cooling Good Coolers for Ryzen 9 5900x (PBO Enabled)

[Big Flubba]

I have gotten a Ryzen 9 5900x recently and I have bought the Be Quiet Dark Rock Pro 4 thinking it was going to be plenty to keep my CPU cool. I was unfortunately wrong. With PBO enabled, -10 offset on the curve,, every fan at 100% and running tests I instantly hit 90c. I currently have it set to stock speeds and not to go any higher until I get me a better cooler. I am looking for a good cooler that offers great performance before I look at looks of it. I want a cooler where I have plenty of headroom for PBO boosting it to its max and more headroom where I don't have to worry about temps being an issue. I have heard mixed stories of others saying that this cooler is beyond good for it and others saying the same cooler causes thermal throttling. If possible I would like some real world examples of the cooler in action. I don't want to be a test dummy with coolers to see which one performs the best.

[Electronics Wizardy]

Whats the temps at stock? How much power is the chip pulling? That feels like there might be a mount issue or thermal paste issue as my nh u14s cooled my 5950x with pbo on, and that should be about as hot, with a worse cooler.

[RONOTHAN##]

10 minutes ago, Big Flubba said:

With PBO enabled, -10 offset on the curve,, every fan at 100% and running tests I instantly hit 90c.

May I ask what tests your running that you hit 90C in instantly, and what case are you using? That cooler should be more than plenty for that chip. This really sounds like a bad thermal paste application.

[An0maly_76]

Could be a lot of things. If you're going to air-cool a Zen3, do it like you mean it, especially with the 5900X / 5950X. I run six 120s with a seventh on the Scythe Mugen 5, which is a bit heavier-duty than the Rock Pro (Rock Pro is a good one, but probably not the best choice for the 5900X).

Mine stays within reasonable limits. 38C-43C idle from a cold boot, 57C-65C under moderate load, and peaks around 87C even with PBO. With PBO off, around 76C. It's hit 90 occasionally, but never longer than a moment or two and has never throttled. Once it's run hard for a bit, it usually hovers in the 44C-46C range, though I did discover my case air filters were a bit dirty and cleaned them today.

 

And that's with a case that's not ideal for airflow (Corsair 4000X), which is another consideration. What are your system's other specs? I have an RTX3060ti that puts out more heat than the processor if you ask me.

Edited July 17, 2022 by An0maly_76
Revised, more info

[RONOTHAN##]

3 minutes ago, An0maly_76 said:

I run six 120s with a seventh on the Scythe Mugen 5, which is a bit heavier-duty than the Rock Pro (Rock Pro is a good one, but probably not the best choice for the 5900X).

No. A Rock Pro is a dual tower air cooler which is better performance than a beefy single tower (Mugen 5). The Rock pro is a better cooler than the Mugen 5. 

[An0maly_76]

7 minutes ago, RONOTHAN## said:

No. A Rock Pro is a dual tower air cooler which is better performance than a beefy single tower (Mugen 5). The Rock pro is a better cooler than the Mugen 5. 

What CFM are the fans between the two? I see what you're talking about, and I was probably thinking about the Rock 4 vs the Rock 4 Pro, but quieter fans usually give up a bit of airflow to accomplish that.

Edited July 17, 2022 by An0maly_76
Revised, more info

[RONOTHAN##]

5 minutes ago, An0maly_76 said:

What CFM are the fans between the two? I see what you're talking about, and I was probably thinking about the Rock 4 vs the Rock 4 Pro, but quieter fans usually give up a bit of airflow to accomplish that.

CFM isn't that particularly useful on air coolers since static pressure is the more optimal statistic (though neither one can tell the whole story), though if you care about CFM it's 50.5 vs. 51.2 (basically the same, within fan to fan variance), but the Dark Rock Pro 4 has 2 fans included as well so that's not the issue.

 

If you're going Dark Rock 4 vs. Mugen 5, then yeah the Mugen 5 is the better cooler (not by a ton, but measurable), but Dark Rock Pro 4 is up there with the NH-D15 in performance. 

[Dogzilla07]

3 hours ago, An0maly_76 said:

What CFM are the fans between the two? I see what you're talking about, and I was probably thinking about the Rock 4 vs the Rock 4 Pro, but quieter fans usually give up a bit of airflow to accomplish that.

Stock Scythe 120 fans are inferior to/roughly match the stock fans on the be quiet Rock series in noise-normalized performance, across almost all of their range. 

