Answer:
P-cores are performance cores and mainly used for traditional CPU work instructions. E-cores are efficiency cores and are mainly used for minor tasks running in the background.
This guide will discuss the concept of P-cores and E-cores on recent Intel CPUs, and how they differ from traditional CPU core architectures.
In a nutshell, P-cores are built for traditional CPU work instructions, while E-cores handle all the other minor tasks in the background. When combined, they provide next-generation multitasking capabilities similar to smartphones, but are much more powerful.
If you are curious about the nature of their design, or if they are really necessary compared to traditional mobile big-little cores, follow the rest of this article.
Table of ContentsShow
Background Development Of P-Cores And E-Cores
Intel’s 12th-gen processors, also known as Alder Lake brought many architectural improvements, as well as officially leaving the 14nm+ process node to the new 10nm Enhanced SuperFin, renamed as Intel 7.
Incorporating two types of cores was part of its key design as a hybrid architecture. This is both a way to push multi-core performance higher while streamlining processes into what Intel refers to as the Thread Director.
The Thread Director is the manager, so to speak, of the combined P-core/E-core configuration of Alder Lake and Raptor Lake CPUs. It uses machine learning to schedule tasks, and assess which type of core should work on which instruction at each precise moment.
In theory, this prevents background tasks from taking up the workflow of the primary P-cores, and can simply be processed by the E-cores without any perceived delays to the system.
What Do P-Cores Do?
P-cores, or Performance cores, are basically the traditional cores of a CPU. They handle all of the major tasks of the system, and are made to work when process-demanding software is executed.
Because they are the main cores, they are built to have higher boost clocks, and are made to handle all the heavier tasks for the computer. Programs such as editing software, graphics renderers, and gaming applications are typically handled by the P-cores.
For Intel 12th-gen (Alder Lake), the P-cores are built using the Golden Cove architecture, and are made for significantly higher IPC (instructions per cycle) than its direct predecessors, Willow Cove (11th-gen mobile) and Cypress Cove (11th-gen desktop).
For Intel 13th-gen (Raptor Lake), the P-cores are built using the Raptor Cove architecture, which is technically a refresh of the Golden Cove architecture which minor tweaks in clock frequency, catch, efficiency, and a new dynamic prefetch algorithm.
What Do E-Cores Do?
E-cores, or Efficiency cores, are secondary cores made to handle everything else that the Thread Director deems less of a priority for P-cores to handle.
Most background processes of a PC’s operating system fall under this, although certain operations such as minor rendering tasks can be given to the E-cores depending on the whole CPU workload.
Because they are not meant for primary tasks, they have lower boost clocks. Thus, they are generally not recommended for gaming applications. They are also built with older architecture, though this is offset by their smaller die size compared to P-cores (four E-cores can fit into the space of a single P-core).
For both Intel 12th Gen and 13th Gen, the E-cores are built using the Gracemont architecture, which is basically a refresh of the Skylake architecture (7th Gen), but is highly miniaturized and is very power efficient.
Do We Really Need P-Cores And E-Cores?
For most modern tasks, not really.
AMD is yet to adopt a similar architectural design as that of Intel and Apple, and yet its Ryzen 5000 and Ryzen 7000 series CPUs can still compete fairly well.
Even some of Intel’s lower-end Alder Lake chips do not have E-cores (Core i5-12400, i3-12100, etc.), and yet they still showcase performance that matches our expectations of a next-generation CPU.
In fact, despite negligible gains, some hardcore gamers completely disable E-cores on higher-end Alder Lake and Raptor Lake CPUs.
That being said, the big-little hybrid architecture has already proven to be a very successful design for more than a decade already in mobile devices.
It is more than a simple case for efficiency, as such architectures have far more granular control in both data processing and power delivery.
And while AMD is yet to develop a hybrid architecture of its own, Intel will push forward to refine the concept for desktops and laptops even further in the foreseeable future.
