AMD Ryzen 9 7950X Cooling Requirements & Thermal Throttling

[marios15]

Has anyone actually measured the surface temperature to see if it matches the reported temps?

 

[TheoneandonlyMrK]

Has anyone actually measured the surface temperature to see if it matches the reported temps?

Why would it.

The temperature sensors are in the chip not on its surface?!.

 

[cadaveca]

Has anyone actually measured the surface temperature to see if it matches the reported temps?

It'd be funny to learn the chips actually ran hotter.

 

[Punkenjoy]

Very good and informative article but I'm missing a key element: noise.
Not a single word about it.

How do you set fan curve with reasonable noise output when you get 95 degree all time?

As someone who is sensitive to noise, I want to tweek and fine-tune noise to preformance. I can't see how it can be done on zen4 cpu's.
In this regard, zen4 is a step back from past platforms and might be the deciding factor when buying my new system.

This is a "design choice" that turn into desing-flaw, if no matter the workload it causing maximum noise and no, setting 40-60% max noise at 95 degree is out of the question as part of what I do is long rendering. I know that on those occasions I will need to cope wite max noise, but at other cases it`s a no-go.

I suggest doing an iso-noise test or noise to preformance curve to see what is the sacrifice that auto 95 degree carry, because according to the article all is sunshine and flowers while disregarding a very important factor.

If you are looking for silent in a PC, i think AIO is the way to go. And on an AIO, you don't set your fan curve based on CPU temps. If you do, you are doing it wrong. You set it to water temp and then, it's the heat output that matter more, not the temperature. if you let your water temp go higher, there will be a higher delta between your room air temperature and water temperature and and a lower fan speed will be able to extract more heat.

But again, they have tested fan speed. Just lookup a review of these cooler, look how loud they are at different fan speed and you will get your results, you don't need to wait for a retest. The information you need is all here.

 

[W1zzard]

Has anyone actually measured the surface temperature to see if it matches the reported temps?

What @TheoneandonlyMrK said.

The temperature of the IHS is MUCH lower. I did a quick test with a thermal camera during first round of Ryzen reviews and the VRMs at 70°C were the hottest part in the socket area

 

[HD64G]

Many thanks to @W1zzard ! You answered many questions in this test. AMD changed the way the CPU perform depending on temps. Also a good aircooler will get us more than 95% performance vs a good AIO.

Now, only efficiency testing remains and as you wrote it's upcoming.

95C is the new 65C - That's just terrible.

I hope you wrote a similar post about 9900K back then...

 

[Dirt Chip]

There are loads of options.

Just set a fan curve based on load, interior case temps, I would use water temperature personally in my rig.

Or just set a noise optimised top end and leave it to AMD to give you the max performance for your setup in that profile.

Well, the only good option as I see it, according to the article, is to set a max fan speed that is about 60% while losing ~1.5% performance.
This will give me peace of mind regarding noise. Still, I dont like it but it`s a compremise I can stend if other factors are in favor of 7950X/7900X and zen4 system as a whole.

I guess what I really don`t like is the dis-correlation between cpu temp and noise. You can work around it but it will never be as good and simple as more heat=more noise.

 

[R0H1T]

Another similar take on this ~

So looks like 150~200W PPT is the sweet spot under that you can get better efficiency, but performance likely suffers. Beyond 200W performance doesn't increase much in the workloads he tested.

 

[r9]

Many thanks to @W1zzard ! You answered many questions in this test. AMD changed the way the CPU perform depending on temps. Also a good aircooler will get us more than 95% performance vs a good AIO.

Now, only efficiency testing remains and as you wrote it's upcoming.


I hope you wrote a similar post about 9900K back then...

It's terrible for any CPU/GPU or any IC chip for that matter.
Even if the CPU can handle it(for how long that's another matter) it will heat at other components around it that don't like 95C.

 

[Dirt Chip]

If you are looking for silent in a PC, i think AIO is the way to go. And on an AIO, you don't set your fan curve based on CPU temps. If you do, you are doing it wrong. You set it to water temp and then, it's the heat output that matter more, not the temperature. if you let your water temp go higher, there will be a higher delta between your room air temperature and water temperature and and a lower fan speed will be able to extract more heat.

