You might know about Moore’s Law, stating that the number of transistors on a circuit board would double every year. Given its longstanding accuracy, the theory has generally remained an acceptable way to measure the rate of technological improvement.
A similar law that you may not have heard before is Dennard scaling. It essentially states that as transistors get smaller, their power usage should drop or at least remain constant. Keeping this in mind, let’s take a look at how Apple’s latest A13 Bionic chip, found in this year’s iPhone 11 models, compares to last year’s A12.
According to Apple, the A13 is 20% faster while using 40% less power, putting it leagues ahead of the industry-leading Snapdragon 855 and Kirin 980 – both of which had no equal prior to the A13’s announcement. In terms of transistor count, the A12 houses 6.9 billion of them, while the A13 has a whopping 8.5 billion.
Impressive, but far from a twofold increase. Even so, will there even be a noticeable performance improvement in day-to-day use? Chances are that last year’s iPhone, or even the one before that, can do all the same stuff with little compromise. So, was Moore wrong? Is the smartphone reaching its speed limit?
Transistors are essentially tiny switches that act as the brains of your smartphone. The smaller you can make them, the more you can fit on a CPU. Through electronics design software such as Upverter, manufacturers can discover new ways to keep improving the architecture of the processor on your smartphone.
But more transistors doesn’t always mean more power. For instance, Apple’s A8 processor has twice as many transistors as the A7, but the performance difference between the two is relatively small. So, what exactly makes smartphones faster?
Understanding Processing Power
The processing power of a CPU is generally defined by its clock speed, measured in GHz. One cycle equals one hertz, which means a 3Ghz CPU can perform three billion cycles per second. So, more GHz means more processing power, right? Unfortunately, it’s a bit more complicated than that.
This is because processors use various tricks to increase the number of cycles they can perform per second. For instance, they can “fetch” new instructions before completing current instructions, thus carrying them out more efficiently. Processors can contain more cores, which helps them avoid bottlenecks and produce less heat.
There are also additional components that influence the overall performance of a smartphone, such as its GPU and RAM. This means that numerous components work in tandem to produce the final result. The CPU is just a small part of what influences a smartphone’s processing power.
What This Means
The modern smartphone is simply far too complex for theories like Dennard scaling and Moore’s Law to still apply. Interestingly, many people don’t know that Moore revised his law in 1995, stating that transistor density will take two years to double instead of one.
The smartphone of tomorrow may not be twice as fast, but it’s still improving. As for how we’ll be able to utilize our newfound processing power, only time will tell.