Intel: Why GPU is the first to be launched at 7nm node? Because it has lower risk

At Intel's Investor Day held on the 8th local time in the United States, the company's CEO Bob Swan and President Murthy Renduchintala talked about Intel's chip development roadmap. In the roadmap, Intel announced the progress of the 7nm process for the first time, and it is expected to produce and launch a 7nm product with EUV in 2021.

       It is understood that Intel's 7nm products will be a combination of the company's lessons learned from its 14nm and 10nm series products. Intel hopes to achieve 2x expansion (Moore's Law), but plans to make intra-node optimizations as part of the roadmap.

Intel announced that the 7nm product to be launched in 2021 will be a new GPGPU based on the Xe graphics architecture. The company said that its Xe products will use EMIB 2D integrated packaging and Foveros 3D hybrid packaging, targeting data center AI and high-performance computing, with two different microarchitectures (from mobile GPGPU), one of which is called Arctic Sound.

      As a major CPU manufacturer, why did Intel choose to launch the GPU first at the 7nm node? Intel Senior Vice President Venkata Renduchintala said that the reason is that when using the new generation of processes, especially EUV lithography, the GPU is launched first because the risk is lower and the defect problem is easier to solve, even if it sacrifices a little performance.