But again, they have tested fan speed. Just lookup a review of these cooler, look how loud they are at different fan speed and you will get your results, you don't need to wait for a retest. The information you need is all here.

AIO will not go inside my system- I dont like the idea of water in my computer. Also, the cost is too much compere to a good fan coolers.
I bouth an NF-U12A and it will need to be enough.

You are right that all the information is there, I just needed some time to see it
The article+forum helped me understand that the 95 degree is just a minor problem with plausible workaround.
Only the easy factors of performance and cost vs. RL need to be sorted out and i`m done.

 

[W1zzard]

Yeah, rendering, photo editing, programming/compiling.

But don't worry about it, I understand there are more aspects to review about these. All I was saying is, if I had to make purchase decision, I would use my head, on top of the figures presented in the article.

All tested on Wraith Spire @ 100%
Compile: jumping between 85 and 94°C
Photoshop: around 75°C, jumping between 60 .. 88°C (this is a scripted workload, with no gaps between actions)

 

[TheDeeGee]

You can run an advanced Ryzen CPU with an NH-D15 on low fan RPM, as shown, and still enjoy its performance with very high clocks

Yeh, well dumping 95C of heat into my room during summer isn't what i'm after either.

Just a terrible design by AMD.

 

[dj-electric]

Yeh, well dumping 95C of heat into my room during summer isn't what i'm after either.

Do you know how the 2nd law of thermodynamics work?
Honestly asking. Because it gives you the answer why you are wrong about worrying about the CPU being 95c or 80c or whatever.
I'll drop a hint, all you should care about is power consumption, not temperature, it matters not in your case.

 

[HTC]

It's terrible for any CPU/GPU or any IC chip for that matter.
Even if the CPU can handle it(for how long that's another matter) it will heat at other components around it that don't like 95C.

If the CPU were monolithic and big, your point would have more merit.

These CPUs use tiny chiplets so the heat is concentrated on very small areas, which is why it's so hard to have lower temperatures.

I wonder if it's possible to check ambient temp inside a case before turning the computer on and after performing one of the more "heavy duty" benchmarks: to see the delta between the temps.

EDIT

Preferably, without a GPU so that one could rule that as a source for the increase in temperature: now that these CPUs have an IGP, that's quite doable.

 

[thestryker6]

Silicon that runs at lower temperature is more efficient, too, maybe that's what happened. I doubt the power sensors have that much thermal drift, still, that's part of the reason I do proper power measurements with test equipment and don't trust the in-chip power sensors.

Yeah it was also running slightly faster too so I assumed it came down to efficiency. I'm sure he'll do more testing as the equipment is finalized, but everything for mounting etc was all WIP or jury rigged for the intial testing.

The fact that you measured temps around the socket and VRMs were hottest really speaks volumes as to the terrible thermal transfer.

 

[bug]

What @TheoneandonlyMrK said.

The temperature of the IHS is MUCH lower. I did a quick test with a thermal camera during first round of Ryzen reviews and the VRMs at 70°C were the hottest part in the socket area

Ok, this is important. Because many water cooling kits don't cool the VRMs, iirc.

 

[W1zzard]

Has anyone actually measured the surface temperature to see if it matches the reported temps?

http://imgur.com/vWIlKQK

This is running Blender @ 100% fan speed on the Wraith Spire

 

[HTC]

Ok, this is important. Because many water cooling kits don't cool the VRMs, iirc.

And with this, my NH-C14 cooler is better than most other air coolers ...

 

[Punkenjoy]

It's terrible for any CPU/GPU or any IC chip for that matter.
Even if the CPU can handle it(for how long that's another matter) it will heat at other components around it that don't like 95C.

Well it's not the 95C that will do that, it's the 200W + that will do that. If you take a soldering iron, the tips get super hot, but you could leave it On in your case and as long you have minimal airflow, it won't heat at all component around.