 

the Dark Rock Pro 4 just has more mass and more heatpipes than Mugen 5, but on the other hand ryzen is so thermally dense, more mass/more heatpipes/more fan performance doesn't make as much benefit on 7nm chiplets than it does on Intel 7 monolithic counterparts.

 

4 hours ago, Big Flubba said:

I want a cooler where I have plenty of headroom for PBO boosting it to its max and more headroom where I don't have to worry about temps being an issue. I have heard mixed stories of others saying that this cooler is beyond good for it and others saying the same cooler causes thermal throttling.

Yeah, exactly, so many mixed stories tells you the issue is somewhere else, and not how good an air cooler/fans are on paper/or even in real world (without any mounting/paste issues). It's a bit of a lottery how good your chip is, and how your motherboard behaves, and how good you undervolt it/curve optimize it.

 

+ LIke mentioned above in this thread, having a vented case, good case fans, good airflow setup. The usually not relevant few degrees paste difference, matters as well. Heck you even have to arbitrary remount the cooler a couple of times just in case. But it's a hassle/lots of things to optimize.

 

If you want a no frills mostly plug&play, with decent headroom cooling that's not water cooling, the only thing that comes to mind is the upcoming copper coldplate IceGiant ProSiphon Elite that LinusTechTips has been showing off lately (but that's not coming out until 2023+). And pls don't get the current one that's available for purchase. It's not gonna do anything

 

There is 1 cooler, that I have a hunch should do better. And that's the Deepcool AK620 that GamersNexus covered (Deepcool did a little bit of Ryzen-specific optimizing for the AK lineup). If i was doing it for me/personally, I'd give it a shot with the AK620, and use Thermal Grizzly Kryonaut Extreme Paste (has to be pink colored).

 

Also please tell us what case, gpu, and case fans you are using.

[Big Flubba]

So after a lot of tweaking and changing settings I have resorted to leaving my findings here. I have power plan set to ultimate performance in Windows. My specs are . . .

https://pcpartpicker.com/list/4NbBhk

 

I have my case fans setup as 3 intakes and one exhaust in the back. I have done 2 applications and mounts on my cooler and I have the mounting hardware and cooler screws tight. I don't have them death grip but they are snug plus a little more. I am using Noctua thermal paste applied using the spread method and a decent layer across the entire chip. I am aiming for maximum performance possible with as much optimizations as I can for cool temps. If some sacrifice is to be made on performance for a big jump in efficiency then I'm all for that. I just don't want to leave untapped performance on the table. I am fine if it is not at the bleeding edge as theirs negative effects. I also want it to be stable too. I know that all of these wants have conflicts, but I want to get the best I can all around. I am fully open to suggestions and am wanting to learn. What should I do?

1152576381_2022-07-1715-51-04.mkv 

 

I have noticed with the scaler at 1x rather then 10x I get better performance at lower temps. I don't know if i should leave it at 1 or 10 when it is supposed to give better performance at 10, but I don't have that thermal headroom.

Edited July 17, 2022 by Big Flubba
Used PCPartPicker

[Big Flubba]

[Big Flubba]

17 hours ago, Electronics Wizardy said:

Whats the temps at stock? How much power is the chip pulling? That feels like there might be a mount issue or thermal paste issue as my nh u14s cooled my 5950x with pbo on, and that should be about as hot, with a worse cooler.

As it turns out I wasn't doing stock. Mounting and thermal paste are good.

 

17 hours ago, RONOTHAN## said:

May I ask what tests your running that you hit 90C in instantly, and what case are you using? That cooler should be more than plenty for that chip. This really sounds like a bad thermal paste application.

Cinebench R23.

 

16 hours ago, An0maly_76 said:

Could be a lot of things. If you're going to air-cool a Zen3, do it like you mean it, especially with the 5900X / 5950X. I run six 120s with a seventh on the Scythe Mugen 5, which is a bit heavier-duty than the Rock Pro (Rock Pro is a good one, but probably not the best choice for the 5900X).

Mine stays within reasonable limits. 38C-43C idle from a cold boot, 57C-65C under moderate load, and peaks around 87C even with PBO. With PBO off, around 76C. It's hit 90 occasionally, but never longer than a moment or two and has never throttled. Once it's run hard for a bit, it usually hovers in the 44C-46C range, though I did discover my case air filters were a bit dirty and cleaned them today.

 

And that's with a case that's not ideal for airflow (Corsair 4000X), which is another consideration. What are your system's other specs? I have an RTX3060ti that puts out more heat than the processor if you ask me.

