5900X Temps - What cooler do you use?

[BornCZ]

Hi, I am cooling my new Ryzen 5900X with Noctua NH-U12S (with 2 fans) and it´s honestly far from ideal. Ideally I would use NH-U12A, but the chromax variant is still too far away and I don´t want anything bigger as far as air coolers go (NH-D15 for example). Now I am deciding if I should try 280mm AIO or stay with air (which I prefer).

 

Idle temps: 41-45c

Cinebench R23 (MC) – 74-75c (20261)

Cinebench R23 (SC) – 68-70c (1587)

Gaming is when it gets really toasty, while playing Death Stranding core temps are around 84-89c and I even saw a spike to 91c! Coming from 6700K these temps have me little worried. In my old system GPU fans (GTX 1070) were the loudest part, but since TUF 3080 is so quiet, I can´t even hear it over the CPU and case fans when gaming.

 

Rest of the system is in my profile and the case (500DX) is well vented (2x NF-A14 front as intake, NF-S12A rear and BQ Purewing 3 top as exhaust). Since the temps were this high I even tried completely disabling PBO and all temps above are with PBO disabled. Everything else is stock and ambient temp is around 23-24c. When stable (non-beta) BIOS with PBO2 will be available for my MB I am going to UV which could help significantly from early reports I saw.

 

My question is. If you have 5900X, what cooler do you use and what temps are you getting? Epecially when gaming (real-world usage).

[IkeaGnome]

Short of going through and resaying stuff a couple people said here, I'd give this a read from yesterday. 5xxx CPUs run hot. By design. You're well within limits and I wouldn't worry too much about it if you're worried about breaking something. If you just want to see smaller numbers, the NH-D15 is a very good way to get those down. If you don't want a large air cooler, I'd look at AIOs that will compete with it.

 

[DailyProcrastinator]

I have a 3900XT, and I get like 70c max under heavy load, but I have low ambient temps, and a much larger air cooler. Ryzen 5xxx series runs hot though.

 

12 minutes ago, BornCZ said:

Hi, I am cooling my new Ryzen 5900X with Noctua NH-U12S (with 2 fans) and it´s honestly far from ideal. Ideally I would use NH-U12A, but the chromax variant is still too far away and I don´t want anything bigger as far as air coolers go (NH-D15 for example). Now I am deciding if I should try 280mm AIO or stay with air (which I prefer).

If you prefer air why do you not want a large air cooler?

 

12 minutes ago, BornCZ said:

Rest of the system is in my profile and the case (500DX) is well vented (2x NF-A14 front as intake, NF-S12A rear and BQ Purewing 3 top as exhaust). 

Honestly try disconnecting your two top exhaust fans that are closest to the front intake, in the past I spent A LOT of time playing with fan curves and setups and found that top exhaust can actually steal from the cool air coming from the front intake fans. That or set them to their absolute lowest operating RPM. Worth a shot. 

 

Also without changing hardware you can decrease temps with a slight under-volt (with offset). 

[Mateyyy]

How's 74-75C at full load "far from ideal"..?

[DailyProcrastinator]

6 minutes ago, Mateyyy said:

How's 74-75C at full load "far from ideal"..?

OP is probably a bit more concerned about gaming temps. 

25 minutes ago, BornCZ said:

Gaming is when it gets really toasty, while playing Death Stranding core temps are around 84-89c and I even saw a spike to 91c!

EDIT: That said with a 5900x and NH-U12S not abnormal temps. 

[Mateyyy]

3 minutes ago, DailyProcrastinator said:

OP is probably a bit more concerned about gaming temps. 

...I'm blind.

 

That's a bit odd though, considering temps in Cinebench R23 are 10C+ lower.

OP, are you seeing the same high temperatures (high 80s, low 90s) if you run a stress test like Prime95 or Linpack?

[DailyProcrastinator]

3 minutes ago, Mateyyy said:

...I'm blind.

It happens hahaha

 

3 minutes ago, Mateyyy said:

That's a bit odd though, considering temps in Cinebench R23 are 10C+ lower.

Likely just sustained load temps vs shorter runs.

 

4 minutes ago, Mateyyy said:

OP, are you seeing the same high temperatures (high 80s, low 90s) if you run a stress test like Prime95 or Linpack?

Agreed.

[BornCZ]

2 hours ago, Voluspa said:

Short of going through and resaying stuff a couple people said here, I'd give this a read from yesterday. 5xxx CPUs run hot. By design. You're well within limits and I wouldn't worry too much about it if you're worried about breaking something. If you just want to see smaller numbers, the NH-D15 is a very good way to get those down. If you don't want a large air cooler, I'd look at AIOs that will compete with it.

 

Thanks, I know I am within limits, but I am more concerned about noise levels. I have PC right next to me on the desk and I use openback headphones. Main goal of this post was comparing my results to others with same CPU but different cooling solutions, to get some baseline.

[BornCZ]

2 hours ago, DailyProcrastinator said:

I have a 3900XT, and I get like 70c max under heavy load, but I have low ambient temps, and a much larger air cooler. Ryzen 5xxx series runs hot though.

 

If you prefer air why do you not want a large air cooler?

 

Honestly try disconnecting your two top exhaust fans that are closest to the front intake, in the past I spent A LOT of time playing with fan curves and setups and found that top exhaust can actually steal from the cool air coming from the front intake fans. That or set them to their absolute lowest operating RPM. Worth a shot. 

