| Best CPU Cooler Performance: Intel Q3-2010 |  |
| Reviews - Featured Reviews: Cooling | |
| Written by Olin Coles | |
| Wednesday, 17 November 2010 | |
|
Page 9 of 16
Zalman CNPS10X-Performa CoolerZalman, the company behind years of quiet cooling products, has taken their Zalman CNPS10X-Extreme design, and given it several tweaks to help focus purpose on specific needs. The Zalman CNPS10X-Performa is designed for serious overclocking projects, and in many ways could be a better performer than the CNPS10X-Extreme. Designed to universally fit all Intel LGA775/1156/1366 and AMD AM2/AM2+/AM3 sockets, the Zalman CNPS10X-Performa CPU cooler sells for $40 at NewEgg.
The Zalman CNPS10X-Performa heatsink keeps the five 6mm copper heatpipe rods used for other CNPS10X-series coolers, but improves the finsink design with more aggressive central bands of rippled aluminum fins and allows a second 120mm fan to be mounted on the backside. These ripples reduce laminar airflow, and create a more efficient heatsink.
Included with all Zalman CNPS10X-series coolers is a small 3.5-gram pouch of Super Thermal Grease ZM-STG2M, the very same TIM that topped the results of our 80-way Thermal Interface Material Performance Test mega-article with an 'A' rating.
Features and Specification
|
|
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.