Temperatures, Noise, and Conclusion
Test Platform | |
---|---|
Processor | Intel Core i7-6800K |
Motherboard | EVGA X99 Micro2 |
Memory | HyperX Fury 16GB (4x4GB) 2400 MHz DDR4 |
Graphics Card | XFX AMD Radeon 5450 (Fanless) |
Storage | OCZ Vertex 460 120GB SSD |
Power Supply | Corsair TX 650W PSU |
OS | Windows 8.1 64-bit |
Temperatures
Load temperatures for these air coolers were generated using the x264 video encoding benchmark running in 64-bit mode. Stress temps were generated using the 64-bit version of the Prime95 benchmark, with the in-place large FFT torture test to force maximum power consumption.
As good as the Le Grand Macho RT performed, the Noctua NH-D14 cooler demonstrated why it has been a fantastic choice for quietly cooling heavy CPU loads for years now. I tested the Scythe Ninja 4 with fan at its "high" position, as this setting provides the best performance without a measurable increase in sound, and it finished just behind the leaders. Just a few degrees behind with the Hyper 212 EVO, which performs so well that it begs the question, why spend more?
Now we'll see how the coolers held up under more strenuous tests, with the Core i7-6800K increased to 4.0 GHz on all cores, with a 25mV voltage bump to stabilize the overclock.
The coolers finished in the same order, though the NH-D14 did extend its lead here with the more challenging loads. The Ninja 4 outperformed the Le Grand Macho RT under a standard x264 encoding benchmark load here, but fell to the larger cooler when pushed with Prime 95's torture test.
Noise
The Ninja 4 is remarkably quiet, failing to register over the noise floor even at the "high" fan setting. (See our review of this cooler here, which "broke" my sound test process with its ultra-quiet operation). It's followed very closely by the NH-D14 and Le Grand Macho RT, both of which produced just enough noise under load to hear, but only at close range. The Hyper 212 EVO obviously measures much louder, and it is; but the 42.8 dBA noise output under full load has a low tone (no fan 'whine') that, while very noticable on my open test bench, isn't annoying.
Test Winner: Noctua NH-D14
The venerable Noctua NH-D14 bested this group in both stock and overclocked temperature benchmarks, and was just a hair behind the ultra-quiet Scythe Ninja 4 in noise output. It could be argued that the NH-D14 had an advantage as the only dual-tower (and dual-fan) cooler in the roundup, and that should be taken into consideration when evaluating these results. I'll be curious to see if other larger coolers can best it in future tests (more air cooler reviews are coming soon).
The other coolers tested all performed extremely well, with Thermalright's Le Grand Macho RT proving that its i5-6600K results were no fluke, and the Scythe Ninja 4 proving that a virtually silent cooler can still tame very heavy CPU loads with the best of them. And, last but not least, we have the incredible value of the Cooler Master Hyper 212 EVO. This cooler, which has long been #1 on Amazon in this category, kept the i7-6800K cool enough in these benchmarks to convince me that $29 would actually be enough of an investment even for a $400 enthusiast processor; even if the noise levels under load are clearly going to be higher than the larger (and more expensive) coolers in this roundup.
Final Thoughts
The Noctua NH-D14 may have come out on top in this roundup, but all four of the coolers on test present their own argument for selection, given their various price-points (and all would receive my enthusiastic endorsement). The NH-D14 did have the advantage of being the only dual-fan option here (as mentioned above), and its price ($79.99 for this SE2011 version with PWM fans) ties it with the Le Grand Macho RT for most expensive cooler in the group. In sharp contrast, Cooler Master's $29 Hyper 212 EVO had a strong enough showing that on price alone it will remain at the top of the list for many buyers.
In the end, there can be only one winner. And though the NH-D14 was officially discontinued with Noctua's NH-D15 now available, it is still readily available in both the SE2011 version here, and the original, less-expensive model without PWM fans ($69.95, Amazon). But you really can't go wrong with any of the powerful coolers in this group, as they all kept the processor well below thermal limits even under the most unrealistic Prime95-induced extremes.
I wish you had included the
I wish you had included the Intel(Delta) TS15A for reference. I bought one for my i7-6700K build and id like to see how it compares to the giant coolers i am always skeptical of.
That’s an LGA1151 cooler.
That's an LGA1151 cooler.
