What is more important — the number of cores or the clock speed of the processor

If you take a look at the current processor market, it may seem that chip companies have decided to enter the Guinness Book of Records. AMD Unveils 32-Core Threadripper 2290WX Processor! «Intel released the Core i7-8086K processor, which can overclock up to 5.0 GHz»! New performance records are set every year, and as a result, it’s easy to get confused about which is better — the number of cores or clock speed?

Let’s try to understand this issue.

A little analogy

Suppose you need to transport a load. Lots of cargo. And there are two options — to transport one very fast car or several slow ones.

Using several slow cars allows you to transport a lot of cargo, but not very quickly. Let’s say he travels from one city to another for three days. But he will come all at once, and that’s great.

One fast car can carry a little less cargo. But from one city to another, she will go all day. Then she, however, will have to return, pick up the rest of the cargo, bring it again, and again, and again.

The ideal solution to this problem is to use several fast machines! Then a lot of cargo will be transported, and it will take only one day. But we live, unfortunately, not in an ideal world.

It would seem, what do the processors have to do with it? Yes, just a “heap of cargo” is a flow of calculations that a chip needs to perform. “A lot of slow machines” is a few low-performance cores. And “one fast” is one high-performance core, with an increased clock frequency.

And the solution to the question “high clock frequency or many cores” depends primarily on the planned scope of use.

How the processor works

The processor is engaged in the fact that it performs the simplest arithmetic calculations. The operating system and software environments turn user commands into the very arithmetic tasks that the chip needs to solve. This is done through interpreters and algorithms that indicate links between high and low level programming languages.

The processor receives a lot of instructions all the time. Each user action requires several calculations at once (or even tens of thousands). And these commands must be executed, otherwise the computer simply will not function.

Teams can be very different. For example, to interpret the HTML code of a page (including this one) into a graphic image on the screen. Or decrypt a compressed ZIP archive. Or even turn something cryptographically encrypted into a display in the clear. In general, the tasks are very different and require different resources to complete them.

So, for example, converting HTML code will require a minimum of resources and can be performed quite quickly even on a relatively slow chip. ZIP archives already require significantly more computer time. And cryptographic algorithms like SHA-256 calculations completely load the unfortunate chip like God knows what, and at the same time cannot be executed on multi-threaded configurations.

The calculation of artificial intelligence — for example, in games — requires a lot of calculations. But they are small, isolated, and therefore can be perfectly executed in a multi-threaded version — on several cores at once. Thanks to this, high performance in games is achieved.

But archiving and unzipping files is an exclusively single-threaded operation. So one powerful chip is required, otherwise this process will take a lot of time. Especially if the archive contains a lot of small heterogeneous files.

And yet, which is better — the number of cores or the clock speed of the processor?

The problem is that computers are used in different ways. In the morning you download firmware from the Internet to your phone, and there are two thousand files of 500 kilobytes in one archive. At lunchtime, you sit and look at cat gifs on the Internet. In the evening you play, fighting against numerous enemies in virtual reality.

That is, in the morning the computer needs a processor with a high clock speed, at lunchtime — just at least some “chip”, and in the evening — with many cores. And «poke» «stones» depending on the intended use is clearly not the best idea.

That is why manufacturers are trying to produce multi-core configurations with high clock speeds. So, for example, the Intel Core i7-8086K (the top one in the Kabu Lake line at the time of writing) is equipped with six processing cores with HyperThreading technology and a base clock speed of 4.0 GHz. He can do everything! Expensive, though — $ 425 at the time of release.

For home use, choosing which is better — clock speed or number of cores — is not worth it. The ideal solution would be to strike a balance. For example, buying some kind of quad-core chip with a base clock speed of 3.0 GHz. Its performance is enough for the vast majority of everyday tasks.