Intel has organized an Intel Architecture Day 2018 for a limited audience, where the company has outlined its future plans. It shared information about a new processor architecture and developments on gpu-level.

In view of the problems surrounding the reduction of transistors, Intel seems to be bringing forward new spearheads. Instead of focusing on clock speed or core numbers, the company wants to make its processors faster in the coming years by improving the architecture. In addition, issues such as software, security, interconnects and memory in the coming years should improve the performance of Intel’s portfolio.

During the Architecture Day, with an extensive report from the neighbors , Intel seemed to admit that the 10nm Cannon Lake generation is a flop. The company would no longer intend to bring the Core generation to market, so that the i3-8121U that has appeared sporadically this year is left alone. The 10nm node, however, would not be skipped, but instead a new generation of 10nm chips would appear, with cores given the codename Sunny Cove. The processors would continue to be called Ice Lake, a name we saw on roadmaps before.

Intel Architecture Day: Sunny Cove

The Sunny Cove architecture is being improved in a number of areas compared to the current Skylake architecture used in all processors to Coffee Lake Refresh. This way Sunny Cove gets a ‘deeper’ architecture, so that more instructions can be executed simultaneously. This is possible through, among other things, larger caches, better branch prediction and important: extra execution units . Where there are eight per core at Skylake, that will be ten per core for Sunny Cove. Intel has also developed new instructions, including for encryption and machine learning.

Intel Architecture Day: Gen11-gpu

The tenth generation gpu, which was shut down in the i3-8121U, would be a dead end, and Intel has indicated to develop the eleventh generation for integrated gpu’s. With a larger L3 cache and many more execution units , from 24 to 64 units, it would enable performance of up to 1.1 Tflops. That would come close to the Vega 8 GPU in AMD’s 2200G apu. In addition, the new GPU receives support for rendering with variable resolution, a technique that Nvidia in RTX cards calls variable rate shading and is called coarse pixel shading with Intel. The Gen11 GPU also gets a new h265 encoder, and hdr tone mapping and adaptive sync for displays.

Finally, Intel has developed a new way to compose chips. Instead of producing one monolithic piece of silicon, it wants to use the Foveros technology. This means that different chiplets, just like AMD that announced for its Epyc chips, can be connected to each other via an interposer. Intel demonstrated that Foveros technology with a 12 by 12 millimeter large SOC that combines two Core processors with two Atom cores, a dram chip and the necessary logic for interconnects and memory and other controllers. It is therefore a complete sip with two powerful and two efficient cores. With this technology, Intel has to compete with Qualcomm, among other things, to make more efficient laptops.

Intel Architecture Day: Foveros