In our cpu block reviews we often talk about the “rotation” of the block. In the old 2012 roundup we defined the blocks as not rotated or rotated 90 degrees, without really showing evidence of what that meant. The definition was based on how the block was labelled with respect to a normal motherboard orientation. This is hard to explain in words but easier to visualize. However this definition was further complicated when Moonman started testing and defined a 90 degree rotation as of the machining of the copper base relative to the die orientation. While this makes sense on one level it can leave the two critical points of information beyond the immediate knowledge of the end user. The end user without googling or delidding will not know what the shape or orientation of the die underneath the IHS is. In addition some CPU blocks don’t show you the orientation of the channel machining without taking the block apart.
Therefore we have decided to rename our previous definitions and show pictures explicitly of what we mean!
This was previously known as “not-rotated”. The manufacturer’s logo is horizontal when the cpu block is mounted on a motherboard in a normal “tower ATX” case. For the Koolance CPU-380 it will therefore look like this:
For the EK Supremacy EVO it would look like this:
This we thing of as the “normal expected” orientation of the block by the end user and therefore we call this a “regular” mount.
For this orientation we steal a surf/skate/snowboarding term which refers to the more unusual right foot forward style. In other words, it’s not what you would expect. However instead of being rotated 180 degrees we rotate the block 90 degrees. Due to symmetry a rotation of 180 degrees should make no difference. For the Koolance CPU-380 therefore it would look like this:
And for the EK Supremacy EVO it would look like this:
How much does orientation matter?
It varies on the block – some blocks are less sensitive to it than others due to bow or design, and some CPUs are more sensitive to it than others. Typically 0-2C temperature differences due to an orientation change might be seen.
Note that as mentioned the “best” orientation depends on the orientation of potentially hidden channels in the block, and the orientation of the die underneath the IHS. Therefore while on the 3930K we saw that the CPU-380 preferred to be “goofy” we see that on the 4770K it prefers to be “regular”. Manufacturers are also not consistent. For example a “regular” EK Supremacy and a “regular” Bitspower Summit do not have the channels running in the same orientation by this definition and therefore will always prefer different orientations.
This is why we now do every cpu block test with both orientations so that a manufacturer doesn’t throw us for a loop by being sneaky!