NVIDIA launched Fermi in March 2010, exactly 2 years ago. It was a big architectural revamp, and NVIDIA themselves said that it was the next major step up in its GPU line following the GeForce 8 series. Now, 2 years later, NVIDIA has released another important architectural update to its GPU line, code-named Kepler.
NVIDIA has focused on efficiency, reducing power consumption and die size, and maintained most of the features and details that we know about Fermi. They stripped Fermi down to the core and built it back up keeping the above restrictions in mind. The end result is NVIDA’s new GPU architecture, Kepler.
NVIDIA has launched Kepler with the current flagship GPU, which is the GTX680. As AnandTech shows, NVIDIA has not only retaken the performance crown with the GK104 based GTX680, but they have done it with superb efficiency.
One thing missing from NVIDIA’s Kepler launch is “Big Kepler”, a large sized 500mm2+ GPU to serve as the top flagship GPU for the consumers, and as the staple GPU for its professional Quadro and Tesla lines. At the size of 294mm2+ the GTX680 is not “Big Kepler”, but we know that the big sized GPU is definitely being worked upon at NVIDIA.
Head on to AnandTech for the extremely detailed review, it’s worth the read. However, to sum it up:
AMD spoiled NVIDIA’s Kepler launch by launching its HD7xxx series 3 months earlier, but at the end of the day, it’s NVIDIA who takes home the performance crown for the highest performing single card GPU.
To quote AnandTech, “The GTX 680 is faster, cooler, and quieter than the Radeon HD 7970. NVIDIA has landed the technical trifecta, and to top it off they’ve priced it comfortably below the competition.”
