hapkido :
Do they not teach science in school anymore?
The point of an experiment is to remove all inconsistencies. For a CPU experiment, that means only the CPU and motherboard should be different. For a motherboard, that means only the motherboard should be different. For a GPU, that means only the GPU should be different. For RAM, that means only RAM should be different. So on and so forth.
Who cares if IB takes 1600Mhz RAM and PD uses 1866Mhz? Not only is RAM not a bottleneck in gaming... or really anything except RAM benchmarks... it has nothing to do with the processor's capability. If you honestly think the choice of RAM makes a fx8350 fall way behind a 3470 and more inline with a 3225, you need to check your scientific method. As long as it's the same on both systems, there is no bias.
No. If you are testing different RAM modules, then you would use same motherboard, same chip, same everything.
If you are testing two chips of the same architecture (e.g. i5 3570K vs i7 3770K), then you would use same motherboard, same memory, same everything.
But if you are testing two chips of
different architectures you must use
different architectural elements such as different motherboards and memory configurations. The Intel ivi is designed for using a 1600MHz memory bus at stock configuration, but the FX is designed to use a 1866MHz memory bus at stock configuration. Caches, internal logic... are designed according to the rest of elements of the
architecture such as stock memory bandwidth, channels...
If you
force the FX chip to use a slower 1600MHz memory then you are artificially favouring the Intel chip which is running at its stock speed.
Memory speed has a important role in any task where the cpu can be bottlenecking. Your claim that memory speed has only importance on "
RAM benchmarks" shows again the little that you know. Maybe you believe that people run RAM benchmarks because are bored but they run those benchmark because RAM speed can affect chip performance.
In fact gaming is also affected by RAM speed. An Intel i7-3770k gets up to a 21% more FPS in skyrim (yes the same game that you cited above) when using faster RAM modules
And AMD chips are more affected by memory bandwidth due to the design of their architecture.
I understand these are not good news for people who wasted money on an expensive i5, but when the benchmarks are made in an unbiased way, the FX-8350 beats any i5 and even beats the i7-3770K,
although current software is not still using all advanced elements found in the AMD architecture design.
As said above, future games will be optimized for AMD architecture thanks to PS4 and other consoles.