Do Intel’s Core i7-3000-series CPUs really need closed-loop liquid cooling? Today we're testing fourteen different LGA 2011-compatible air coolers on an overclocked Core i7-3960X in order to determine whose is the most effective.
It seems strange to us that Intel no longer includes a cooler with its fastest retail CPUs. This is, after all, the same company that seems to insist that we include our original boxed cooler any time we send in one of our purchased processors in for a warranty exchange. Instead, it now suggests that enthusiasts spend extra money on something that resembles Antec’s Kühler H20 620, but actually costs as much as the larger Kühler H20 920: Intel's own BXRTS2011LQ sealed liquid CPU cooling system.
The rationale behind this move does make sense in that the Core i7-3960X and Core i7-3930K are multiplier-unlocked products that specifically target overclocking enthusiasts. Enthusiasts typically won’t settle for anything as underpowered as its RTS2011AC downdraft cooler, and Intel doesn’t think hanging two pounds of copper and aluminum up to six inches from the surface of a motherboard is a great idea. And while that kind of leverage tugging on your expensive motherboard is probably ill-advised under the harsh conditions of shipping a machine cross-country, systems that are handled gently and don't get moved often tolerate the weight of big, heavy coolers without a problem.
We invited every major CPU cooler manufacturer (that's right, all of them) to show off their LGA 2011-compatible heat sinks, and fourteen responded with products ranging from $30 to $99. That takes either a great deal of confidence or a great leap of faith from the budget cooler companies, since everyone participating in this piece knew that we'd be basing our evaluation on overclocking. The wide price range allows us to really test how much we need to spend on cooling, and how much we can benefit from spending more money than necessary.
Following up on our earlier picture story, here’s a brief overview of each cooler’s basic features, grouped by price first and arranged alphabetically. Please pay special attention to each cooler’s base height, which is measured with the intake fan at its highest position, along with thickness and offset. If the fan overhangs your DIMM slots, its base height (plus around 0.3" for the CPU/LGA package) represents your maximum supported DIMM height.

| LGA 2011 CPU Cooler Features, Sub-$50 Models | |||||
|---|---|---|---|---|---|
| Arctic Cooling Freezer i30 | Cooler Master Hyper 212 Evo | Corsair A70 | Enermax ETS-T40 | SilenX EFZ-120HA5 | |
| Height | 6.3" | 6.3" | 6.3" | 6.3" | 6.2" |
| Width | 5.5" | 4.7" | 4.9" | 5.5" | 4.9" |
| Rad. Thickness | 2.6" | 2.0" | 2.8" | 2.8" | 2.3" |
| Total Thickness | 3.8" | 3.1" | 5.1" | 3.6" | 3.4" |
| Base Height | 1.2" | 1.4" | 1.5" | 1.5" | 1.4" |
| Assy. Offset | 0.5" Forward | 1.1" Forward | Centered | 1.0" Forward | 1.1" Forward |
| Cooling Fans | 1 x 120 x 25 mm | 1 x 120 x 25 mm | 2 x 120 x 25 mm | 1 x 120 x 25 mm | 1 x 120 x 25 mm |
| Connectors | 1 x PWM | 1 x PWM | 2 x Three-Pin | 1 x PWM | 1 x Three-Pin |
| Weight | 32 Ounces | 22 Ounces | 41 Ounces | 23 Ounces | 23 Ounces |
| Web Price | $40 | $35 | $35 | $30 | $33 |

| LGA 2011 CPU Cooler Features, $50-79 Models | ||||
|---|---|---|---|---|
| Akasa Venom Voodoo | Coolink Corator DS | Gelid GX-7 Rev. 2 | Xigmatek Venus SD1266 | |
| Height | 6.5" | 6.2" | 6.3" | 6.4" |
| Width | 5.1" | 5.5" | 5.1" | 5.2" |
| Rad. Thickness | 2.8" | 2x 1.8" | 2.5" | 2.8" |
| Total Thickness | 5.1" | 4.7" | 3.5" | 3.9" |
| Base Height | 1.3" | 1.5" | 1.4" | 1.5" |
| Assy. Offset | Centered | Centered | 0.4" Forward | 1.1" Forward |
| Cooling Fans | 2 x 120 x 25 mm | 1 x 120 x 25 mm | 1 x 120 x 25 mm | 1 x 120 x 25 mm |
| Connectors | 2 x PWM | 1 x PWM | 1 x PWM | 1 x PWM |
| Weight | 37 Ounces | 35 Ounces | 23 Ounces | 33 Ounces |
| Web Price | $55 | $50 | $65 | $65 |
.

