Temperature Performance

We took temperature readings as close to the BGA memory ICs as possible, as well as centrally on the heatsink or heatspreader to determine two things:

1. What the overall cooling performance is at a reference standard of 800MHz ~1.8V, and their overclocked and overvolted values. This will see how well the cooling technologies perform for the job tasked to them.

2. What the difference in temperature is between the heat producing parts and heat dissipating parts - regardless of actual temperature this should be as uniform as possible to highlight the actual efficiency of the heatsink or heatspreader.

Bearing in mind that between the four memory sets they all use different IC's so the power output will be slightly different, however at the baseline of ~1.8V and 800MHz the Corsair DHX heatsinks basically match the OCZ Flex II heatsinks - there's but a mere fraction between them.

The Flex II is better than the original FlexXLC though, but by only around a single degree Centigrade, and naturally everything eclipses the basic heatspreaders that are fitted to the Crucial Ballistix.

As for the differences between the memory ICs and heatsink the original FlexXLC actually came out with the smallest difference, and the Corsair DHX and Flex II with a slightly larger, and almost identical difference.

At an overclocked speed and voltage, the OCZ Flex II comes out slightly better than the Corsair DHX by a couple of degrees - in fact, there's very little difference between this and its idle temperature. Will watercooling actually make that much of a difference then? The original FlexXLC is unsurprisingly hotter running at a much higher 2.32V because it's a 2GB kit that uses the older Micron D9 ICs. All of them are still far, far lower than the Crucial Ballistix heatspreaders which top over 50C when applied with even just a moderate overvoltage.