The CPU will adjust it's power usage based on the cooling capacity it have. if you have a bad cooler, it not be able to use as much power and it will release less heat in it's surrounding. If you have more cooling, more of that heat will be extracted from the CPU to the surrounding air.

If your case have very bad airflow, the efficiency of that cooler will drop, the CPU will lower it's power usage to keep the 95 C° and less heat will be dumb in the case. If you have a good airflow in your case, it will just be extracted.

AIO will not go inside my system- I dont like the idea of water in my computer. Also, the cost is too much compere to a good fan coolers.
I bouth an NF-U12A and it will need to be enough.

You are right that all the information is there, I just needed some time to see it
The article+forum helped me understand that the 95 degree is just a minor problem with plausible workaround.
Only the easy factors of performance and cost vs. RL need to be sorted out and i`m done.

Yeah, well AIO is a matter of taste, they are very good for many things but if you aren't comfortable i understand that. There are still plenty of very good air cooler. A good air cooler in a good case go a long way.

I still prefer AIO and i am looking at building a custom loop, but i can see that it's not for everyone

Yeh, well dumping 95C of heat into my room during summer isn't what i'm after either.

Just a terrible design by AMD.

You dump watt in your room. (Watt are a measure of energy per second). You don't dump Temperature (a measurement of how excited atoms are) in your room.

2 CPU, running at 200w will dump the exact same heat in the room even if one run at 55°C and the other at 100°C.

running non stop a Soldering Iron where the tips is at 350°C wouldn't have any impact in your room temperature because it does it at such a low power (between 20-50w depending on the model).

 

[A Computer Guy]

Another thing people might learn in the kitchen is the notion of thermal capacity and energy.

Boil water (100 °C / 212 °F ) and stick your hand in it. You will probably remove it in a tenth of a second and immediately seek a cold water bath for your hand.

Preheat oven (100 °C / 212 °F ) and stick your hand in it. You can probably leave it in there for a while. Not a big deal.

Boiling water and steam are the same temperature (100 °C / 212 °F ). Which one is a nastier burn?

Heat oil to (175 °C / 350 °F ). Heat oven to same temperature. Throw in some chicken legs. Will they be done in the same amount of time?

A lot of people here simply don't understand basic concepts of heat because they are poor observers. Unfortunately at this point, the opportunity to teach them how to be good observers is probably over for most of them.

As we have seen, an entire article about heat was written and a lot of people still don't get it. It's not like a Q&A forum discussion is going to fix deficiencies in observational skills.

Your making me hungry. I need a boiled egg, baked pastry, and some coffee.

This would be 100% a no go as silent enthausiast.

There is a reason i went with a overkill NH-D15 on my Non-K 11700, it means it only has to spin 600 RPM to keep it 45-55C during gaming.

Why can't you just fixed your fan speed to 600 RPM with Zen4 then?

 

[natero62]

naah, so much good defends,even seen 95c temp..i dont get it...really.

when 12900k goes high oc'd near 98c its hot and so on.. and thats it...and it is 10nm cpu.

now when 5nm ryzen 7950 goes 95c, even any cooling its call ok, 'now 95c is 65c' come on....noway i say.

95c is 95c,thats it, i seen and feel just alot amd sympaty and gift for ryzen, its not earn it, i think..
i cant help it...again, its 5nm cpu.

i dont like that even you use 420 premium water cool loop, no help,so its clear ,ryzen7000 series plan and build is fail, period.and, amd know it.

i hope techpowerup can compare it 12900k same details,and again,12800k is 10nm cpu,so its 13900k cpu.

also, if it even beat clear 10nm 12900k, i except that high temperatures,, but no, not even that happen.

i see only fails ryzen 7000 series for that high temp.

its not mean its lausy cpu,no,but its not good one,nope

 

[Dirt Chip]

Yeh, well dumping 95C of heat into my room during summer isn't what i'm after either.

Just a terrible design by AMD.