| LGA 2011 CPU Cooler Features, $80-100 Models | |||||
|---|---|---|---|---|---|
| Deepcool Assassin | Noctua NH-D14 | Phanteks PH-TC14PE | Thermalright Archon SB-E | Zalman CNPS12X | |
| Height | 6.4" | 6.4" | 6.5" | 7.0" | 6.2" |
| Width | 5.8" | 5.9" | 5.8" | 6.7" | 6.0" |
| Rad. Thickness | 2x 2.0" | 2x 2.0" | 2x 2.1" | 2.1" | 2x 2.2" |
| Total Thickness | 6.0" | 6.2" | 6.3" | 3.1" | 5.2" |
| Base Height | 1.7" | 1.8" | 1.6" | 1.3" | 1.2" |
| Assy. Offset | 1.0" Forward | 1.0" Forward | 1.0" Forward | 1.0" Forward | 0.25" Left |
| Cooling Fans | 1 x 140 x 25 mm 1 x 120 x 25 mm | 1 x 140 x 25 mm 1 x 120 x 25 mm | 2 x 140 x 25 mm | 1 x 150 x 25 mm | 3 x 140 x 25 mm |
| Connectors | 1 x PWM 1 x Three-Pin | 2 x PWM | 2 x Three-Pin | 1 x PWM | 1 x Three-Pin |
| Weight | 37 Ounces | 45 Ounces | 47 Ounces | 36 Ounces | 36 Ounces |
| Web Price | $80 | $85 | $90 | $85 | $99 |
Since higher fan speeds typically increase cooling capacity at low cost, some of the more expensive models attempt to provide the best balance of cooling and noise. Today’s tests includes both heat and noise measurements, and we’ll even compare heat to noise, heat to price, and heat/noise to price!
- LGA 2011: No Boxed Solution, Pick Your Own Instead
- Installing Akasa's Venom Voodoo
- Installing Arctic Cooling's Freezer i30
- Installing Cooler Master's Hyper 212 Evo
- Installing Coolink's Corator DS
- Installing Corsair's Air Series A70
- Installing Deepcool's (Gamerstorm) Assassin
- Installing Enermax's ETS-T40
- Installing Gelid's GX-7
- Installing Noctua's NH-D14
- Installing Phantek's PH-TC14PE
- Installing SilenX's EFZ-120HA5
- Installing Thermalright's Archon SB-E
- Installing Xigmatek's Venus XP-SD1266
- Installing Zalman's CNPS12X
- Test Settings And Benchmarks
- Benchmark Results: Cooling And Fan Speed
- Benchmark Results: Noise And Acoustic Efficiency
- Which LGA 2011 Cooler Would We Buy?
By that logic, having no cooler at all is the most efficient... 0g/200c= 0 efficiency ratio. And a dead CPU.
Also, the higher the temperature (Bad), the lower the ratio, which doesn't make sense.
You would also need to use the ambient temperature delta rather than the absolute temperature in any sort of ratio for the results to be meaningful.
I think that the majority of people don't care how heavy their cooler is, only about the price.
On a side note i wish the "ZALMAN CNPS9900MAX-R" was reviewed as well considering on newegg it is compatible with LGA 2011, and I've been eye balling that one since it came out for my next build.
