AMD and NVIDIA GPUs are still based on the TSMC’s 28nm process, although, digging architecture can significantly improve power consumption and heat, but from Apple A8 processor transistor doubling the size of the core area is also 13 percent smaller examples of view, 20nm in reducing costs, improving transistor density is still worth the wait. Latest news show that, NVIDIA’s first 20nm process products is likely to be the next-generation Tegra processor, codenamed Erista, Maxwell architecture will be integrated as a GPU core, Coming in 2015.
NVIDIA’s first 20nm process products are the next generation of Tegra, Coming in 2015
Fudzilla news exposed that NVIDIA citing informants in developing 20nm process products, the first appearance would be from the mobile processor, which is published by the General Assembly this year GTC Erista new generation Tegra processor architecture, the current Tegra processor architecture codenamed Logan while Erista is named after Wolverine (Logan) son, just father son pass, turn it up this year.
Earlier, Tegra Logan after there is a Spider-Man “Parker” and Iron Man “Stark” code (Stark is good since tony stark is a scientist), but eventually host a son Logan.
For now the Specifications of Next-generation Tegra processor are unclear, but according to NVIDIA announced before the news, Erista the GPU will be upgraded from the current Kepler to Maxwell architecture, more outstanding performance per watt ratio. Judging from the indicators given NVIDIA, Tegra 4 per watt ratio (single-precision floating-point GFLOPS /W) than 20GFLOPS / W, Tegra K1 can reach 40GFLOPS / W (Tegra K1 floating point performance is said to reach 326GFLOPS) while Erista will exceed 60GFLOPS / W, under the same power performance than Tegra K1 upgrade at least 50%.
In addition, TSMC (TSMC), although production of the 20nm process, but the process of producing this generation can basically monopolized by Apple and Qualcomm, and it is more suitable for 20nm mobile processor , process life cycle will be very short (28nm process relative terms) , TSMC production next year will be more advanced 16nm FinFET technology. AMD’s next-generation X86 CPU, NVIDIA’s Pascal and other heavy products will turn FinFET process, leaving not much time 20nm process, AMD, NVIDIA GPU probably true, as rumors of a new generation, as pointed directly across the 20nm process stages, etc. 16nm FinFET process until next year, a comprehensive upgrade..