Noctua NH-U14S Getting 90C at 4GHz

[tomdean]

Good p95 info! In short, do you agree that temps are normal for the p95 load and his cooler, maMac?


Tomdean, why do you not want more exhaust than intake? The dust difference is negligible in most cases. But that said, if your temls dropped about 2C removing the side panel, it doesn't seem like it's struggling for airflow. I'd still add at least one fan for exhaust, however. Cases aren't really sealed so 'pressure' differences are negligible (air flow, CFM, isn't ).

Please read my post. I specifically wrote " I do not want a negative pressure in the case."
I was lookig for help. Guess this is the wrong place.

 

[freeagent]

Another thing I noticed was the size of the cooler.. its pretty small.. I know they say it doesn't matter, its how you use it.. but it just looks like it could use a hamburger. I wouldn't use that on my 5900X even if it fit.

 

[EarthDog]

Please read my post. I specifically wrote " I do not want a negative pressure in the case."
I was lookig for help. Guess this is the wrong place.

I read it. Apologies if asking why was offensive. But you do understand my advice to add one fan (even two) still keeps you 'positive' anyway.......you have 4 intakes. Feels like I'm helping within your parameters, bud.

If you want more flow without fans, free the exhaust (remove filters, grating, etc.)

 

[tomdean]

I finally got back.
After some more testing with the side panel off, the 100% load temps are in the mid 70's C.

The Be Quiet! Dark Base 900 Full Tower case has a filter and vents on the bottom. It is unlikely that I can make much of a negative pressure in the case.
I ordered two 140mm fans - they came yesterday. Now, to mount them on the case and run some tests. Tomorrow.
BTW, I received Windows 10 and plan to install that so I can run in the same mode it seems like the rest of the group does. That should help.

Thanks for the patience and help.

 

[EarthDog]

After some more testing with the side panel off, the 100% load temps are in the mid 70's C.

Im not even going to ask how you have a 20C difference now, from 2C earlier in the thread..or what test you're running now. If 70C is your stress test load temp like post 1, I wouldn't change a thing..

Be sure your testing is consistent. Use the same test...same time periods (30+ for akr cooling), and be sure the temps comes back to your starting point before beginning another test. Looking forward to seeing what you come up with.

 

[tomdean]

I finally got around to testing the system with the case closed and two 140mm fans exhausting on the top.
Looks like the system is 4C hotter with the case closed. Maybe need one more fan on top...



The case open run was without the two 140mm fans on top.
Both tests were run with 100% load.

 

[freeagent]

High performance and silence almost never go hand in hand. You gotta move some air.

 

[EarthDog]

Case off is typically going to be better unless you have a super high airflow case (you do not - it's a 'closed' front with mesh vents). Reviews confirm this.

If you were testing temps with the side off/on, the only variable that should change is the side on/off. You changed the fans so the datasets aren't giving you an accurate answer (accurately anyway). In the end, to me, it's not worth a retest, I doubt it would make much of a difference (and that knowledge gains you little), but that's empirical testing.

Aren't you watercooling the CPU now anyway? You want airflow and don't care about noise so........... just install them!

 

[tomdean]

I have not received all the parts to install the water cooling.

Funny. The last part is from Auburn, WA, just across the sound. To get here, you travel South to Tacoma, hang a right, and go back North. FedEx missed the turn in Tacoma. If you do that, you wind up in Oregon, California, or Cabo. Fedex caught the problem, I hope, just South of Portland. Maybe I will get it today!

I plan to install water next week.

 

[roidroo]

I mean. I'm not surprised.
280W TDP with a 'smaller' air cooler. Sounds about right with P95?

 

[trents]

tomdean, it's not clear to me if you know this but CPUs will begin to throttle as they approach max safe temps.

 

[Keith Myers]

If you intend to run a compute intensive workload at 90-95% utilization on Ryzen or Epyc/TR, then you need to forget using any sort of automatic clocking or boosting. The cpu will always throttle its clocks down to stay within the power budget and maintain safe temps.

Trying to cool a 24/7 200W or greater load calls for water cooling. I run all my hosts with fixed clocks on all cores at reasonable voltages to maintain stability and keep the temps down to sensible values. I can get much better overall clocks compared to any autoboost algorithm. Using @tomdean's example of P95 and having the clocks fall down to 3.5Ghz on small FFT's, I can run all-cores at 4.4Ghz and keep the temps under 80° C. on my 5950X for example.

Same goes for my Epyc 7402P and 7443P. I can run all cores at almost the spec boost clock of 3.35Ghz on the 7402P and setting the cTDP and PPL limits to 200W and end up with all cores running ~3.34-3.35Ghz and keep the temps to around 45° C. using a water blocked cpu and 360 rad.

When I ran my TR 2920X at 4.1Ghz all-core I could keep temps below 62° C. Just have to make allowances for the power budget and what that requires for cooling. Asking air-cooling to handle that amount of heat dissipation is a frustrating exercise.