 

Also without changing hardware you can decrease temps with a slight under-volt (with offset). 

It´silly but I just don´t like the look of them. They just obstruct everything in the case and usually have problems with RAM clearence.

 

Top fan shouldn´t be a problem. I have only one at the top and its at the very rear end of the case, nowhere near the front fans.

[BornCZ]

2 hours ago, DailyProcrastinator said:

OP is probably a bit more concerned about gaming temps. 

EDIT: That said with a 5900x and NH-U12S not abnormal temps. 

Yeah, I thought that it could be normal temps for this combo, but I was hoping I could get some experience from other 5900x owners with different coolers to compare.

[BornCZ]

2 hours ago, DailyProcrastinator said:

Likely just sustained load temps vs shorter runs.

Not sure, that this will be the case. Cinebench run is like 8-10 mins and I get 85+ temps within a couple minutes after playing the game (and 80+ within first minute).

[BornCZ]

3 hours ago, Mateyyy said:

That's a bit odd though, considering temps in Cinebench R23 are 10C+ lower.

OP, are you seeing the same high temperatures (high 80s, low 90s) if you run a stress test like Prime95 or Linpack?

I don´t find overall temps bad considering relatively small air cooler, but this difference also seems odd to me. I haven´t tried either one yet. 

 

EDIT: I tried at least Time Spy Stress Test that I have at hand and which should emulate load similar to gaming and temps were hovering around 74-75c.

[IkeaGnome]

1 hour ago, BornCZ said:

I don´t find overall temps bad considering relatively small air cooler, but this difference also seems odd to me. I haven´t tried either one yet. 

 

EDIT: I tried at least Time Spy Stress Test that I have at hand and which should emulate load similar to gaming and temps were hovering around 74-75c.

I wonder if it's hotter in gaming because of core utilization. Is Death Stranding using less cores at a higher utilization rate?

[BornCZ]

11 hours ago, Voluspa said:

I wonder if it's hotter in gaming because of core utilization. Is Death Stranding using less cores at a higher utilization rate?

I am pretty sure thats the reason. CCD1 temps are lot higher than CCD2 temps.

 

EDIT: I did some more testing and I it seems that Death Stranding alone is the "culprit". I tried older engine game (Witcher 3) and the temps were much lower (75c vs 85c).

 

Difference between Witcher 3 and Death Stranding was obvious.

 

Witcher 3 puts heavier load on 4 cores, which maxed utilization on 100% and were boosting up to 4951 MHz while other cores were hovering around 3000-3500 MHz.

 

Death Stranding on the other hand spread the load across almost all cores and not a single core ever hit a 100% utilizitation. Most of the time all the cores were running uniformly on 4399 MHz. I am wondering if the temp bump can be caused by Denuvo which DS uses.

 

EDIT2: Seems like Denuvo was removed in December patch so its likely not the culprit.

[AngryBeaver]

21 hours ago, BornCZ said:

Hi, I am cooling my new Ryzen 5900X with Noctua NH-U12S (with 2 fans) and it´s honestly far from ideal. Ideally I would use NH-U12A, but the chromax variant is still too far away and I don´t want anything bigger as far as air coolers go (NH-D15 for example). Now I am deciding if I should try 280mm AIO or stay with air (which I prefer).

 

Idle temps: 41-45c

Cinebench R23 (MC) – 74-75c (20261)

Cinebench R23 (SC) – 68-70c (1587)

Gaming is when it gets really toasty, while playing Death Stranding core temps are around 84-89c and I even saw a spike to 91c! Coming from 6700K these temps have me little worried. In my old system GPU fans (GTX 1070) were the loudest part, but since TUF 3080 is so quiet, I can´t even hear it over the CPU and case fans when gaming.

 

Rest of the system is in my profile and the case (500DX) is well vented (2x NF-A14 front as intake, NF-S12A rear and BQ Purewing 3 top as exhaust). Since the temps were this high I even tried completely disabling PBO and all temps above are with PBO disabled. Everything else is stock and ambient temp is around 23-24c. When stable (non-beta) BIOS with PBO2 will be available for my MB I am going to UV which could help significantly from early reports I saw.

 

My question is. If you have 5900X, what cooler do you use and what temps are you getting? Epecially when gaming (real-world usage).

My r23mc is right at 23000

My r23 sc is around 1630.

 

My temps were hitting mid 70s on mc and 45-50 ST, but when swapping out video cards I noticed the tension bolts on my cpu block weren't bottomed out... so not sure if the new springs after a few days had more play or if I just didn't do it the first time, but since my temps are 3-5 c cooler so I barely hit 70s under MC.

 

Now my chip after undervolting and PBO adjustments is hitting 5100+ single core and 4.7-4.8ghz MC.

[BornCZ]

On 1/19/2021 at 6:59 PM, AngryBeaver said:

My r23mc is right at 23000

My r23 sc is around 1630.

 

My temps were hitting mid 70s on mc and 45-50 ST, but when swapping out video cards I noticed the tension bolts on my cpu block weren't bottomed out... so not sure if the new springs after a few days had more play or if I just didn't do it the first time, but since my temps are 3-5 c cooler so I barely hit 70s under MC.

 

Now my chip after undervolting and PBO adjustments is hitting 5100+ single core and 4.7-4.8ghz MC.

That are impressive numbers for 5900x even with custom loop. I tried undervolting as well, but I only get around 4.95 on SC, and around 4.2 on MC during CB r23.