| Best CPU Cooler Performance: Intel Q3-2010 |  |
| Reviews - Featured Reviews: Cooling | |
| Written by Olin Coles | |
| Wednesday, 17 November 2010 | |
|
Page 4 of 16
Prolimatech Megahalems HeatsinkWhat happens when the best engineers from Thermalright decide that it's time to leave a stagnant company who refuses their ideas? They gather their intellect and begin Prolimatech. Unlike other companies that have formed from the intellect of another, Prolimatech didn't decide to leave their former home and release a mainstream product. The Prolimatech Megahalems offers everything former projects (like the Thermalright Ultra-120 eXtreme) featured, but then adds new engineering where there were improvements needed. Very recently Benchmark Reviews tested the ProlimaTech Super Mega heatsink directly against the original Megahalems model. In that review, which used the results from two independent tests, the Super Mega trailed behind the Megahalems and reinforced our decision to continue using the original version for future projects.
The Prolimatech Megahalems uses a very dense array of aluminum fins which are split down the middle to form two separate heatsink halves. The Megahalems is comprised of six heat-pipe rods, which span to each side and offer twelve total cooling ends. The nickel-plated copper base secures firmly to the CPU with a proprietary mounting clip system. This mounting system, which I am pleased to report is the most effective clip mechanism I've ever used, offers tremendous contact pressure while completely removing the chance of processor movement or sliding. AMD users must pay an extra $10 for the AM2/AM2+/AM3 mounting kit.
Aside from a very impressive alloy mounting system, each outer fin plate has been shaped to accept a wire fan frame so that up to two 120mm or 140mm fans may be attached. While the fins are no wider than a 120mm fan, the attachments will secure larger fans to the heatsink. Benchmark Reviews has tested the Prolimatech Megahalems with a single Yate Loon high-output fan and two silent Noctua Fans, and it always performs at the top of our results. The ProlimaTech Megahalems revision B is available for $62 without cooling fan at FrozenCPU and NewEgg.
Megahalems Specifications
|
|
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.