I’m sure he was referring to
I’m sure he was referring to the 2011 counterpart the TS13A
You should check out
You should check out AnandTech’s recent “Battle of the Stock Coolers” article. They compare (among other things) the TS15A, AMD Wraith, and a Hyper 212 Evo for reference. Long story short: besides the Wraith, nothing comes close to the Evo in neither thermals or acoustics, and all Intel coolers perform roughly the same when paired with their intended wattage (i.e. a 65W Intel stock cooler at a 65W load results in the same temps and noise as a 120w cooler at 120W load).
http://www.anandtech.com/show/10500/stock-cooler-roundup-intel-amd-vs-evo-212
I wonder what the difference
I wonder what the difference in performance would be if you added a second fan to the CM Hyper 212 EVO? It’s bound to help some but is it enough to notice?
on my old 2500k running @ 4.8
on my old 2500k running @ 4.8 and 1.5 volts it makes almost no difference (it cannot keep up at 5.0 GHz). as I adjust the voltage down the best it will do is between 1-3C. In the end all that it really did is allow me to keep both of the fans running at a lower rpm to achieve the same cfm of air flow.
eventually you will just reach the heat dissipation ability of the amount of copper and aluminum that you have with a hyper 212.
What about the SilentiumPC
What about the SilentiumPC Fortis III cooler? It’s even cheaper than the Hyper 212 Evo. I wonder how it compares to it
I’ve never heard of it.
I've never heard of it. Searching Newegg and Amazon show no results… Appears not to be available in the U.S.A.
http://www.kitguru.net/compon
http://www.kitguru.net/components/cooling/anton-shilov/intel-rolls-out-its-most-advanced-air-cooler-to-date-ts15a-for-lga1151-processors/
http://supremelaw.org/systems/heatsinks/warning.htm
And, compare the
And, compare the spring-loaded machine screws
on these Intel HSF units:
http://www.kitguru.net/components/cpu/anton-shilov/intel-demos-its-cooling-solution-for-core-i7-5000-haswell-e-chips/
Is it a problem that the 212
Is it a problem that the 212 design is not completely flat on the plate that contacts the CPU? I had to reapply a couple times, finally settling on the dap in the center method, but my CPU always throttles down under load due to max temps. My next cooler will be a different brand for sure, or one with the thermal paste already applied.
Which CPU is it that’s
Which CPU is it that's throttling? I have found that AMD FX CPUs throttle based on some apparently pre-defined perameters (doesn't seem absolutely dependant on temps, either). In my use of the EVO with Intel CPUs I haven't had any throttling, but I haven't had temps hot enough to cause it (have never even seen 90 C with it).
Still, the EVO certainly isn't perfectly flat, and I imagine there are variances between units.
I’ve seen a test done
I’ve seen a test done comparing flat vs convex vs concave cooler surfaces and the convex one was actually the best. In other words, if the cooler’s mating surface is designed to contact in the center instead of being flat – and some of them are – that is a good thing. Temp variations were small so I wouldn’t worry about it unless the outside of the cooler is touching before the center (concave). If that’s the case you have a defective unit.
Otherwise, the only reason you should see thermal throttling with a 212 EVO is that 1) You forgot to install thermal paste or incorrectly applied it, or 2) You forgot to plug in the fan, or 3) your overclock attempt is too high. The 212 EVO isn’t intended to be wring all the possible OC you can get but is generally good enough for moderate overclocks.
Other possibilities – questionable CPU, bad cooling in the rest of the case. Really, considering my experience and the huge number of people that love the EVO I think it’s the least likely source of your problem. But then I’m not seeing your situation in person so can only guess. 🙂
The 212 is one of the best
The 212 is one of the best budget coolers on the market, but it’s also a pain in the ass to install. Lol I would gladly pay an extra 5-10 bucks for a better mounting solution.
Thanks for this review, glad
Thanks for this review, glad to see my good old 212+ is up to date. My 2 cents about this HSF is to change the single rattly fan to two decent push-pull fans. I have this config since a Core 2 Duo many many years ago and the cooler seems to be fine for a Broadwell-E, how good of an investment is that?
I think, also size, weight
I think, also size, weight and aesthetic important paremeters. 1240 gr and dual fun ugly Noctua not winner overall !! imo.
and not as silent as ninja.
Thanks for the great review!
Thanks for the great review! Very informative look at these coolers.
Thanks!
Thanks!
why not tests against other
why not tests against other old monsters like cm v10?
My collection is limited to
My collection is limited to more recent coolers. Besides those tested, I have a couple of low-profile air coolers and some AIO liquid coolers on hand.
Would you run some of those
Would you run some of those AIO liquid coolers through the same tests in this review, to compare their performance vs. the air coolers? That would be some good perspective.
Typo in test platform specs
Typo in test platform specs box, should say i7
Fixed! I don’t think Intel
Fixed! I don't think Intel has announced any i5 processors for X99 🙂