Although Nvidia managed to get Fermi up and running and, by now, provided us with a wide range of solutions based on this architecture, nobody can deny the fact this DirectX 11 GPU was late to market so don't blame me for being somewhat skeptical when looking at this Nvidia slide that states their next generation graphics core, dubbed Kepler, would start shipping in 2011.

Based on the 28nm manufacturing process, this processor should deliver three to four times the performance per watt compared to the Fermi architecture, while also coming in as a more power friendly alternative.

The slide also presents a GPU codenamed Maxwell, that is supposed to arrive in 2013 and deliver up to a sixteen-fold increase in performance per watt compared to Fermi.

This sounds like good news indeed since we all need some competition in the graphics market, although I am not so sure what to make of Nvidia's announcement after the whole wooden screw Fermi fiasco from last year. 

Taking a closer looks at the slide reveal yet another fact that makes me wonder when is Nvidia actually going to release this since Fermi is displayed as being available in 2009 and, as we all know, the first graphics cards based on this architecture were launched a full three months later.

TSMC could also add to the problems encountered by Nvidia with the Kepler chip since we all know how much time it took for this Taiwanese foundry to perfect their 40nm manufacturing process, ATI's Evergreen family of graphics chips being plagued by availability problems for quite some time, and, if the same thing will happen with their 28nm process, Nvidia is going to have a serious setback on their hands.

End of it all, I certainly hope that Nvidia manages to do what has set to accomplish with the Kepler GPU, but until I see the first solutions with my own eyes I remain somewhat skeptical, given NVIDIA's least-than-immaculate past. (via engadget)