Nvidia ampere, leaked information on rtx 3080 / ti, 3070 and more

A series of leaks regarding Nvidia Ampere have emerged online with a lot of information, which we will go on to reveal below.

Nvidia Ampere, The RTX 3080 Ti would be 40% faster than the RTX 2080 Ti

According to KittyCorgi on Twitter, the GeForce lineup would be powered by a total of 5 Ampere models and their respective SKUs with various core configurations. The flagship of the lineup would be the GA102 which would be the successor to the TU102 GPU. The rest of the GPUs and their predecessors are mentioned below:

• GA102 - TU102 - GP102GA103 - No predecessor GA104 - TU104 - GP104GA106 - TU106 - GP106GA107 - TU117 - GP107

It is mentioned that all GPUs will be made with Samsung's 10nm process node (8LPP). SLI through a state-of-the-art NVLINK interconnect will be available but only on top-tier GA102-based graphics cards. Currently, this limitation is only below the TU104-based GPUs, but it looks like NVIDIA will impose more limitations with next-gen Ampere parts.

NVIDIA would also enable RTX (Real-time Ray Tracing) on ​​all GPUs, including low-end GA107 parts, which means there are some major architectural improvements to RT design on Ampere GPUs, allowing for higher performance Ray Tracing in real time on new cards.

Once again, it is mentioned that FP32 units in Ampere would double, but they do not translate into an increase in rasterization performance of 2x or twice the increase in the number of CUDA cores. It is also mentioned that Ampere GPUs will take advantage of the performance of PCIe Gen 4.0.

RTX 3080 Ti (GA102)

The GPU will have 84 SM, which is equivalent to 5376 CUDA cores. This is 16% more CUDA cores than the TU102 GPU used in Titan RTX. The GPU would be capable of supporting up to 12GB of VRAM through a 384-bit bus interface.

Performance would be up to 40% faster than the RTX 2080 Ti.

RTX 3080

It will have a total of 60 SM with 64 CUDA cores each, which means 3840 CUDA cores. The card would have a 320-bit bus interface, which means it could support 10 or 20 GB of GDDR6 memory.

Its performance is speculated to be above the RTX 2080 Ti by 10%.

RTX 3070 (GA104)

The successor to the TU106-based GeForce RTX 2070 is suggested to have the same number of CUDA cores as the RTX 2080, in 3072. The card is expected to have a 256-bit bus interface that, once again, would allow 8GB or up to 16GB of GDDR6 memory.

Speculation is that it is 5% slower than the RTX 2080 Ti.

Lastly, we have the GA106 and GA107 GPUs, which will target the consumer and budget segments. The GA106 GPU would have 1920 CUDA cores with up to 6 GB of memory through a 192-bit bus interface, while the GA107 GPU would have 1280 CUDA cores with up to 4 GB of memory through a 128-bit bus interface.. We will keep you informed.

Wccftech font