You dont dump celsius degrees. You dump heat. They are NOT the same.
Celsius degrees is a unit we use to mesure the magnitude of heat, not it`s amount. You can have constant very small amount of heat at 95 degree that will not change your room temperature over a year or the same constant 95 degree but at a huge amount that will knock you out in a a few minutes.
The amount of heat is given by the CPU TDP\PPT (wattage) and the duration of the process.
For a given TDP/PPT and time you dump the same amount of heat, no matter the celsius degrees.

The cooler you choose will determine how effectively you move the heat from the CPU\computer to the room.
A very strong cooler will do it quickly while staying cool, a medium cooler will do it slower and will get somewhat worm in the process and a weak cooler will do it slowly and get very hot but in all cases all coolers will absorb the same amount of heat from the CPU and eventually dump the same amount of heat to the room.
The celsius degrees that the CPU reach is relevant to heat output (it is relevant to frequencies, performance and throttling)- only the amount of heat it's producing and it`s not the same thing.

 

[HD64G]

It's terrible for any CPU/GPU or any IC chip for that matter.
Even if the CPU can handle it(for how long that's another matter) it will heat at other components around it that don't like 95C.

Did you see that the post @W1zzard earlier wrote about Zen4 chip that it emits 70C heat at worst?

http://imgur.com/vWIlKQK

This is running Blender @ 100% fan speed on the Wraith Spire

 

[TheoneandonlyMrK]

It's terrible for any CPU/GPU or any IC chip for that matter.
Even if the CPU can handle it(for how long that's another matter) it will heat at other components around it that don't like 95C.

@birdie too.

Wtaf are you two on about, 95 is high is it balls.

Heat dissipation is not new, chip's including GPU'S, vrm and many more have done year's at these high temperatures in millions of applications yearly.

Just because you like idle temps at all times doesn't make you right, because you aren't.

And some of us that actually used they're pc AT these temperatures already Know it's fine.

Like Intel Knows it's fine hence 12900K and kf and KS, and RL will be the same.

Stick to quad core non k then and be done with your BS.

 

[Tom Yum]

It's terrible for any CPU/GPU or any IC chip for that matter.
Even if the CPU can handle it(for how long that's another matter) it will heat at other components around it that don't like 95C.

Yeh, well dumping 95C of heat into my room during summer isn't what i'm after either.

Just a terrible design by AMD.

My case in point. Again, temperature doesn't equal heat flow. The temperature of an IC is a product of its heat dissipation and the resistance to heat flow ( or thermal resistance) of its cooling system (usually measured in delta milli degrees Celsius/watt, with the delta being from ambient temperature).

A CPU consuming 100W power will dissipate 100W of heat, period. It's temperature doesn't change the amount of heat dissipated. If the cooling system has a heat flow (thermal) resistance of 0.5 degrees Celsius/watt, and the ambient temp is 30C, then the CPU will be 80C. If the CPU boosts power draw to 130W, temperature will rise to 95C, everything else being equal. A cooling systems thermal resistance can change depending on airflow, as the testing showed. Only passive systems have static thermal resistance.

Zen 4 boosts based on temperature. If it is at 80C at 100W and it has frequency and voltage headroom, it will boost clocks and voltage until it hits 95C. In the example above, once it hits 130W sustained power consumption (=heat dissipation). What surrounding components 'see' though is not 95C, it is the 130W of heat being dissipated into the ambient air around them. The temperature of the ambient air rises (because it is absorbing 130W of heat, minus whatever heat transfer outside the system via case airflow), and those surrounding components heat to different temperatures based on their own heat dissipation and thermal resistance.

If the above isn't intuitive, think about your phone running an intensive game for 10 min. If you checked you'd likely see the phone SoC operating at 90-95C, yet does it heat the room? No, because it is only dissipating 3-5W of heat. The thermal resistance of the SoC's cooling is high (~tens of degrees Celsius/watt), because it has no airflow and little surface area to radiate heat with. So it's SoC temp is high. But you don't worry about it overheating your laptop sitting nearby.

Long story short, power consumption is what matters, not temperature. Worried about system components or room temps? Focus on reducing power consumption, not component temperatures.