The final piece of your computer that we need to cover is “volatile memory”. We know from earlier that memory is a term for anything that can store data. “Non-volatile” meant that the data was not lost when you turn off your computer. “Volatile” means the opposite; the data is lost when you turn off your computer.

It may not be immediately clear why we would want volatile memory. Choosing between losing your data and not losing your data seems simple. However, there is more to this choice than just volatility.

One drawback to the hard drive that we haven’t seen yet is that it is slow. It’s hard to give exact numbers because the technology is constantly improving. But if we rewind to roughly the year 2000, we can look at some approximations.¹

Suppose we wanted to read some data that is already stored in memory. In 2000, most people had hard disk drives as their non-volatile memory. You could read from these in approximately 20,000,000 nanoseconds. In comparison, volatile memory would have only taken 250,000 nanoseconds, 80 times faster. The solid state drives that were coming closed the gap significantly, coming in at 1,000,000 nanoseconds, only 4 times slower than volatile memory.

Rephrased in a more human scale, if reading from volatile memory takes 1 day, then a solid state drive would take 4 days. That’s not too bad, but it is noticeable. Hard disk drives on the other hand would take 80 days. That is a large enough difference that I would do everything I could to use volatile memory.

RAM

In a perfect world, we would have hardware that provides non-volatile memory and is fast. But for now, we are stuck with two separate pieces of hardware: the hard drive, which doesn’t lose data but is slow, and RAM, which is fast but loses data.

RAM stands for “Random Access Memory”; it is the hardware that provides volatile memory. As the name suggests, it is exceptionally good at “random access”. Hard drives work best when you read/write data in one spot, such as in a single file. But RAM still performs well when you read from one spot, and then write somewhere else, and then read from another random location.

RAM is our workspace when we are programming. We have access to this great big storage space that we can use however we would like to. Often we use it to store intermediate results when we are working. For example, I was taught to multiply two digit numbers in a multi-step process. It looks something like this,

  23
x 44
----

First, we multiply with the one’s digit of 44.

  1
  23
x 44
----
  92

Then we add a 0 to the second line and multiply with the ten’s digit.

  1
  23
x 44
----
  92
 920

Finally, we add the two results together for the result,

  23
x 44
----
  92
+920
----
1012

When we are programming, we need some place to put 92 and 920. They aren’t part of the result, but they need to be stored somewhere while we are calculating the next steps. Once we have the result, the space that held 92 and 920 can be cleared and used for something else.

Vim

Let’s look at one more example which is a bit more complex.

When you type a new letter into Vim, what happens?

One possible implementation would be to update the file on the hard drive every time you type. The moment you press e in insert mode, an instruction is sent to the CPU to add an e to the file, and then the CPU sends a request to the hard drive to update the contents of the file.

This would work, but it could lead to an unpleasant user experience. You probably don’t usually notice a lag between pressing a key and seeing it appear on the screen. But if Vim had to update the file on the hard drive every time you added or deleted a letter, that lag might become noticeable. It might even become so noticeable that you decide to try a different text editor.

An alternative implementation would be to keep the contents of the file in RAM. Now when you type an e, instead of sending an instruction through the CPU to the hard drive, it sends an instruction through the CPU to RAM. This makes the user experience fast, so you don’t notice any lag between typing and seeing the results on the screen.

Our end goal is to modify the file on the hard drive, so we also need a method to update the file with the changes we made in RAM. When you enter the :w command, you are asking Vim to “write” the changes from RAM onto the hard drive.

This process of moving changes from volatile memory to non-volatile memory is common; although it is usually called “saving” instead of “writing”. You may have seen warnings before that say something like,

Are you sure you want to quit without saving? All unsaved progress will be lost.

That warning message is trying to tell you that you made changes in volatile memory but did not update the hard drive. If you quit without writing or saving, those changes will be lost.

You may have also run into this if your computer lost power while you were working. If the power goes out or your battery runs out in the middle of working on something, you may find that the next time you go back your most recent changes are missing. This is the same issue. Your changes were in volatile memory and were lost when the power went out.

Your imagination is the only limitation on how you use RAM. It is your workspace to use however you would like when you are programming. Based on the name, we know that random access memory is used to store data, like any other type of memory. But we don’t have the same tools to work with RAM that we have for the hard drive. There is no filesystem for creating and destroying files or directories, and there is no Vim to let us write the data we want.

Instead, we have to work more closely with the hardware and define how we want to use it. But to do that, we’ll need to learn a bit more about how RAM works.