After spending almost two years in promoting the efficient and powerful Maxwell architecture, NVIDIA is expected to unveil its next generation graphics card to counter back on rival AMD, the new GTX 1080 and 1070 GPU will run on the new Pascal architecture and is set to be announced by NVIDIA CEO Jen-Hsun-Huang at the upcoming GPU Technology Conference next month, while the launch will be made at Computex that is set to happen at June this year. The Pascal architecture is code named GP100 and the chip will be based on TSMC’s 16nm FinFET manufacturing process, down from the 28nm process from the current Maxwell architecutre.
The Pascal architecture will introduce improved power efficiency while being able to pump out two times performance per watt, it will also have native support for DirectX 12 feature level 12_1 or higher and features 4-Hi HBM2 stacks for a total 16GB of VRAM and 8-Hi stacks for up to 32GB for the professional compute SKUs. With a maximum TDP of 250W, the flagship ‘GTX 1080’ is expected to boast 8GB of GDDR5/GDDR5X memory and will merely require a single 8-pin connector, other details such as the core and memory clock are still unknown, but we can definitely expect a slight increase this time.
Will NVIDIA be able to put up a good fight with its new GPUs over its rival? By the way, could NVIDIA be also using the name ‘GTX 1080’ as well? We will find out really soon as April approaches.
SOURCE: WCCFTECH
