An Overview of PC Memory Types – CompTIA A+ 220-801: 1.3

| September 10, 2012

Your computer contains many types of memory. In this video, you’ll learn about ROM, SRAM, SDRAM, and other memory types. You’ll also learn an easy way to calculate the memory bus speed or memory throughput rates of DDR, DDR2, or DDR3 memory.

<< Previous Video: Motherboard Jumpers and ConnectorsNext: Understanding PC Memory >>

Review Quiz: An Overview of PC Memory Types

Shaded items are complete.

One of the key components inside of our computers is the memory. We’re usually referring to memory as the Random Access Memory, the amount of total RAM storage that you might have for that computer. But there’s all kinds of different memory inside of your computer.

When we’re commonly talking about memory, we’re usually talking about the RAM. But there may be occasions where we’re talking about certain types of memory that aren’t specifically associated with the RAM types that we’ll discuss in this video.

Another thing to keep in mind is your RAM memory is not the same as your hard drive storage or other storage device that you might have on your computer. Of course, we’re using the same type of measurement system for those. We talk about our memory sticks being a 4 gig or an 8 gig memory stick. And we’re talking about hard drive space also in gigabytes. So it’s very common to mix those up, if you aren’t exactly sure what you’re talking about.

But make sure when somebody says I have 8 gig in my computer, are they referring to how much memory they have in their computer or are they referring to some type of storage device inside of your computer? Sometimes people will use those interchangeably. And of course, those are very, very different things.

Memory is so important in our computers because the only time you can perform a calculation on anything is if you get that into the memory of your computer. So it’s very common for us to pull files off of our hard drive. And the only time we can execute them or perform any type of calculation is when we have it in active memory.

And we might perform some calculations. We might modify a graphics file. We might work on a spreadsheet. And then we will save that back to the hard drive for later. But the only time we can ever work with it, the only time we can ever execute a file, the only time we can run a program is when we have that information in the memory of your computer.

One of the kinds of memory we often don’t directly refer to is memory like read-only memory. This is very common to see when we are storing our BIOS information for instance onto our computer. This BIOS is something that we generally don’t change. And so it is a read-only memory in the way that we use it.

Read-only memory has many different types that you would see in a computer. A very old type is a PROM, a programmable read-only memory. That’s one where we would program the chip itself. We would fasten it to the motherboard. And it’s always going to have that particular code burned into that particular chip.

There’s also an EPROM, which is one that is the same as a PROM, but we can erase it if we have the right equipment. We would remove that chip from the motherboard. We would replace it into our burning device. So we can put different types of information on that chip. And then we would replace it onto the motherboard. Again, you’d have to physically remove it from the motherboard generally to have that happen.

In most of our computers today, especially when we’re dealing with the BIOS configuration of our computer, we’re talking about EEPROM. This is electrically erasable programmable read-only memory. And that is memory that we can simply take an application and change the contents of that memory just by running an application. It makes it very, very simple for us to upgrade the BIOS of our computer because we can do all of that without ever opening the case.

When we talk about CPUs and we talk about processors inside of our computer, we often refer to static RAM. This is not read-only memory. And it’s not the memory stick RAM that we would put in our computer. This is a very, very specialized type of memory that is very, very fast. It’s so fast that it’s exactly what we would like to use in the calculation process inside of our processors.

The problem is that this very, very fast memory is very, very expensive. Otherwise, we’d be using it for everything. We can only use a small part of it because of the costs involved with this static RAM.

It’s also relatively large. We’re looking at a CPU die right here. The static RAM are these large sections where all of this looks very similar. It’s like we’re looking at something from a plane.

And every time you see these little fields of information, those are generally static RAM sections. Usually it’s a layer 1, a layer 2, a layer 3 cache that is right there on the CPU itself. The static RAM is named static because we don’t have to constantly refresh the contents of the memory for it to be available to us. As long as we have power to our computer, our static memory will keep whatever we put inside of it.

It’s still volatile memory. If we were power off our computer, anything within the static RAM goes away. But as long as we can change it and provide constant power to these chips, it will always maintain that information within the static RAM.

Whenever we’re commonly talking about the memory inside of our computer, we’re almost always talking about dynamic random access memory. This is dynamic memory because we constantly have to refresh this memory to maintain the information that’s inside of it. If we were not going to refresh this data, it would ultimately disappear and completely wipe out everything that was inside of this memory.

That’s a little different than the static memory we were just looking at, where you write it and as long as there’s a constant power, it’s always going to be there. That’s not the case with dynamic memory. We always have to provide those updates. Otherwise, all of the information inside of it disappears.

One type of dynamic random access memory is one called synchronous dynamic random access memory or SDRAM. SDRAM is not static RAM. We have an “S” in front of that. But the S in this case stands for synchronous dynamic random access memory. It’s synchronous because the memory itself is synchronized with the clock rate of the memory bus. We have to make sure those are synchronized together so that they stay in sync and we can constantly refresh and access all of the information that we need from inside of that memory.

