| Best CPU Cooler Performance: Intel Q3-2010 |  |
| Reviews - Featured Reviews: Cooling | |
| Written by Olin Coles | |
| Wednesday, 17 November 2010 | |
|
Page 12 of 16
Heatsink Performance: Stock Cooling FanThis section concentrates on the cooling performance of each kit with the included fan. Manufacturer-supplied stock cooling fans usually offer either extremely high airflow or incredibly low noise, so there's a lot riding on what's packaged with the kit. Sure, there's added importance on the cooler's design and construction, but at the stock level these factors really don't carry tremendous weight. These results best compare products right out of the box. For the purpose of this article, all CPU-coolers have been orientated so that heatpipes span horizontally from front-to-rear with fans exhausting upward and not top-to-bottom with fans blowing towards the rear of the computer case. While the difference may be minimal, performance enthusiasts will want to use every technique available to reach the highest overclock possible. Benchmark Reviews has tested several new products against some proven top-performers from previous tests in this section, all using stock fans on the Intel LGA1366 socket. The average temperature difference (core temp minus ambient temp) is noted beside each heatsink:
Best CPU Cooler: Stock Fan PerformanceLeading the pack is Thermalright's Venomous-X heatsink kit, which adds a basic 120mm cooling fan with the RT version of the kit. After five recorded test runs on the overclocked/overvolted i7-930, the Venomous-X averaged 38.98°C above the 20~21°C ambient room temperature. Next came the Zalman CNPS10X-Performa heatsink, which cooled to 39.32°C over ambient with stock fan and trailed behind by a mere 0.34°C on average. Xigmatek's Aegir SD128264 heatsink was only 1°C behind the leader, cooling to 39.78°C in stock form. Schythe's jumbo-sized Mugen 2 heatsink kept cool using the included Slip-Stream fan, and averaged 41.39°C over. The Titan EVO NK85TZCS2 heatink comes with a fairly powerful fan, and cooled to 41.83°C over ambient on average. The next 'step' in heatsink performance occurs with the Cooler Master N620, a cooler that uses twin 120mm cooling fans in push/pull configuration, and cooled down to 42.95°C over ambient. The Xigmatek Balder recorded 43.18°C over ambient temperatures, nearly matched by the Scythe Ninja 3 at 43.25°C over, and Cogage TRUE Spirit with 43.40°C. The Scythe Rasetsu trailed behind and cooled to 43.98°C over ambient with the included Slip-Stream fan. Set to 'High', the CoolIt Vantage ALC all-in-one liquid cooler reduced overclocked temperatures down to 46.89°C over ambient, and remained extremely quiet during operation. The Deep Cool Gamer Storm heatsink offered very aggressive looks, but the stock cooling fan was silent and didn't stir up much air as a result, giving the Gamer Storm 49.32°C over the ambient room temperature. Taken as a whole, every single heatsink tested here performed very well with our heavily-overclocked Intel Core i7-930 processor running at 1.40 volts vCore. In the next section, we re-test the collection with a pair of Noctua NF-P12 cooling fans in push-pull configuration...
|
|
Comments
Guys where is V6GT? from precedent review was a very good performer,
This is a good one guys. Something that I can link to and refer people to that have questions about CPU cooling solutions. While there is always gonna be someone who brings up questions about why wasn't my 'Blah-Blah Cooler' included, you have a good assortment here and tested them all fairly.
Thank You
in this review was used different hardware...so the results could be different.....anyway ...
Other reviews show it as a good HSF, but will be great to see it included on a comprehensive comparison as this one
:(
Are there any plans to record noise levels? Obviously (based on your review) using two Noctua fans is a silent or near silent solution, and strapping on that Yate Loon is a recipe for distracting the neighbors, but what about the stock fans? I'm willing to spend money on a decent solution (namely the SCMG-2100) but may not have the budget for two super fancy fans at ~$15 a pop. Plus, if spending $30 is only going to get me 2db less than the included Slipstream, it's just not worth it.
Have Fun!
