|Best CPU Cooler Performance Q2-2010|
|Reviews - Featured Reviews: Cooling|
|Written by Olin Coles - Testing by David Ramsey|
|Tuesday, 27 July 2010|
Page 13 of 14
CPU Cooler Final Thoughts
There is one minor drawback to using the Core i7 or Phenom II processors which affects overclockers: the difference in CPU cooler mounting dimensions. Many overclockers and enthusiasts have grown to cherish their favorite cooler, and trust them to cool the hottest system they can build. The problem is that now many manufacturers are offering free adapter kits, or include an adapter with their current model coolers, which leads to bigger problems because of processor differences. For all of our LGA1366 test products, we used the Xigmatek ACK-I7361 or ACK-I7363 CrossBow bolt-through mounting kits whenever possible.
Heatsinks made for the old LGA775 platform are designed for use with a Core 2 (Duo or Quad) or Pentium 4 and D processor with an integrated heat-spreader measuring 28.5 x 28.5mm (812.25mm total area), but the LGA1366 socket requires a much larger 32 x 35mm (1120mm total area) footprint to accommodate the extra 591 'pins'. Then there's the LGA1156 socket, which measure 30mm square for 900mm of area. If you use an LGA775 or LGA1156 cooler on a LGA1366 socket, your missing out on up to 38% (307.75mm) of the contact surface. Additionally, the cores are located in slightly difference locations; the Core 2 Quad is slightly spaced away from the center, while the Core i7 is concentrated there.
The Phenom II processor series from AMD offer a very large 37.31 x 37.31mm (1392.04mm total area) integrated heat-spreader surface, which is the largest processor surface I can recall since the original Intel Pentium (I) days. Compared to Intel's Core 2 Duo and Quad processors which measure 28.5 x 28.5mm, the Phenom II offers over 71% more contact surface area. If you compare the latest Intel Core i7 processors which measure 32 x 35mm, then the Phenom II series offers 24% more contact surface area. For overclockers, this will mean a much larger area to cool, but also much more manageable temperatures.
There are a lot of different products out there, and believe it or not we exclude a few from each article because they don't stack up well at all. So this is why you may not see some of the coolers other sites have tested in our results. Because of space and time limitations it's just simply not feasible to review them all, but it's certainly worth mentioning which products should be avoided. So I began to carefully think about it and nearly constructed a real-time chart which places products into different levels of performance. That's when I realized that performance is relative, too, and what performs well today might be considered low-end only a year from now. Perhaps the best method for testing is to use a synthetic system to generate the same exact load for each and every test conducted. This would stand the test of time much better than any computer system or processor platform would, because temperature is a static measurement, but it wouldn't take into account the differences seen between processor model architecture.
The synthetic test unit might generate 250W of thermal energy, but every CPU series has a different layout and might not mate perfectly to a particular cooler. This brings me to my final point: there's a cooler for every processor and purpose. The ordinary casual computer user is fine with the included thermal cooling solution that comes with the retail processor kit. Systems built with a Core 2 Duo processor and three-piped HDT cooler (like the HDT-S1283 or Vendetta 2) will not be cooled the same as a Core 2 Quad processor because of where the cores align with the heat-pipes. Likewise, coolers built around the Core 2 LGA775 design may not perform well at all with the Core i7 or Phenom II platforms. This is why the research is so critical, and understanding the product is important.