I've installed the cooler on my AM2 board, utilizing the mounting clips the unit shipped with, which require a good bit of pressure to attach to the stock retention bracket. The factory AMD coolers have a locking lever to apply the force to the clip, which is much easier to use. All I can say is, thankfully we have chips with IHS, since putting this kind of uneven force on an unprotected die would almost certainly crack the edges. Once the clips are in place, the heatsink is practically unmovable. The motherboard bends slightly around the CPU socket, but with little weight from the cooler itself there shouldn't be much concern about breaking something.
The system specs are as follows:
Biostar TF570 AM2 Motherboard
AMD 6400+ X2 90nm Windsor
2x2GB OCZ Fatal1ty PC6400
The CPU socket is oriented perfectly to allow proper placement of the XT-1264 so that it is directly in line with the rear case exhaust fan. This particular case utilizes a BTX format, however most ATX style cases should offer similar layouts for the motherboard in relation to the rear fan. In the event your socket is turned 90° the next best scenario would likely be to exhaust the cooler up, towards the location of either a top-mount power supply or blowhole fan.
This board does have a 4-pin header for the PWM fan, however in prior tests no change could be effected using the controls built into the BIOS. Due to this restriction, all tests done with the included fan were run at maximum speed, which at a rated 38dBA is certainly not the quietest 120mm fan in use. While I wouldn't exactly call it noisy, it was definitely noticeable from outside the case, and more than I care to listen to on a daily basis. With 101CFM however, the Kingwin fan does provide a lot of performance to go with the XT-1264 cooler.
Two CPU states were used to test the XT-1264, both with the stock voltage and an overclocked setting that uses 1.5V, which on this 90nm die processor really cranks out the heat. The Kingwin was able to keeps things under control and in fact, under 50° C even with 100% load on both cores. I also tried swapping out the fan for a much quieter one with less airflow, and although load temperatures went up several degrees, I was still able to maintain these settings without any crashes or instability. OCCT was used to test and verify the settings.
Let's wrap things up with some final thoughts and conclusion.