Why does the Apple M1 chip perform better than Intel?
Because the M1 chip is a “System On a Chip” (SoC) and is designed to be task-specific and engineered to work with the specific operating system, unlike the Intel chips, which are “off the shelf” parts. The M1 is the world’s 1st chip to use a 5nm process with 16B transistors. Each transistor is half the size of Intel’s, thus 2x frequency at half the power. That is the usual benefits of Moore’s Law, which Intel has ironically fallen off of.
The CPU, the GPU & the RAM are all in one chip. This makes you don’t need external electrical connections running all over the printed circuit. That makes the electronics slower with capacitance resistance & inductance in the printed circuit board tracks. In short, the CPU (central processing unit), GPU (graphics processor), DRAM (memory), and a few other processors share the power and architecture of the same chip. The result is somewhat counter-intuitive – most people would assume separate components would be better-performing. The opposite is true with this chip.
Intel processors are designed to be off-the-shelf products that can be used with a wide range of systems. They have minimal optimization based on the hardware they are working with and have a significant communication time since they are integrated through yet another generic hardware (motherboard).
Its integrated design allows the various processors to communicate with each other much faster. There’s less physical space that signals have to travel, and its integrated design makes memory available to both processors. In layman’s terms, all the necessary resources are within reach. You can reach for a cell phone in your pocket much faster than having to run to a separate room to answer the telephone mounted on a wall. It requires less energy for the various system components to communicate with each other.
With the efficiency comes more speed, less power consumption, and less heat produced. However, there are some drawbacks as well. For example, you can’t upgrade your RAM, and certain apps you could run on an Intel-based Mac (including Windows x64) will not run.
In short, it’s like a mega iPad or iPhone that is designed to run macOS. Since it was designed specifically for the task, it can do it better than most Intel chips. The whole computer is actually on a singular chip. This chip is optimized to work on single hardware to not be used with other computers as Intel can. This gives Apple the chance to customize the hardware based on the specific device’s needs, which provides much better optimization. They make sure that the operating system, which Apple also developed, is optimized to be used with M1.
Apple has been working on ARM since before 1993, far longer than anyone other than ARM themselves. Apple helped Acorn Computers co-found ARM. Apple supplied millions in investment and owned 20% of ARM shares. Apple was the first company to actually use an ARM processor in a consumer product, in the 1993 Apple Newton MessagePad.
Apple has been developing ARM as a next-generation processor, but Intel has not changed their old fashioned thinking. So their processors (and AMD) are still entirely dependent on aged designs, which are huge, run extremely hot, and use massive amounts of power.
Apple expended billions of dollars to assemble a world-class chip design team and put decades into creating a processor (family) that uses the latest 21st Century techniques and thinking. The task is to deliver a faster, smaller, and less heat produced than all of Intel’s and AMD’s ancient offerings.
Apple has developed a better processor because Apple cares enough to invent into creating the kind of processor that Apple customers need and want. This isn’t really new as Apple has been focusing on optimizing their hardware specifically for their operating system (and the other way around) to provide a better customer experience. Since M1 is actually built on ARM architecture, there are some latency advantages and a much more limited range of use. But it isn’t really an issue as Apple isn’t focused on supporting various devices but rather only their own.