You must not buy Intel CPUs older than 12th Gen — Let's talk E-Core & P-Core
Intel’s latest Alder Lake processors have two types of CPU cores: E-Cores and P-Cores. The CPU cores in your computer have evolved at a consistent rate over the years. We started with single-core CPUs, but that quickly evolved into multithreading. From there we have shifted to multi-core configurations, beginning with dual-core designs and progressing to quad-core, octa-core, and beyond.
Going multi-core with Intel® Core™ 2 Duo was huge. Now, the next jump is multiples of different cores, also known as the big.LITTLE architecture. Alder Lake, or 12th Gen Intel® Core™, features two different types of CPU cores- one for outright performance and another for efficiency. But what exactly are the Intel E-Core and the P-Core? And, maybe more significantly, why should you care?
Performance Cores, or P-cores are all about raw speed, meant for reducing latency and pushing the limits of single threaded performance. They are the most powerful cores on Intel’s two distinct core layouts. These are the ones that will consume the most energy, run at the highest clock speeds, and crush through instructions and tasks in general. These are the “main” cores in the chip that do most of the heavy lifting. Demanding tasks, such as gaming or heavy processing loads, will often be handled by P-Cores, as would other workloads that benefit from single-core performance in general.
P-core’s micro architecture, code-named Golden Cove, is Intel’s most powerful x86 design yet. Incredibly, Golden Cove improves upon performance over last generation’s Cypress Cove (the micro architecture in Rocket Lake) by an average of 19% at the same frequency. Intel has achieved it by making Golden Cove architectures wider, deeper and smarter than previous generation.
They have smartly added more decoders, execution ports, and physical registers. New Gen processors have also got bigger buffer for reordering instructions and a larger scheduling window. To add more to it, Golden Cove has smarter predictive and reactive features, including caches armed with all new prefetch engines that can observe program behaviour and actually predict future memory access patterns, massively cutting down on latency. All of that together means speed. And lots of it.
12th Gen would be impressive enough on the strength of this new P-core architecture alone, but it also adds something extraordinary with efficient cores.
E-cores, code-named Gracemont, give Alder Lake its flexibility. Essentially, they do the most they possibly can, while taking up the least amount of power and space on the die. Four Alder Lake E-cores occupy close to the same space as a single P-core, giving Alder Lake a huge amount of scalability across all form factors. Where P-cores are all about powering up performance, E-cores are built for efficient multitasking.
E-cores do this with a deep front end, wide back end and intelligent branch predictors that reduce latency, keeping unnecessary power use to a minimum. Clustered in groups of four, E-cores share L2 caches of either two or four megabytes, depending on the configuration. I want to make this clear. These E-cores pack a considerable punch. Compared against Skylake (10th Gen), the Alder Lake E-core offers the same general purpose energy performance per clock at a fraction of the power.
Now, for those not familiar, the Skylake micro architecture powered Intel desktop CPUs up until 10th Gen. Let’s put some numbers behind us. An Alder Lake E-core can deliver equivalent performance at 40% less power than Skylake. Or conversely, one E-core delivers 40% more performance than Skylake at the same power. The Alder Lake E-cores don’t feature Hyper-Threading like the P-cores do, but that doesn’t hurt the equation. Four E-cores running four threads provides 80% more performance than a Skylake set up with dual cores and four threads via hyper-threading. And if power efficiency is a priority, then E-cores can also match Skylake’s throughput while consuming 80% less power.
How well do P-Cores and E-Cores complement each other? In a nutshell, pretty well. There is an additional intelligent design in making them work best together, and that’s Intel® Thread Director.
This is a dynamic system that ensures the right threads are assigned to the right cores at the right time. Conventional approaches rely on static rules that don’t adapt to changing conditions. But Thread Director is a dynamic system that uses real-time feedback to help the OS Scheduler make better decisions about how to allocate work.
When you launch a performance-intensive application, like a game or video editing software, the associated threads immediately go to the P-cores. As background tasks, like email and cloud syncing begin, they’re scheduled on the E-cores to maintain maximum performance for your primary tasks.
But, what happens when something more important arises, like a thread that requires AI instructions? If all the P-cores are occupied, Thread Director provides a hint to the OS that a higher priority thread is ready and suggests which thread moved to the E-cores to make room. We can also make these transitions when a thread on the P-cores goes into a spinning state. When that happens, Thread Director notifies the OS to shift a thread to the E-cores, allowing a more performant one to take its place.
So, are Hybrid CPU Layouts the Way of the Future? While the concept of P-Cores and E-Cores is not new in the tech world, it is new to the x86 architecture, and Intel is finding incredible results with it. Core counts on its chips have increased, as has performance. They’re one of the most important developments in PCs in years, even in their initial iteration, and we can’t wait to see how they improve in the future.
