I know console gamers don't understand this, but Mhz is not an objective measurement of performance for different architecture.

a 2.6ghzamd cpu even with the same number of cores and the same cache size per core as an intel cpu with the same 'specs' will run slower.

ram isn't just 'mbs' either. There are ram timings to consider etc. the core speed if the GPU is almost the most irrelevant spec as well given the shader heavy nature of modern games. fillrates and bandwidth are usually more important but again, architecture is where the difference is made up.

This equaltion is for tflops of the 500 series GPUs

"Each Streaming Multiprocessor(SM) in the GPU of GF110 architecture contains 32 SPs and 4 SFUs. Each Streaming Multiprocessor(SM) in the GPU of GF114/116/118 architecture contains 48 SPs and 8 SFUs. Each SP can fulfil up to two single precision operations FMA per clock. Each SFU can fulfil up to four operations SF per clock. The approximate ratio of operations FMA to operations SF is equal 4:1. The theoretical shader performance in single-precision floating point operations(FMA)[FLOPSsp, GFLOPS] of the graphics card with shader count [n] and shader frequency [f, GHz], is estimated by the following: FLOPSsp ≈ f × n × 2. Alternative formula: FLOPSsp ≈ f × m × (32 SPs × 2(FMA)). [m] - SM count. Total Processing Power: FLOPSsp ≈ f × m × (32 SPs × 2(FMA) + 4 × 4 SFUs) or FLOPSsp ≈ f × n × 2.5."

This equation is different for different cards, even though they use the same core components.

Saying 1 GPU is better than another based on the specs on paper is meaningless. That's what benchmarks are for. To test what these e-peen press releases actually mean in terms of real world performance.

Fanboys always quote PS3 ram speed...but they don't know that even when coupled with the slowest ram money can buy or the fastest, a high end GPU will still only gain/lose less than 1% performance.