Every once in a while and repeatedly the same question arises again with these two types of processor architectures. The excuse now lies with Apple leaving Intel behind and rushing for its own ARM chips, and although the decision is complex, many do not understand how they have taken this path. Therefore, we will try to compare ARM vs X86 in its characteristics, approaches and basic concepts to understand this jump.
The main problem that we are going to face in this article is concepts, which is at the same time the key that Apple solves for the jump. And the fact is that the company has not disputed the performance crown in several sectors for a long time, but is focusing on its own and above all more efficient ecosystem.
ARM vs X86, a story that has no end
We assume that we all know where the name X86 (Intel 8086) comes from and based on this we have to understand where computing comes from. X86 was based on CISC technology and architecture, where it has always been characterized by having very extensive sets of instructions to solve more and more complex operations (hence its name).
The evolution of CISC architecture as a concept and basis was to move from these already complex instructions to simpler ones and this is where ARM comes in. Both Intel and AMD realized that they had to create microinstructions to simplify tasks and thereby gain performance in certain parts, always without losing the characteristic performance of CISC in complex operations.
As we know for sure, CISC spent some time competing with another architecture called RISC, much simpler than the first one, but which pushed micro instructions to end up surpassing it in all aspects. Although it was believed abandoned as such, RISC took off again from the hand of ARM or Advanced RISC Machine , a company created to license the manufacture of processors with this technology, occupying a gap that both Intel and AMD were completely ignoring due to the fight by the crown they have always had.
Although many predicted an ARM failure, the company knew how to enter several markets, among them smartphones, ovens, microwaves, televisions and a large etc … Which has led the company to ever higher levels to the point that Apple is about to change its philosophy based on processors licensed with ARM technology.
ARM vs X86: space, features and consumption
As we see, ARM vs X86 are antonyms rather than rivals, since as a rule they do not usually bother, at least until recently, something that we will discuss later. Both CPU architectures had a well-defined market: X86 was much faster, but it also takes up much more silicon and is much higher in consumption.
ARM instead seeks efficiency, a balance between consumption and performance based on fewer stages of processing, but cores with lower speed and divided into two very specific specifications: high performance and low performance, the so-called big.LITTLE architecture .
While a desktop or server CPU can go between 65 watts and almost 300 watts at peak, an ARM processor will be above 5 or 7 watts and 150 in the case of servers. That enhances the core factors. Intel and AMD opt for large-scale complex cores, so in the same space as an ARM they can include fewer of them, but very fast and solvent.
On the other hand, ARM starts with the advantage of simplicity and with it the size of the cores is considerably reduced compared to their rivals, therefore they are capable of accommodating a greater number of these although they are slower and at lower average speed, achieving that workloads are divided and can be further optimized in certain environments such as servers.
They don’t use the same software, it has to be carried
Although efforts by both Apple and Microsoft are heading in the right direction to achieve software hegemony that runs on both ARM and X86, it seems that those in Cupertino are going with some advantage, driven of course by leaving Intel permanently.
Operating systems need to be created and optimized largely for each of the architectures, especially if we want to get the most out of X86 or ARM. In this aspect it is easier to program for ARM than for X86, since the latter requires more complex compilers, better compaction of the code and better debugging of the same due to the number of instructions they manage to carry out.
In addition, it is the case that the overall ARM ecosystem is much smaller than that of X86, although Android and iOS have helped reduce the gap. So the question is clear, are they really competition? Years ago the answer would have been an almost resounding no, but currently this cannot be affirmed and less every day.
ARM goes for servers, Intel goes for smartphones?
ARM is trying to get fully into the world of servers through various companies and currently giants like CloudFlare have been testing the benefits of their architectures. Although they are opening a hole and they have to improve, today ARM is feasible for business environments and data centers that are very focused on certain workloads, where curiously neither Intel nor AMD can compete.
We speak for example of OpenSSL , certain compression algorithms, NGINX and the like. They are obviously very specific tasks, but ARM here is competitive by its nature, on the other hand, in tasks such as Regexp, Go Crypto etc … ARM has nothing to do.
If ARM is entering this sector, Intel wants to go straight to that of its rival with Lakefield, architecture based on its own concept of big.LITTLE with 3D stacking and only 7 watts of consumption, which is a reality and has already been presented.
Interestingly, Apple is now leaving Intel and focusing on its new chips that promise up to 12 low-power cores, just as the blues are going to implement the first major architectural change since Skylake and have marked a roadmap at a devilish pace. Therefore, the story between ARM vs X86 is not even at its zenith, it is just beginning, perhaps in two decades we can see where it ends.
