NVIDIA GT300 Architecture Detailed

For a good while now, people have been speculating on the bits and pieces of NVIDIA's next-generation GPU architecture, the much-anticipated GT300, which the company has been consistently keeping under wraps. Rumored to be scheduled for release sometime by the end of this year, or early next year, the new GT300 GPU continues to make the headlines as more details become available on the Internet. The most recent piece of information, which has been made public on a forum page, includes a diagram of what is speculated as NVIDIA's next-generation graphics processing unit.

 

Unlike previous rumors, it looks like the new diagram also adds a bit of confusion, as it talks about a new GF100 code name, which makes this the third, unofficial code name for NVIDIA's next-generation graphics processing unit. The previous naming schemes include GT300 and Fermi, with the latter having surfaced just recently. It could be that, by the time of the official debut of this new GPU, the Santa Clara, California-based chip maker will further change the naming of its new card.

 

There have been a couple of speculations made on the side of the newly surfaced diagram, but everything is still pretty much a rumor at this time. According to a couple of following posts, on the left bottom corner, the diagram reads “3.xB transistors, 40nm @TSMC,” with the x being speculated as a 2. This tells us, if correct, that the GF100 will turn out as a monster of a card, boasting over three billion transistors and taking advantage of the next-generation 40nm process technology that has been developed by TSMC and already adopted by AMD in its latest line of GPUs.

 

Everything else is also speculation, which means it's best to wait until the graphics maker will officially debut the new cards, thus providing the world with a perspective on what this architecture has to offer.