something is off, I'd be curious if my pump is running at maximum potential. I use an EK 360 and it runs at full speed (2600RPM) at all times. I noticed you're also using a custom profile in Ryzen Master? You may want to revert all that to defaults using the reset button in the bottom left.

 

Idle with a -18 all core undervolt using the curve optimizer.

 

This is the update after resetting, the custom profile was the undervolt

[Zando_]

Just now, Cfischer6797 said:

 Wouldn't one tube have hot water heading to the radiator from the block, and the other tube have cooled down water heading to the block?

The coolant temp should equalize pretty quickly as it runs around the loop, constantly.  It shouldn't stop and pick up a bunch of heat in one spot, that means your pump has stopped, which either means something isn't plugged in (if it's on the CPU fan header then your mobo would throw a CPU fan error on boot, if not then it probably won't say anything) or the pump is dead. 

[An0maly_76]

A pic of your cooling setup might help in pinpointing any trouble spots. My 5900X ranges 42C idle to 75C-86C full load, PBO dependent -- air-cooled.

Edited April 25, 2022 by An0maly_76
Revised, more info

[Cfischer6797]

Just now, Zando_ said:

The coolant temp should equalize pretty quickly as it runs around the loop, constantly.  It shouldn't stop and pick up a bunch of heat in one spot, that means your pump has stopped, which either means something isn't plugged in (if it's on the CPU fan header then your mobo would throw a CPU fan error on boot, if not then it probably won't say anything) or the pump is dead. 

So it has to be working to a degree, I shut down my computer, unplugged the pump and temps skyrocketed so fast in the BIOS. So it has to be working somewhat

[RMTM]

What is your room temperature?

 

This might be another case of gunk build-up inside the liquid lines in the AIO. Causing blocked or restricted liquid flow. Resulting to cpu high temperatures.

 

I suggest for you to stop benchmarking it while you are having these cooling issues figured-out. Your idle temps can already tell you if you have cooling issues.

 

I suggest for you to just get an air-cooler for your cpu instead. There are many options, I can help you out with. 

[Cfischer6797]

2 minutes ago, An0maly_76 said:

A pic of your cooling setup might help in pinpointing any trouble spots. My 5900X ranges 42C idle to 75C-86C full load, PBO dependent.

 

[Cfischer6797]

2 minutes ago, RMTM said:

What is your room temperature?

 

This might be another case of gunk build-up inside the liquid lines in the AIO. Causing blocked or restricted liquid flow. Resulting to cpu high temperatures.

 

I suggest for you to stop benchmarking it while you are having these cooling issues figured-out. Your idle temps can already tell you if you have cooling issues.

 

I suggest for you to just get an air-cooler for your cpu instead. There are many options, I can help you out with. 

Room temp is 72F

[An0maly_76]

That case honestly doesn't appear well-suited for cooling. And that's probably a lot of your problem. It can't cool if there's nowhere for the heat to go.

 

My setup -

 

Corsair 4000X case - includes three SP120 fans

Added three LL120 fans (wasn't aware these didn't match the included fans, but they work)

Corsair iCUE Commander Core XT (two thermal probes in CPU cooler)

Scythe Mugen 5 Black Edition

Custom fan curve -- 50-75 percent 30C-70C - 85%-90% above 70. Never tops 76C without PBO, 86C with PBO.

Edited April 25, 2022 by An0maly_76
Revised

[GuiltySpark_]

The key point to not lose sight of here is this is a new problem that did not exist before. The current setup was perfectly adequate and working well so any suggestions to physically change anything would be losing sight of the problem. 

 

I know we've mentioned it a few times but I don't know if you've fully answered it. Have you verified the pump RPM? The actual number. Lian Li doesn't actually seem to publish the RPM range on their site but some googling says it should be run in the 3000RPM range.

[191x7]

I would try reorienting the cooler, have it work in a different possition, might help clear the partial blockage. 

[Cfischer6797]

1 minute ago, GuiltySpark_ said:

The key point to not lose sight of here is this is a new problem that did not exist before. The current setup was perfectly adequate and working well so any suggestions to physically change anything would be losing sight of the problem. 

 

I know we've mentioned it a few times but I don't know if you've fully answered it. Have you verified the pump RPM? The actual number.

Per Open hardware monitor it as at 1500 RPM right now

[Zando_]

4 minutes ago, An0maly_76 said:

That case honestly doesn't appear well-suited for cooling. And that's probably a lot of your problem. It can't cool if there's nowhere for the heat to go.

Case has plenty of airflow, it's set up for AIOs/Custom loops over air coolers, thus why the intakes are on the side and not the front. It just needs to pull fresh air into the case, the AIO fans will push it out. I would reccomend using the AIO as an intake and the side as an exhaust to maximize fresh air to the radiator though, with fans on the bottom to pull fresh air into the GPU it should be a good setup, but same concept as what OP already has, just flipped. 

2 minutes ago, GuiltySpark_ said:

I know we've mentioned it a few times but I don't know if you've fully answered it. Have you verified the pump RPM? The actual number. Lian Li doesn't actually seem to publish the RPM range on their site but some googling says it should be run in the 3000RPM range.

^^^ Assuming it's an Asetek AIO (almost all of them are due to Asetek having nearly all the patents), it should run about the same as any other AIO. 

 

EDIT to avoid double post:

Just now, Cfischer6797 said:

Per Open hardware monitor it as at 1500 RPM right now

If you have the ability to set it to max rpm, do so and see if there's any difference in temps. 

[RMTM]

6 minutes ago, Cfischer6797 said:

Room temp is 72F

Your room is cold enough. That rules it out.

 

I suggest for you to change your AIO already. The Noctua's NH-U12A is a good option for you. Or an NH-D15, if your RAM allows clearance for double fan.

 

I have an 11700K overclocked (41%) in a room at 86F(30c) temp and it barely hit 65c on gaming. Idle is at 39-40c. Cinebench23.2 only reaches 85c on multi-core overclocked at 41%(~5100mhz) in same room temp 30c.

[Cfischer6797]

4 minutes ago, An0maly_76 said:

That case honestly doesn't appear well-suited for cooling. And that's probably a lot of your problem. It can't cool if there's nowhere for the heat to go.

That wouldnt explain why it use to max out at 75c a while ago, it has plenty of room for the air to escape, plus my GPU never hits above 65c full load. so that seems to show plenty of air flow throughout