It's been a while since we tested an overclocked i7 CPU with the stock cooling solution... generally because it's too dangerous. Here's an example from the last time we did: benchmarkreviews.com/index.php?option=com_content&task=view&id=371&Itemid=62&limit=1&limitstart=11
Scythe Mugen 2, default fan at maximum speed
i7 950, 4GHZ, 1.4Vcore
Artic Silver 5
Everest 5.5
Ambient room temp: 23c
Stress CPU and FPU selected and run for 30 min. All temps are from the statistics panel, from the average column and with an ambient room temp of 23.
CPU: 57,6
Core1: 79,9
Core2: 77,3
Core3: 76,4
Core4: 73,1
I don't know how you get your temp 41.39, but I assume you add all core temps, divide by four and subtract ambient room temp? In my case that's 53,7c. That's 12.3c more than your reading. Can that simply be explained by different motherboards?
You've got the right idea: Core1-4 are averaged and then subtracted from the ambient temperature. I don't take the maximum Core readings, however, I allow Everest/Aida64 to average them for the final ten minutes of the test run.
Motherboard sensors and BIOS calibration are different between all motherboards; even those of the same model. Another big difference is voltage regulation, where some motherboards are analog while others (such as MSI) are digital and more precise. This makes it difficult to match results between systems. What's important is scaling between heatsinks tested on the same system.
Scaling between heatsinks then, this might be interesting.
Prime95 blending for 1h (1.33125Vcore, 4GHz) causes a max temp of 94c with my Xigmatek HDT-S1283. With the Scythe Mugen 2 it's 88c. The readings are the max temp for the hottest core. The max temp of the coolest core was 84 with my S1283 and 81 with the Scythe Mugen 2.
6c gained on the hottest core. 3c gained on the coolest core.
That's not really a whole lot gained when you take into account that HDT coolers should be really bad with i7, and Scythe Mugen 2 is the runner up in "Best CPU Cooler Performance: Intel Q3-2010".
Am I expecting too much?
Installation and TIM application play a big role, and if you're using too much paste then your temperatures would be higher than mine. You also need to take into account the heatsink orientation (vertical/horizontal), which makes a considerable difference.
In my second post however I wasn't trying to compare against your results but my own. I used Prime blending to generate the heat, and then noticed the max temps with Real Temp.
The Xigmatek S1283 generated these max temps: 94, 92, 90, 84
The Scythe Mugen 2 generated these: 88, 86, 84, 81
Is that what your would expect?
Correct me if I'm wrong, but the fact that my Rampage III Extreme reports temps so much higher than your motherboard, wouldn't that hold my OC back? I mean, you can push your Vcore so much more before reaching TJmax than I, because your temps are reported lower.
Right now with a Vcore at 1.35625 my max core temp is 91, only 9 degrees from max. I'm doubting that I'll be able to reach 1.4v. Now if my reported temps were some 12c lower, that would leave me with some headroom.
But changing motherboard can too.
So my question is: Shouldn't lower reported temperatures result in possible higher OC's? If yes, and you can have lower reported temps by changing motherboard, then it should be possible to reach higher OC's just by changing motherboard.
So manufactures could use the fact that a given product is known for low readings as a sales argument. Seems a little silly.
At any rate, make sure that the heatpipe rods travel as horizontal as possible. You might already have them that way... on the Scythe Mugen 2 the heatpipes should be on the left and right, as should the fans. On HDT coolers it's usually opposite.
Given your findings that different CPU's have different heat signatures
I am not finding any comparative test past the last Q3 2010 and it is June 2011.Did you stop them?
I am looking for a good cooler for the Intel LGA 1055 platform i5 2500K CPU
thnx
While Benchmark Reviews has published several CPU cooler articles, we have not produced another comparitive roundup like the 'Best CPU Cooler Performance' series. This is for two reasons: 1) the latest Intel CPUs (Sandy Bridge) run significantly cooler, and do not require the same high-end heatsink as before, and 2) there haven't been that many new heatsinks introduced.