Intel, one of the biggest players in the processor world, has received a patent that opens new doors in architectural design. This technology, called “Software Defined Super Cores,” could radically change processor performance in the future. With this new approach, multiple cores will be able to act as a single, large, and powerful core.
Intel is ushering in the Super Core era
The basis of this new system is that, instead of single cores with high frequencies and high power consumption, multiple cores collaborate to perform the same task. This results in more efficient performance gains, especially in single-threaded applications.
Workload is distributed among the cores through dedicated instructions and shared memory. This saves energy and improves overall performance. Cores can dynamically switch between “normal mode” and “super core mode” depending on the operating scenario.
This new method brings to mind AMD’s “Clustered Multi-Threading” technology, which was once used in the Bulldozer architecture. However, Intel’s approach differs by linking cores together through software and lightweight hardware support, rather than separating them into modules.
There’s also speculation that the patent may be linked to Intel’s “Royal Core” project, which it previously canceled due to its high cost. Unlike the Royal Core project, which aimed to significantly increase single-core performance at the time, the new patent appears to offer a more flexible solution and aims to strike a better balance between performance and efficiency.
So, when will this technology reach users? According to emerging information, Intel is expected to use this new feature in processors it will introduce in the second half of the decade following Arrow Lake. This move will allow the company to scale single-core performance through a different assembly method, rather than simply producing larger cores.
{{user}} {{datetime}}
{{text}}