Question Cooler Master Hyper 212 vs Stock Ryzen 5600/X cooler.

[brad1138]

I built 3 new computers for my family. I put Ryzen 5600 in 2 of them and 5600X in mine. I don't really plan on overclocking, especially after reading how it doesn't do so much with new gen CPUs. (I may play around with it a bit)

Anyway, I bought a Cooler Master Hyper 212 for mine and put stock in the others. I ran Prime 95 on all of them, expecting to see the 212 keeping my 5600X cooler then the 5600s, but it didn't. The peak temps on my 5600X/212 was a degree or 2 hotter at the end, than the 5600/stock.

So either the 5600X generally runs a lot hotter than the 5600, and the 212 kept it close, or the stock fan is nearly as good as the 212...

I am just curious, if nothing else, the 212 looks better, but if I had known the stock fan did that well, I could have saved ~$50...

I could put the stock cooler back in my rig and take readings again, but I don't really want to bother.

 

[hotaru.hino]

A few questions I have here:

• What was the temperature of the CPUs?
• What was the final steady state CPU frequency?
• What case do you have and what's the fan configuration in it?

And then there's the trouble that the Ryzen 5000 series die layout is not a rectangle in the middle, but one (or two) closer to the corners and one to the side. Which means a small dollop in the middle may not provide enough coverage.

However, if anything, the 5600X runs slightly faster than the 5600, which means it'll pull more power.

 

[brad1138]

A few questions I have here:

• What was the temperature of the CPUs?
• What was the final steady state CPU frequency?
• What case do you have and what's the fan configuration in it?

And then there's the trouble that the Ryzen 5000 series die layout is not a rectangle in the middle, but one (or two) closer to the corners and one to the side. Which means a small dollop in the middle may not provide enough coverage.

However, if anything, the 5600X runs slightly faster than the 5600, which means it'll pull more power.

5600 / 5600X

76 / 79 degrees (read from CPU)
76 / 78 Deg (read from MoBo)
4,450 / 4,650

There are variables, cases are not the same, but have similar cooling (1 intake & 1 120mm rear exhaust fan, + PS). Both were cool, at idle before test in a 65 degree room with moving air.

I know having variables like that makes it hard to compare, but the 2ish degree difference is constant.

When I am bored with time one of these days, I'll swap back the stock cooler and see. That it the only real way to test.

I guess I was just hoping someone might say they have noticed the 5600X runs 10 deg (or so) hotter. Then I would feel better about the 212.

 

[brad1138]

Also, stock fan ramped up to ~2,500 RPM, 212 up to about 1,900.

 

[hotaru.hino]

I'm leaning on this not really being a problem and falls within what's expected:

• The temperatures are no where near the limit (90C for the 5600, 95C for the 5600X)
• The clock speeds are actually higher than expected, as all-core workloads should be reducing the clock speed. Though it may be the thing you're looking at that's reporting clock speed is only reporting one clock speed and it happens to be the fastest one at the time.

Supposedly though the 5600X (and presumably the 5600) start dialing down clock speeds at 80C.

Although I'm also not a fan of using Prime95 to test thermals because it's not a realistic workload for most people. While I tend to use Cinebench, I'd call that barely any better because I also don't do 3D modeling/rendering. Try running something you'd actually do that would work the CPU and see if it makes a difference.

 

[brad1138]

I found this

View: https://youtu.be/omNtKUVIChg

which is very close to what I have. And I understand your question above now, about CPU freq. I'll need to run it again and observe better. And if I run cinbench, it will give me something for comparison. I am running a B550 Gigabyte MoBo also.

I'll run with the cover off.

 

[Inthrutheoutdoor]

Well, IMHO, worrying over a 2-3 degree difference in temps is about as silly as paying $50 for a 212... last time I looked (a month or so ago) they were running about $25-28 most places, at least here in the US anyways.... I know everything has went up in price recently but that's no reason to overpay for stuff that is basically basic in nature as the 212 is..

I used 212's on some of my 1st builds many moons ago and they did ok all things considered, but that was back when they were usually around $13-17, and cpu's didn't suck up nearly 200 (or moar) watts either....

 

[brad1138]

I never said I was "worrying" about it. Just curious, I thought it would be 5-10 cooler than stock.

This is the best price I could find

Are you a human?

www.newegg.com

EDIT: It was $50, just sold out.

 

[brad1138]

Looking around, 5600X idle temp with stock cooler is generally 45-50 degrees. Mine is idling under 30, so that is good. The 5600 is idling at about 30 with stock cooler. So maybe it just runs cooler.

 

[Karadjgne]

Idle temps in a Ryzen are pretty much meaningless. They aren't Intels. With Intel, at idle all the cores are cut equally in power and frequency, but each core is active so receives a portion of the load. The temp you see is the hottest read core at time of reading, which is barely above the other cores, having a somewhat equal load.

Ryzen does the opposite. Shuts down all the cores completely, except one. That single core sees the entire workload for everything, so is by far the hottest core, if the reading happens to hit when work is being done, you'll see a 40ish temp. If it happens to hit during a service startup, you'll see a 60ish temp as that one core is under a higher startup load. It's not representative of the entire cpu, which is far closer to 30ish. To prevent overuse, that single core is rotated to the next in line.

The amd coolers range from @ 130w to 180w, the Stealth being on the short side. The hyper212 is a 140w cooler. On a 65w cpu they'll be equitable in ability, but different in efficiency. The 120mm fan will spin slower and on that heatsink is more efficient, which makes it more effective at a lower volume. Because the fans can only slow so much, you'll get both fans at @ the same rpm at idle, but the higher efficiency of the hyper212 means lower average temps. At max, the efficiency and lower rpm is replaced by brute force of the higher rpm of the amd cooler, so ends up about the same temps.

5600x and 5600 are close enough in power usage to make no effective difference to a cpu cooler.

 

[brad1138]

I ran Cinebench R20 on both, with interesting results. With the stock cooler in the above video, it hit 80+ degrees after a 10 minute run of CB R20. And it started off running the CPUs at about 4.3GHz, dropping to 4.2GHz by the end. Final CB score was ~4200.

With the 212, mine peaked out at 63 degrees after 12+ minutes, and the CPU freq held at ~4.45 GHz the entire time. Final CB score ~4400.

The 5600 w/ stock cooler peaked at 69 degrees, but was only running at 3.85 GHz, which it held the entire 12+ minutes. Final CB score ~3900.

So, with that info, I am very happy with the 212, I can see it made a real world (more or less) difference. Almost 20 degrees cooler than the stock cooler in the video, and mine was in a case.

 

[Karadjgne]

Yes. Ryzen are dynamic cpus, Intel are static. When it comes to boost levels a Ryzen will boost according to loads, temps and voltages. As any of those go up, an algorithm starts dropping individual core boosts by 20-50MHz±, so with temps of 80+, you'll see a lower boost rate as compared to temps in the 60's, which affects the scores in judgement software like Cinebench or TimeSpy or GeekBench.