I have 4120mm fans 3 intakes (2 on the front 1 on the top with the filter on), and 1 exhaust.

[An0maly_76]

Might try making the top fan an exhaust and see if that helps. A single exhaust fan can only do so much. I have three front intake, three exhaust (two top, one rear) and have no issues.

Edited July 17, 2022 by An0maly_76
Revised, more info

[RONOTHAN##]

6 minutes ago, Big Flubba said:

Cinebench R23

That really shouldn't be happening. It should be somewhere around 85C in R23. 

 

7 minutes ago, Big Flubba said:

Mounting and thermal paste are good.

Are you sure about that? I would remount the cooler and check the thermal paste application anyway, this really sounds like a bad application. 

 

49 minutes ago, Big Flubba said:

I have noticed with the scaler at 1x rather then 10x I get better performance at lower temps. I don't know if i should leave it at 1 or 10 when it is supposed to give better performance at 10, but I don't have that thermal headroom.

The PBO scalar is weird, it tends to sweet spot pretty hard on the lower end. The scalar being out of whack can be your entire issue

[Big Flubba]

5 minutes ago, RONOTHAN## said:

Are you sure about that? I would remount the cooler and check the thermal paste application anyway, this really sounds like a bad application.

 

57 minutes ago, Big Flubba said:

I have my case fans setup as 3 intakes and one exhaust in the back. I have done 2 applications and mounts on my cooler and I have the mounting hardware and cooler screws tight. I don't have them death grip but they are snug plus a little more. I am using Noctua thermal paste applied using the spread method and a decent layer across the entire chip.

 

[Big Flubba]

10 minutes ago, An0maly_76 said:

Might try making the top fan an exhaust and see if that helps. A single exhaust fan can only do so much. I have three front intake, three exhaust (two top, one rear) and have no issues.

It may, but the method to my madness is to have as much fresh air as possible coming in and having positive pressure. the rest of the positive pressure can escape out of the PCI Express slots and such.

[Big Flubba]

10 minutes ago, RONOTHAN## said:

That really shouldn't be happening. It should be somewhere around 85C in R23. 

Let me test it.

[Big Flubba]

5 minutes ago, Big Flubba said:

Let me test it.

Same thing is happening.

[Big Flubba]

It goes right up to 90.

[Big Flubba]

21 minutes ago, An0maly_76 said:

Might try making the top fan an exhaust and see if that helps. A single exhaust fan can only do so much. I have three front intake, three exhaust (two top, one rear) and have no issues.

Can you share your bios settings?

[Big Flubba]

3 minutes ago, Big Flubba said:

Same thing is happening.

I changed the customized pbo scalier and pbo scalier to 1x.

[An0maly_76]

20 minutes ago, Big Flubba said:

It may, but the method to my madness is to have as much fresh air as possible coming in and having positive pressure. the rest of the positive pressure can escape out of the PCI Express slots and such.

You can have as much potential for fresh air intake as you want, it means nothing if there is not enough exhaust flow, because the heat is not being exhausted. You can never take in more than you can push out, and a single 120 pushing out and a few minor holes in some expansion slots are never going to max out the three 120s you have pulling in. Simple physics.

Edited July 17, 2022 by An0maly_76
Revised, more info

[Big Flubba]

41 minutes ago, An0maly_76 said:

You can have as much potential for fresh air intake as you want, it means nothing if there is not enough exhaust flow, because the heat is not being exhausted. You can never take in more than you can push out, and a single 120 pushing out and a few minor holes in some expansion slots are never going to max out the three 120s you have pulling in. Simple physics.

I made it an exhaust.

[Big Flubba]

1 hour ago, RONOTHAN## said:

That really shouldn't be happening. It should be somewhere around 85C in R23.

I don't know why it is that hot.

[freeagent]

I use a Thermalright FC140 on my 5900X in a Meshify C. It idles in the mid to upper 20s and loads in the low to mid 80s with a load from Linpack Xtreme. I have 4x Intakes and no exhaust on my case.. I have close to 528cfm of air going through the Meshify C, and roughly 268cfm through the cooler.. works great. 

[Big Flubba]

16 minutes ago, freeagent said:

I use a Thermalright FC140 on my 5900X in a Meshify C. It idles in the mid to upper 20s and loads in the low to mid 80s with a load from Linpack Xtreme. I have 4x Intakes and no exhaust on my case.. I have close to 528cfm of air going through the Meshify C, and roughly 268cfm through the cooler.. works great. 

Can you share your settings in the bios and benchmarks and such?