I've heard nothing but great things from CM's Hyper 212+ and Evo variants. Might I add the other powerhouse in affordable cooling, Thermalright's TRUE Spirit 120 for just about the same price. It seems the Hyper 212+ and TRUE Spirit have fallen off in Heatsink/Fan comparison charts despite kicking some serious ass against there competition price rise, and can even hang very well against high-end coolers costing 2-3 times as much.
I realize you have to compare modern products to modern products for the sake of it, but just a FYI for those not familiar with the cooling scene. Don't ever count out a product that first debut 2-3 years ago, they can still hang, the good ones at least.
Water cooling is a lot more expensive and a lot more complicated. Water cooling systems are not maintenance free and always add the risk of a leaking pipe.
$80-$90 is a small price to pay for getting a quieter PC without resorting to water cooling.
Example:
Noctua NH-D14 weighs 900g without fans and it did 45c at full load.
900g/45c= 20.00 efficiency ratio.
Panteck PH-TC14PE weighs 970g without fans, performing at 46c.
970g/46c= 21.09 efficiency ratio.
CM Hyper 212 EVO Weighs 580g with fan, performing at 51c.
580g/51c= 11.37 efficiency ratio.
Ideally, the lower the ratio, the more efficient a cpu cooler is. Other charts count as well when making a final decision.
I checked newegg and amazon
seems like a good replacement for my sunbeam core contact heatsing
By that logic, having no cooler at all is the most efficient... 0g/200c= 0 efficiency ratio. And a dead CPU.
Also, the higher the temperature (Bad), the lower the ratio, which doesn't make sense.
You would also need to use the ambient temperature delta rather than the absolute temperature in any sort of ratio for the results to be meaningful.
I think that the majority of people don't care how heavy their cooler is, only about the price.
I am not saying that i need mass to make my decision. And i agree that surface is important. All i am saying is that i want to see cpu heatsinks to be more efficient or equal at cooling with less metal.
Great cooler and I am happy for the fact I choose it over others in the same price range
Also, controlling fan speeds with A70 is simple with the Asrock MOBO. Just don't plug it into a CPU header--plug it into a 3-in Chassis header and let Speedfan set the speed (with voltage regulation) based on CPU temp. With the inline resisters, I can get RPMs down to 1000-1100, which seems pretty quiet. Seems to have little effect on cooling - weird thing about the A70 is that higher CFMs don't change much unless you hit the 130+ levels. Maybe it doesn't conduct well enough?
Seems like a very biased review all around. If you're gonna give a prize to the Noctua NH-D14 despite its price and size on the basis that it's one of the top performers and it's quiet, you should have also tested the A70 at the lower fan speed and accordingly given it OR the Hyper 212 Evo a prize for being the best bang-for-buck for cooling performance or noise-to-performance-to-value ratio.
People that are gonna buy the i7-3820 probably care a little about value, too, and the money saved from getting a cheaper cooler could go somewhere else--like a faster/higher capacity SSD, more RAM, a more feature-rich motherboard, etc.
Any chance of re-testing this?
Many thanks for the article, very helpful.
I remember them at one time having some of the best coolers around.
No, biased as in "you used a crappy testing methodology (by not using the included inline resistor for the A70 which would in turn lower fan noise)" and gave the top-performer the only award because, according to you, people buying i7s have completely disposable incomes. According to you, the people buying into this platform don't give two craps about performance for the money, and it doesn't matter that they could use the money saved from going with a less expensive cooler for something else like a better SSD or a more feature-rich motherboard.
Again, if you're gonna give the Noctua NH-D14 an award, you need to give one either to the Hyper 212 Evo or the Corsair A70, too, on the basis that they deliver the best bang-for-buck. With the A70 we don't even see an accurate comparison regarding fan noise, so it's impossible to know if it would've been the jack of all trades.
I've been using Thermalright coolers in my builds recently, however on one build, I combined a Thermalright CPU cooler with a quiet Scythe fan, and several quiet Scythe case fans in an NZXT Gamma chasis. The result is much quieter than I expected, however, the system is well cooled.