Mini-PCs with powerful Intel-AMD chip combination come in a few months

Mini-PCs with powerful Intel-AMD combo chip coming in 2018

Mainboard with Intel AMD combo chip

(Picture: Chiphell)

In the first half of 2018, Intel plans to use the first core processor with AMD GPU and small form factor PCs for PC gamers.

Shortly after Intel announced a combi processor with AMD graphics unit in early 2018, images of a motherboard with that very semi-custom chip are now surfaced. The motherboard is designed for future Intel NUC systems, ie particularly compact complete systems. Such mini-PCs are to appear in the first half of 2018 and be game fit thanks to the 3D performance of the AMD GPU.

Under the code name "Hades Canyon-VR" and "Hades Canyon" diving such Gamer Minis on an unofficial roadmap graphics, which became known in September 2017th At that time it was not clear, however, that Intel intends to place the AMD GPU on the same chip package as the CPU.

Intel Intel CPU, AMD GPU and HBM2 memory on a package(Picture: Intel)

The jointly developed well under the code name Kaby Lake-G from AMD and Intel processor combination with Core i CPU and Radeon GPU will also drive flat gaming notebooks in addition to mini-PCs. The 3D performance of the graphics unit is expected to exceed the current Intel processor graphics clear and sufficient for the requirements of design and rendering programs, according to AMD's vice president Scott Herkelmann.

connected to the carrier board by EMIB

Intel processor and AMD graphics unit sitting on a carrier board (package) and are embedded multi-die interconnect Bridge (EMIB) connected. In contrast to the known about by Vega or Fiji GPUs interposers sit between the Dice only small connector - a total of Intel technology is less expensive than full surface into the substrate incorporated interposer.

In addition to the CPU and GPU, there are still stacked memory chips of the type high bandwidth memory 2 on the package. About a particular framework that was developed by Intel, processor, graphics processor and memory HBM2 information coordinate with each other. So energy saving mechanisms will be implemented around, changed performance states and read temperatures. Such as to meet the different load scenarios playing, calculating or render optimally - according to Intel utilization states between CPU and GPU also should be possible to move automatically.