NVIDIA Next-Gen Ampere-Based GeForce GPU Made By Samsung On 10nm Node, Indicates Massive Leak

NVIDIA’s next-generation GeForce lineup, although yet to be officially announced and launched, has started appearing in persistent leaks. The latest reports about the upcoming premium NVIDIA GeForce Graphics Cards have offered multiple details, including the process node, architecture, specifications, RAM configurations, and performance gains.

The NVIDIA’s next-gen GPUs will be based on Ampere microarchitecture, which is succeeding the Turing microarchitecture. This has been corroborated on multiple occasions. However, it appears that NVIDIA’s upcoming GeForce Graphics Cards will have GPUs manufactured by Samsung. These next-gen chips will be manufactured on the 10nm 8LPP Fabrication Node, unlike TSMC’s 7nm EUV Production Process, claim the new rumors.

NVIDIA’s Entire GPU Lineup Will Be Capable Of Supporting Ray-Tracing For Hyper-Realistic Gaming Experience:

The latest reports strongly indicate Samsung’s new 10nm process node will be utilized for the fabrication of NVIDIA’s next-gen GeForce GPUs. It was previously believed that NVIDIA was relying solely on Taiwan’s TSMC and its 7nm EUV process, but now it seems NVIDIA has approached Samsung and will rely on its 8LPP technology instead.

The previous reports aren’t inaccurate. It seems NVIDIA’s GPUs could indeed be manufactured on the 7nm Node at TSMC. However, it could be possible that the HPC oriented GA100 GPU is based on TSMC’s 7nm process while the GeForce GPUs are to be fabricated at Samsung’s (10nm / 8LPP) fab. Incidentally, the same Samsung’s process is also being used to produce the Orion SOC, and the System on Chip reportedly packs NVIDIA’s Ampere-based GPU.

NVIDIA GeForce Ampere GPU Lineup For Gamers Includes GA102, GA013, GA104, GA016 & GA107:

According to the latest leak about NVIDIA’s GeForce Graphics Cards for serious gamers, the lineup would be powered by a total of 5 Ampere GPUs and their respective SKUs featuring various core configurations. The highest-end GPU would be the GA102, which would obviously succeed the current-gen Turing-based TU102 GPU. Barring the NVIDIA GA103, which is an entirely new SKU, the rest of the Ampere-based GPUs have the same naming scheme inspired by their Turing-based predecessors.

  • NVIDIA GA102 – TU102 – GP102
  • NVIDIA GA103 – No Predecessor
  • NVIDIA GA104 – TU104 – GP104
  • NVIDIA GA106 – TU106 – GP106
  • NVIDIA GA107 – TU117 – GP107

Currently, only the highest-end NVIDIA GeForce SKU, the GA102 based graphics cards, will have SLI through a next-generation NVLINK interconnect. However, all the upcoming Turing-based GPUs will be made on the Samsung 10nm (8LPP) process node, and reportedly support Ray Tracing, down to the lower-end GA107 parts. This would significantly boost the appeal of even entry-level NVIDIA GeForce Cards which compete against AMD.

The rumors indicate that the new Ampere GPU would not only improve the rasterization and shading performance but also deliver a substantial boost over Turing in Ray Tracing performance. Turing was the first generation to even support Ray Tracing. Hence it is quite possible that Ampere will be significantly optimized. This would be achieved with a higher number of RT and Tensor cores.

The upcoming NVIDIA GeForce GPUs can also leverage a lot of performance benefits from PCIe Gen 4.0. The standard, which AMD has adopted, and Intel is struggling with, is finetuned for next-gen gaming platforms. The significantly higher bandwidth from the Gen 4.0 protocol, plus NVLINK, would be an attractive feature.

NVIDIA Ampere GPUs Performance Gains To Be 50 Percent Higher Than Turing-based Graphics Cards?

According to the Taipei Times, NVIDIA’s Ampere GPU’s overall performance gain is expected to be 50 percent higher than existing Turing GPUs. The boost will be maintained even while offering twice the power efficiency. The report clearly mentions Ampere-based GPUs would have half the power consumption compared to existing Turing GPUs.

[Image Credit: WCCFTech]
If the reports are accurate, the next-gen NVIDIA GeForce RTX Graphics Cards, even the aggressively priced SKUs, could easily support 4K 60 FPS Ray-Traced gaming from the updated architecture. The next-gen Ampere-based NVIDIA GPUs in the top-end include GeForce RTX 3080 Ti, RTX 3080, RTX 3070. The flagship GeForce GPU is rumored to be the GA102 and will be the successor to the Turing based TU102 GPU. In other words, the NVIDIA Titan RTX and RTX 2080 Ti graphics card should be succeeded by Titan and GeForce RTX 3080 Ti Graphics Card.

NVIDIA Titan And GeForce RTX 3080 Ti Graphics Card Specifications, Features:

The Ampere-based RTX 3080 Ti Graphics Card will feature 84 SMs which equals 5376 CUDA cores. This is a 16 percent jump compared to the current generation TU102 GPU featured on the Titan RTX 2080 Ti. The GPU would be able to support up to 12 GB of VRAM across a 384-bit bus interface.

Even if NVIDIA maintains its current pattern, the upcoming GeForce RTX 3080 Ti could have a full fat 5376 core config on the Titan, while the RTX 3080 Ti may feature 5120 CUDA cores. The GA102 GPU is reportedly up to 40 percent faster than the RTX 2080 Ti.

Alap Naik Desai
A B.Tech Plastics (UDCT) and a Windows enthusiast. Optimizing the OS, exploring software, searching and deploying solutions to strange and weird issues is Alap's main interest.