Nvidia will be moving on from the GTX moniker altogether as the next flagship graphics card from them would be the GeForce X80 series.
If rumours are to be believed, the GeForce X80 will be built based on the “Pascal” architecture. Based on the performance-segment “GP104” and high-end “GP100” chips, the GeForce X80 series is expected to consist of the performance-segment GeForce X80, the high-end GeForce X80 Ti, and the enthusiast GeForce X80 TITAN.
The GP104 silicon is expected to feature roughly 4,096 CUDA cores, 256 TMUs, 128 ROPs, and a GDDR5X memory interface with 384 GB/s memory bandwidth. Texture and pixel fillrates on the GP104 is rated to be 33% higher than the GM200-based GTX TITAN X. The GP104 chip will be built on the 16nm FinFET process and will have a TDP of 175W.
On the other hand, the GP100 will be built on the same 16nm FinFET process as the GP104 and will have a TDP of 225W. That said, the chip will feature memory controllers that are rumoured to support both GDDR5X and HBM2 memory interfaces. The X80 Ti is expected to get 5,120 CUDA cores, 320 TMUs, 160ROPs, and a 512-bit wide GDDR5X memory interface that holds 8GB of memory with a bandwidth of 512GB/s.
The X80 TITAN will feature 6,144 CUDA cores, 384 TMUs, 192ROPs, and a 4096-bit wide HBM2 memory interface that will hold 16GB of memory with a memory bandwidth of 1TB/s. Both the X80 Ti and the X80 TITAN will double the pixel and texture fillrates from the GTX 980 Ti and the GTX TITAN X respectively.
Source: TechPowerUp
