Hello, everyone! I’ve been doing some research into computer hardware lately, and one thing that I keep coming across is this idea that the speed of a processor, while important, isn’t the only thing that affects a computer’s overall performance. I’m having a bit of a hard time wrapping my head around this because I always thought that a faster processor meant a faster computer. Can anyone explain why this isn’t necessarily the case? I’m really interested to learn more about this!
In: 97
Think of your computer as a tiny office full of people.
The CPU is the executive and computation department. It crunches numbers, but it has very limited space for materials. The CPU also has executive capacity to control the rest of the computer.
Memory is the rapid-access records department. Because the CPU has limited space, it needs a place it can quickly access. The memory department is full shelves that are very close to the CPU department. The memory department is in charge of storing and fetching stuff from the shelves, but the CPU department tells them where to put things. When the office closes for the evening (when the computer is shut down), everything in the memory department is sent to the shredder.
The SSD/Hard Disk departments are warehouses for long term storage. The CPU sends data here that needs to be maintained, even when the office closes. The SSD department is quite a bit faster at retrieving materials than the Hard Disk department, but the Hard Disk department tends to have more square footage.
The Graphics Card department is in charge of what is shown to customers. Everything a customer see goes through this department. Pretty much all Graphics Card departments can handle basic requests like displaying photos and video, but it takes specialized staff to animate game instructions sent from the CPU department in real time.
So given what we know about this office, let’s think about how each affects speed.
An better CPU department means you can crunch numbers faster, but since there is limited space there, what happens if the CPU is done calculating, but the Memory department isn’t big enough to hold the result? The CPU will have to send some of the data off to the SSD warehouse before it can send more results to Memory. The SSD department is pretty quick, but you’re still dealing with a warehouse, and the SSD department is a warehouse, so it’s in a separate building. The CPU ends up waiting on the Memory department to clear out space so it can send over more results.
From the example above, we can see how a fast SSD/Hard Disk has a cascading effect on Memory, and therefore the CPU. If the CPU asks the Memory department to clear space, that data has to go to the SSD/Hard Disk warehouse. So the entire operating can end up waiting on the warehouse.
The Graphics Card department is a little bit special. These departments vary pretty widely depending upon what type of tasks the office will perform. Pretty much all of these departments have projectors for photos and film, but gaming is different. Gaming comes in as a set of descriptions of the scene, and the Graphics Card department has to literally draw each one. To do this, they employe a massive number of workers. They split the scene up into little parts, and dispatch each part to a small working group so they all work in parallel. The results are sent back to the head of the department to be assembled.
Because of the complexity of this work, some Graphics Card departments have their own, separate Memory department. If your department drawing a very large, complex scene multiple times per second, you need *very* rapid access to storage so you can rapidly work through frames of the scene and put them in storage to be displayed on the screen very quickly.
So if the Graphics Card department isn’t able to keep up with the requests from the CPU (who is managing everything), the customer will have to wait for that department to catch up.
Hopefully the picture is becoming clear that a computer is a complex system with interconnected dependencies. It’s worth noting though that there is some flexibility in that interconnected nature. You can usually identify one department that is most frequently holding the rest up. Upgrading that *single* component *will* have the net effect of speeding up the entire computer.
A good example of this is moving from a hard disk to an SSD. Hard disks are literally metal boxes with spinning plates in them. These plates work a bit like an old vinyl record. Data is encoded in rings of magnetic charges on the plate. If you want to read data from a hard disk, a little arm has to sweep over to the ring where the data is stored. If the data is spread out across multiple tracks, this little arm has to move around. That little arm can only move back and forth so fast before its inertia becomes too great. Wave your arm back and forth really quickly. Faster… Faster! Try to do that 100 times per second. See how much effort it takes? That’s going on inside a hard disk all the time.
An SSD, by contrast, has no moving parts. Every internal component is made from a semiconductor. The only movement is at an atomic level, and atoms have very, very little mass. This means they have almost no inertia, so they can switch much faster than a tiny metal arm.
A lot of a computer’s time is spent waiting on long term storage. If you’re using a hard disk for long term storage, then you’ll spend even more time waiting on that aspect of the computer’s operation. Upgrading to an SSD will often result in a massive improvement in performance because SSDs are >10 times faster than hard disks when it comes to grabbing random stuff from the warehouse.
Contrast this with a CPU upgrade. Most modern CPUs are incredibly fast and are actually multiple CPUs packaged as one CPU. They can compute so quickly that most modern CPUs spend the majority of their time sitting idle. Tasks that do require the CPU execute more quickly, but instead of waiting 300 milliseconds, we wait 250 milliseconds. It can be difficult to notice the difference.
Latest Answers