Table of Contents
Intel is one of the leading manufacturers of processors for desktops, laptops, servers, and other devices. The company has been releasing new generations of processors every year, each with improved performance, features, and efficiency. The latest generation of Intel processors is the 12th Gen Alder Lake series, which launched in late 2021 and introduced a hybrid architecture that combines high-performance cores with high-efficiency cores on a single chip.
However, Intel is not resting on its laurels and is already working on its next-generation processors, codenamed Meteor Lake. These processors are expected to debut in late 2023 or early 2024 and will bring significant changes to Intel’s CPU design and technology. Here are some of the key things you need to know about Intel’s 14th Gen processors.
A New Node: Intel 4
One of the most important aspects of any processor is the node or process technology that it is built on. The node determines how small and dense the transistors are on a chip, which affects its performance, power consumption, heat generation, and cost. Smaller nodes generally allow for faster and more efficient chips that can fit more cores and cache memory.
Intel has been struggling with its node development in recent years, falling behind its main rival AMD and other chipmakers like Samsung and TSMC. Intel’s current Alder Lake processors are based on Intel’s 10nm Enhanced SuperFin (ESF) node, which is roughly equivalent to TSMC’s 7nm node that AMD uses for its Ryzen processors.
However, Intel plans to leapfrog its competitors with its next node, called Intel 4 (formerly known as 7nm). This node will use extreme ultraviolet (EUV) lithography, a cutting-edge technology that allows for more precise patterning of transistors. Intel claims that this node will offer a 20% improvement in performance per watt over ESF , which means that Meteor Lake chips could run faster and cooler than Alder Lake chips at the same power level.
However, there is a catch: Intel will not use this node for all parts of Meteor Lake. Instead, it will use a chiplet design that splits the processor into different dies that are connected by an interconnect fabric called Foveros . The main compute die will be fabricated on Intel 4 and will contain the CPU cores and cache memory. The other dies will include an integrated GPU (iGPU) die made on TSMC’s 5nm node and an input/output (IO) die made on TSMC’s 6nm node. This way,
Intel can leverage different nodes for different functions and optimize each die separately.
A New Architecture: Redwood Cove + Crestmont
Another major change that Meteor Lake will bring is a new CPU architecture that will replace Alder Lake’s hybrid design. Alder Lake uses two types of cores: Golden Cove cores for high-performance tasks like gaming and content creation; Gracemont cores for high-efficiency tasks like web browsing and background processes. These cores can work together dynamically depending on the workload thanks to a technology called Thread Director.
Meteor Lake will also use a hybrid design but with different core architectures: Redwood Cove cores for high-performance; Crestmont cores for high-efficiency . These cores are expected to be based on Golden Cove and Gracemont respectively but with some enhancements such as higher IPC (instructions per cycle), better AI capabilities, and support for DDR5 memory .
The exact configuration of these cores is not yet confirmed but some rumors suggest that Meteor Lake will feature up to 22 CPU cores: six Redwood Cove cores
and 16 Crestmont cores. This would be an increase from Alder Lake’s maximum configuration of 16 CPU cores: eight Golden Cove cores and eight Gracemont
cores[^2^. However, these numbers may vary depending on the SKU (stock keeping unit) or model of Meteor Lake processor.
A New Socket: LGA1851
One thing that PC enthusiasts should be aware of when planning their upgrades is the socket compatibility between processors and motherboards. Yes, I do. The new LGA 1851 socket will house the new Intel Meteor Lake-S desktop CPUs. The socket is the exact dimensions as the LGA-1700 socket, with one slight difference. The IHS of the socket protrudes higher than previous sockets, which could ensure adequate room for Intel’s new multi-tile technology. The new Socket LGA1851 platform is expected to power Intel’s “Meteor Lake-S” and “Arrow Lake-S” microarchitectures.
Do You Know a New Socket LGA1851?
If you are a PC enthusiast or a gamer, you might be familiar with Intel’s desktop processor sockets. These are the physical interfaces that connect the CPU to the motherboard. Intel has been using different sockets for different generations of its processors, and it seems that they are planning to introduce a new one soon.
The new socket is called LGA1851, and it will replace the current LGA1700 socket that debuted with the 12th Gen Core “Alder Lake-S” processors in late 2021. The LGA1851 socket will be compatible with Intel’s next-gen desktop platforms, namely “Meteor Lake-S” and “Arrow Lake-S”, which are expected to launch in 2023 and 2024 respectively.
But what is so special about this new socket? And why does Intel need to change it again? Let’s find out.
What is LGA1851?
LGA stands for Land Grid Array, which means that the CPU has an array of contact pads on its bottom surface, while the socket has an array of pins that match those pads. The CPU is inserted into the socket and secured by a retention mechanism. This design allows for more pins than other types of sockets, such as PGA (Pin Grid Array) or BGA (Ball Grid Array).
The LGA1851 socket has 1851 pins, which is 151 more than the LGA1700 socket. However, the physical size of the socket and the CPU package remains unchanged at 45 x 37.5 mm. The extra pins are achieved by reducing the size of the “courtyard”, which is the central area of the land grid that lacks pins.
The reason for adding more pins is to support Intel’s new multi-tile technology, which will debut with “Meteor Lake-S” processors. These chips will feature chiplets built on various foundry nodes, such as Intel 4 (formerly known as 7 nm) or TSMC N5 (5 nm). The chiplets will include different components of the processor, such as CPU cores, iGPU, memory controller, and uncore. The chiplets will communicate with each other through an interconnect fabric called Foveros.
The multi-tile technology will allow Intel to leverage its IDM 2.0 strategy, which involves using both its own fabs and external foundries to produce its chips. This way, Intel can optimize each chiplet for performance, power efficiency, and cost. It can also mix and match different types of cores and graphics units to create various SKUs for different market segments.
What are Meteor Lake-S and Arrow Lake-S?
“Meteor Lake-S” and “Arrow Lake-S” are codenames for Intel’s upcoming desktop processor microarchitectures based on multi-tile technology. They will succeed “Raptor Lake-S”, which is expected to launch in late 2022 as a refresh of “Alder Lake-S”.
“Meteor Lake-S” will be part of Intel’s 14th Gen Core family (or possibly a new branding), while “Arrow Lake-S” will be part of its 15th Gen Core family. Both microarchitectures will feature hybrid designs with performance cores (P-cores) and efficiency cores (E-cores), similar to “Alder Lake-S” and “Raptor Lake-S”. However, they will use new core architectures that offer higher IPC (instructions per cycle) improvements.
“Meteor Lake-S” will have up to six P-cores based on “Redwood Cove” architecture and up to sixteen E-cores based on “Crestmont” architecture. The P-cores will be built on Intel 4 node (comparable to TSMC N5), while
the E-cores will be built on TSMC N5 node. The iGPU will use Xe-LPG graphics architecture with four Xe Cores (64 EUs or execution units), which could support DirectX 12 Ultimate features. The iGPU will be built on TSMC N6 node (6 nm).
“Arrow Lake-S” will have up to eight P-cores based on an unnamed architecture that succeeds “Redwood Cove”, while retaining sixteen E-cores based on “Crestmont”. The P-cores will be built on Intel’s next-gen node called 20A (which could rival TSMC’s sub-2 nm nodes).