Our modern CPUs include advanced technologies for computations, graphics, virtualization, and memory management. In this video, you’ll learn about processor cores, caches, overclocking, and more.
<< Previous Video: Installing BIOS Upgrades Next: CPU Cooling >>
If you’ve ever seen a picture of a CPU or installed one in your computer, you know that it’s a single chip. There’s a lot of things happening inside of that chip. And in this video, we’ll look at those CPU features. Older CPUs were designed to process one transaction at a time, but we’ve found ways to increase the number of CPU use that are available on one single processor. We call these separate CPUs separate corners of the processor.
And there could be dual core, four cores, eight cores, or even more in modern PCs. The block diagram that I have here shows four different cores, and each core has its own CPU. This means that four separate transactions could be occurring simultaneously.
In my diagram, I’m also showing a level 1 and a level 2 CPU cache that is independent to each individual core. This particular processor also has a level 3 cache that has a shared cache for all four cores on this processor. If you were to look very closely at the die of a CPU, you can almost pick out the individual cores, or you may be able to see where cache memory might be stored on the CPU itself. If you’re interested in seeing more of these CPU dies, you could perform a Google search for this or go to Intel or AMD’s websites to be able to find the CPU die for your particular processor.
Here’s a block diagram of one of those CPU dies. Here’s our four cores that are in this processor. This particular processor also has a graphics processor associated it. There’s a memory controller as well. And here’s our level 3 cache– that cache memory that’s in this particular CPU.
Let’s look at level 1, level 2, and these level 3 caches to see how they help improve the performance of this processor. This cache memory is very fast memory, and it’s meant as a temporary holding space for the transactions that are going through the processor. There’s very commonly different levels of cache on processors, and the number of levels and where they’re located will depend on the model of the CPU.
Level 1 cache is usually the one that is closest to the processor itself. It is providing that first check or first storage of data. The level 2 cache may also be part of the same course, such as the diagram that we have here. It’s the secondary level or secondary storage of cache memory. And level 3 cache these days is still very often on the same physical processor, although in my particular case, it’s showing this as a shared cache across all four of those cores.
We tend to use virtualization extensively in our data centers and even on our desktop computers these days. Virtualization allows us to run multiple operating systems on the same piece of hardware simultaneously. This means that I could be running the Mac OS operating system on my desktop, but also have a window where I’m running the Windows operating system and a Linux operating system all at the same time.
When you’re virtualizing operating systems, there’s a lot of things happening behind the scenes. Each part of memory that is associated with each operating system has to be kept separate. There’s also separate storage and separate transactions that are occurring, and all of these are independent to each individual virtualized system. You could certainly perform all of this virtualization in software and have the operating system manage exactly what data should be associated with which virtualized system.
But as virtualization evolved, we needed faster and more efficient ways to handle these separations between these virtual machines. So now inside of our processors, we have specialized hardware that’s focused on making virtualization more efficient and much easier to manage. If you’re using an Intel processor, this hardware is called Intel’s Virtualization Technology or VT. And if you’re using an AMD processor, AMD calls this AMD virtualization or simply AMD-V.
I mentioned earlier that the core of a CPU is designed to execute one transaction at a time. Information is moved into a CPU. The CPU performs the transaction, and the results of that transaction are moved out of the CPU. The process of moving information in and out of the CPU, though, takes time. And while that information is being transferred in and out of that CPU, the CPU is sitting idle.
Well, instead of it being idle, engineers have found a way to run additional transactions at the same time. This technology is called threading. Hyper technology or HTT where one single CPU can act as if it is multiple CPUs.
So while the transfer process is happening in or out of a CPU, another transaction could be executing information at the same time. In reality, this does not operate as if there are two separate CPUs, but it does increase the overall efficiency. And you can see an improvement of around 15% to 30%. The operating system that you’re using has to support hyperthreading. But these days, that’s almost any modern operating system.
So as long as you’re using any operating system that is newer than Windows XP, you’re able to support hyperthreading technology. If you look at the specifications of a CPU, there’s always going to be a processor speed value listed. For example, this CPU has a processor speed of 3.80 gigahertz.
This is referring to the total number of transactions that can occur in a single second. Older CPUs were rated in megahertz, which is millions of transactions in a single second. But modern CPUs are often rated in gigahertz, which is how many billions of transactions can occur in a single second. Traditionally, we’ve looked at the processor speed of a CPU to determine just how fast a computer might be.
But in reality, there are many different components that all work together to determine the speed of a computer. For example, clock speed is certainly important, but you also have to examine the overall architecture of the CPU, the bus speeds, the bus widths. Are there any caches that are part of the CPU?
And, of course, how the operating system will take advantage of these capabilities of the CPU. This is one of the reasons that computer manufacturers have moved away from promoting processor speed as being equal to the overall performance of a computer. But this also means that there is no broadly accepted single value that we can associate with a computer that allows us to compare and contrast the different capabilities or the different performance between systems.
In reality, you’ll want to use a benchmark that works best for you. If you’re someone who performs a lot of high-end transactions, you want to try that across multiple systems. If your requirements are more for browsing and email, then you’ll have a different set of benchmarks that you might use.
When Intel and AMD build their processors, they build and test them to operate at a particular clock speed. But some computer enthusiasts have found that they can increase this clock speed and get a little more performance out of their CPUs. This process of increasing the clock speed is called overclocking.
Overclocking allows you to get more performance, but it also is going to use more power. It will certainly create more heat from your CPU. And at some point as you get higher and higher and higher numbers for clock speed, you’ll start to find that your system becomes unstable. Not all motherboards allow you to perform these experiments on your CPU and change the clock speeds.
But if your motherboard does, then there’s probably a setting in your BIOS called Base Clock or BCLK. This also requires that the CPU that you’re using allows you to increase the clock speed. Not all CPUs are unlocked, so you have to make sure not only that you have the right motherboard, that you’re using a CPU that can support overclocking.
Once you’ve increased the clock, you can then perform a series of stability tests– run some tests overnight to see how stable your computer acts at this higher clock speed. As you can imagine, the manufacturers of these CPUs do not support overclocking, so you will void the warranty if you try to overclock your system. And of course, increasing the power and heat can not only damage your CPU but other components as well.
On older computers, we use separate video cards to provide the video display that we see on our screens. But as our technologies have improved, we’ve been able to take a lot of the video technology and integrate it directly into our processors. On many computers, you’ll find this graphics processor is built into the CPU, so you don’t necessarily need a completely separate video card in your system. If you’re doing video editing on your computer or you’re a gamer or you need some higher end graphics, then you still might want to have a separate independent graphics card in your computer. But if you’re simply browsing the internet, watching some streaming video, or checking your email, then the integrated graphics processor CPU might be the best choice.
Two of the very big players in consumer CPUs are Intel and AMD, and there are advantages and disadvantages for both of those companies’ products. But the differences are slowly merging together– becomes more and more difficult to decide whether the right chip for you is Intel or whether the right chip might be AMD or whether it wouldn’t matter at all depending on your requirements.
One of the requirements, obviously, is going to be cost. The cost of an AMD tends to be a bit less expensive than Intel. And that allows them to sell many more entry-level PCs than you might find with Intel chips.
On the laptop side, Intel has historically had a larger portfolio of processors to choose from, and that certainly helps the laptop manufacturers with their options. But this is a very dynamic segment of the industry, and we tend to see changes happening all the time. You want to find the chips and the solution that works right for what you need to do on your computer.