End of the “myth”: Intel wants to prove that ARM is not superior to x86

It's a cliche that comes up regularly: x86 processors will be threatened by the ARM architecture,”Better and more efficientThe x86 instruction set is used by Intel and AMD probably At the heart of your processor if you're using Windows, and still powering the vast majority of data centers and supercomputers around the world.
But with many players — like Qualcomm in the general public, but also Ampere and Nvidia in data centers — starting to question Intel and AMD's near-monopoly on chips, it's time to consider that assertion.

Apple M1 phone revealed, not Snapdragon

Given the success of the Snapdragon at the Computex show in Taipei – more than 22 designs from the majority of manufacturers – some commentators claim that Lunar Lake is a response to that ARM chip that promises mountains and wonders. These comments seem to ignore the development time of the components. Eric Jehun, Intel engineer responsible for designing the SoC, explains: “Discussions about the design of Lunar Lake occurred long before Qualcomm announced the development of its chip. The real impetus was Apple's M1.“

Launched at the end of 2020, Apple's in-house chip ended up completely knocking Intel out of its PCs. Based on the ARM instruction set, its microarchitecture was developed entirely in-house. Understand that instead of buying the generic basic plans that ARM designed for its customers, Apple acquired a license that allowed it to delve into the heart of the language to organize its chip as it wanted, and in effect create its own microarchitecture, which it then integrated into the SoC.

Instruction set ≠ Microarchitecture ≠ SoC ≠ Burning node

There is a lot of semantic confusion in the terminology surrounding central processing units (CPUs), especially word structures that are often misused. In order to understand this, we need to delve deeper into our central processing units, which are central processing units (CPUs), also called microprocessors in French. Responsible for managing the entire system – unlike, say, a GPU or NPU, which are more computationally intensive beasts or accelerators – a CPU has basic instructions as its basic building block, which are collected in a kind of dictionary, called the Instruction Set Architecture in English (ISA): Instruction set.

These instructions are part of the basic vocabulary of the chip. But it is not because the words in Camus's work or in a political treatise all come from the French dictionary that makes the two texts similar! Just as these two examples are different in substance and form, two chips designed with the same instruction set have different mechanisms by design. Logical design on the one hand – the interactions between different computational units – and physical design on the other hand. The organization of instruction sets, how they are interconnected, and how information flows within them is called microarchitecture. From this microarchitecture arise many characteristics of the CPU: target frequencies (minimum and maximum), number of cores, etc.

Therefore this fine structure does not represent the final component. Once that is determined, you should select the CPU size according to your needs. The CPU today is only a small part of the chips: You only have to look at the chip diagram to realize that the part of the CPU we are talking about no longer occupies only a small part of the processors in our computers. Processors whose real name is now System on a Chip, i.e. System on a Chip (SoC). SoC mapping makes it possible to specify the part(s) of the CPU, GPU, and now the NPU, and the parts dedicated to exchanging data on and off the chip (cache, fabric, etc.).

Finally, a strong argument has been waved in favor of ARM chips lately: engraving precision. When Apple launches its M1 phone at the end of 2020, it will have Intel's 11th generation Core mobile chips in front of it. But while the M1 is etched at 5nm, Intel's chips are etched at 10nm. However, engraving precision is the key to reducing energy consumption. Obviously, although the x86 processor is constantly being singled out for its low power efficiency, this is actually not the case. And it's not just Intel that says this…

When the competitor's mind checks the validity of Intel's words

According to Intel's Eric Jehun, “x86 can be just as efficient as ARM if used well“. Words that could be interpreted as a simple defense on the part of an industrialist who has recently experienced technical burnout. But regardless of the fact that this message was not sent by another marketer, but by one of Intel's chip design luminaries, it has been verified These words were true last year… through competition!

A competition embodied by Gerard Williams, the godfather of the Apple M1 and Snapdragon X Elite CPUs. The chief engineer, whom we met last October at the Snapdragon Summit, actually contradicted us when we mentioned ARM's potential superiority over x86. “Any ISA is good“, Confirms. “Again, it all depends on your starting point, and for x86, it's in watts, not milliwatts.“It is the starting point for designing the microarchitectures that J. Williams identified at the time as the Achilles' heel with respect to ARM, not its DNA.

After being confronted with these comments we collected last year, Arik Gihon agrees: “This is absolutely true. It's not the ISA that's at stake at all, it's the exact architecture, its implementation, and the burning node“.admitting that”Apple has raised the bar in many areas, including energy“Arik Jehun confirms that”Competition is good for Intel.“ The first tests of Meteor Lake, scheduled for the end of the year, will show whether Intel can rise to the challenge and prove that x86 can indeed be as efficient as ARM… to be continued.