Welcome to the fourth installment in the All About Computer Memory and How It Works series here on Be Your Own IT. In this installment, we are going to take a closer look at your computer's system RAM. You are already most likely aware that your computer needs system RAM to run correctly and that the more system RAM you have available for your central processing unit to use, the faster and more efficient your computer system will run. Knowing that is sometimes enough for most people, but if you really want to be your own IT, then you need to understand how system RAM works in more in depth detail.
The speed of your RAM is controlled by two things – one is called bus width and the other is called bus speed. The bus width of your RAM refers to the number of bits that can be sent to the central processing unit in your PC at the same time, and the bus speed refers to the number of times that a group of those bits can be sent to the central processing unit per each second. Every time data travels from memory to the CPU, it is called a bus cycle. You can think of bus speed as describing how wide a road is or how many cars can fit on that road side by side while they drive along. You can think of bus speed as how fast the cars on that road are able to travel and complete their intended trips.
When you buy new system RAM, the number of bits your RAM has makes a huge difference. A 32 bit stick of system RAM can send three times as much data as a 16 bit stick of RAM, even though the 32 bit stick of RAM is technically only twice the size of the 16 bit stick. This really boosts the amount of data that can be passed to the CPU in a second, and the more data that can be passed over means that your CPU will not be idling around as often, waiting for the information it needs to continue.
But RAM does not always perform at the speeds its specs say it will. There are many other factors that go into determining how fast your system RAM will actually be. One of these things is something called latency, which refers to how long (technically, how many clock cycles) is needed to actually read a bit of information. RAM that is rated at 100 MHz, for example, may be technically capable of sending a bit of information to the CPU in 0.00000001 seconds, but it may actually take 0.00000005 seconds to start the process of reading the information for that first bit.
Computers compensate for latency in a couple of different ways that work together to speed up the process. A process called “pipelining” organizes data pick up into something resembling an assembly line, where the memory controller will read one or more words from memory at the same time while sending the current word or words to the CPU, and then writes one or more words to memory cells – all at the same time.
Pipelining is combined with burst mode to reduce latency. Burst mode is a process that lets the CPU anticipate that whatever information it needs next will come from the same basic area as the current information it is using hails from. This means that latency will only affect the first bit of data, and that consecutive readings of subsequent bits will be much faster.
To make the most of your system, only buy RAM that your system is actually capable of handling. You can pay for super fast RAM, but if the rest of your system is unable to use its top speeds, then the RAM will only be able to perform as fast as the other components can.
Stay tuned to the next installment of this series, where we will talk about caching, another way your computer uses memory to make your user experience faster and more efficient.
