The rumor mill has been talking about something, that if true would be a pretty big deal in the CPU industry. According to Bitsandchips.it, Intel has started work on a brand new x86 architecture that will succeed the current “Core” generation, which has been in place since Sandy Bridge. If this turns out to be true it will be a pretty big move and one that has been long overdue. It is apparently supposed to happen to in the 2019-2020 time frame. This new architecture will be more of a lean and mean approach, which we will explain below.
When it comes to CPU architectures Intel really has not shifted from its primary architecture in quite a while. Intel processors have remained 100% backwards compatible with all previous iterations. When we talk about Intel’s PAO cadence it is essentially the same underlying x86 architecture expanded with new features with every iteration. So if the rumors are true, this might actually change. Intel could introduce an x86 architecture that we would consider to be truly different as backward compatibility is no longer assured.
Why would Intel do this? Intel could save precious die space by removing the hardware for legacy SIMDs and other legacy features. It would make for a more leaner and meaner chip, providing a bigger bang for lower resources on-die. The architecture would be used to effectively replace the current “Core” series on Desktop and even the enterprise market.
If these rumors turn out to be true, Tiger Lake would be the last iteration of the current x86 architecture from Intel. Toger Lake is expected to arrive in 2019, so we can expect the new x86 implementation in 2020. Intel has already taken steps towards a more efficient architecture, which we first saw with Skylake. Intel removed the Fully Integrated Voltage Regulator (FIVR) on this architecture and showed a much more mature outlook on on-package economics.
So will Intel’s new lean and mean design effect us? Likely not. There might be a rare few, but if you are someone who uses legacy SIMDs these can easily be emulated if there ever is a call by the OS. So Intel’s move to free up space on the die for smaller and more efficient cores will make for a much more streamlined x86 architecture.