It’s actually a topic I have been putting some thought in. As of late nvidia has been putting a ton of energy with packing clocks with pipes. Pretty strait forward right? While in the camp on the side of the pond, we have ATI that has been writing each and every part of the their software engine with precision in mind, not just raw pipes and clock speeds in mind. Yes ATI has notoriously ran higher clocked cores. But for everything software aspect, ATI has a custom coded and written program to handle the instructions per clock for that individual item. Where as Nvidia acts like the proverbial melting pot, and just takes it as it goes, sometimes ending shoving square blocks in round holes, thus having performance hiccups and sometimes out and out performance "deaths" as of late in their newest cards. Why are we hearing of more and more 7xxx series cards eating it? Simply because, the dump truck method to handling instructions is becoming obsolete VERY fast, with more and more shaders per clock, much higher resolutions and pixel output, the strain on precision rises with each new app that is released. Sadly for the guys that wish to stay in the Nvidia camp are losing ground once again, the lofty dominance that 6xxx series has given us is about to fall on its face again. ATI is doing what they did 3 years ago with the 9700 and 9800 cards. They cut back on pipes and pixel clocks, in favor of compatible and precise software instructions. With DirectX10 at the doorstep, who knows what new problems and issues will arise with the current formula for fast graphic processing? Only time will tell, but the ultimate question, will Nvidia realize too late? Will this turn into another FX-series face plant? Or will the Nvidia behemoth ride on, crushing everything in its path? Only time will tell.