We don’t see SDRAM on a lot of our modern systems. Generally, we see DDR type memory. We’ll talk about that in a moment. But if you are going to acquire or replace some existing SDRAM inside of one of your computers, you’ll notice that the SDRAM is labeled with the same speed as your memory clock bus.

So if you have a memory clock bust of a 133 megahertz clock, you’re looking for SDRAM that is a PC133. And generally you’ll see that on the package that you purchase. Sometimes the memory itself, like this one, actually has listed on it PC133. That way I know the SDRAM I have here is designed for a memory clock bus rate of 133 megahertz.

As our computers matured through the years, we needed to make sure that the speed of our memory was keeping up with the faster and faster processors that were created. By the time the Pentium IV came along, we were hitting the limits of what that traditional SDRAM was going to be able to do for us. So a third party called Rambus created a new type of memory that could go much faster. And they partnered with Intel so that you could create some motherboards that would require this Rambus dynamic random access memory.

The advantage of course is that this memory is much faster than the older SDRAM. One of the disadvantages of course is that because it was created by a third party and a number of licensing requirements were there, that it was more expensive than traditional memory that we were using in the past. And so now you had a balancing act of needing to go faster, but this Rambus memory also required additional money. And that was a balancing act that most people had to determine if it made sense for them.

Not all systems use this Rambus memory. And as you’ll see in a moment, we evolved to have even faster memory types that were not created by the Rambus Incorporated.

Generally, our modern memory types are based on double data rate synchronous dynamic random access memory. This DDR SDRAM is one that we see very, very commonly. The first one that we had come out was simply DDR RAM. And we’re going to talk about that in just a moment.

It was able to go twice the rate of the older SDRAM, thus the double data rate associated with it. If we were to look at, for instance, a memory bus clock of 100 megahertz, we used what we called a bus clock multiplier of 1. We multiply that by the clock rate. This is a dual rate memory, so we’d multiply it by 2. And then we’d multiply by 64 to represent the number of bits that you can transfer in a single cycle.

And of course we wanted that represented as bytes, so we would divide it by the number of bits per byte. So that gives us this calculation of 12,800 divided by 8, which meant the total peak data rate for a 100 megahertz bus clock rate was 1,600 megabytes per second as a peak data rate. With the DDR memory, we would represent that as a PC-1600. Notice now, that the representation of the memory model did not match the memory clock rate.

So what we had was what I like to call a magic number. If we were to take the memory clock rate for DDR and multiply by 16, we would get the peak data rate. And conversely, if we had a PC-1600 data rate and you were trying to determine what the original memory bus clock rate was, you could simply take that 1600 number and divide it by 16. And ultimately, that would give you 100 megahertz.

The next generation of DDR memory was called DDR2. This is the double data rate 2 SDRAM. and there were a number of enhancements associated with it. We had the ability to handle more interference. We had buffers built into this. There were drivers off-chip. It was the next generation of DDR memory.

It also was twice as fast. So we would take our original 100 megahertz bus clock rate– notice that our bus clock multiplier changes to 2. All of the numbers stay the same after that. A dual rate of 2, 64 bits transferred, and you divide everything by 8. So your total number is 25,600 divided by 8, which is a total peak data rate of 3,200 megabytes per second.

DDR2 memory is designated by PC2. it has that 2 right after it. And the 3,200 is referring to the peak data rate. The magic number here for DDR2 memory then would be the number 32. If we take that 100 megahertz and simply multiply it by 32, we can skip all of these other equations and we can simply come up with 3,200. And conversely, we can take 3,200, divide it by 32 and get our memory bus clock rate of 100 megahertz.

The latest generation of DDR memory is DDR3. And as you can expect, it’s twice the data rate of DDR2. We even have much larger modules that we can use in our computers today. The formula, very, very similar. The only thing you’re really changing is the bus clock multiplier, which means you can take your 100 megahertz memory clock rate. And once you do the equation, you come up with a peak data rate of 6,400 megabytes per second.

And we can see here that we put a 3 on the end of the PC3 to designate that this is DDR3 memory and then our peak data rate of 6,400. That means that magic number that you can use, instead of going through that big set of equations, is that you can take the 100 megahertz clock rate and simply multiply it by 64 to come up with 6,400 megabytes per second as your peak data rate. And of course, you can conversely do this by dividing 64 from your 6,400 peak data rate to give you 100 megahertz.

So if you’re given a question where you need to calculate either your peak data rate or you need to calculate the original memory bus clock rate of a particular set of DDR, DDR2, DDR3 memory, just remember your magic numbers. And you’ll simply multiply or divide for the appropriate magic number of the memory type that you’re using.

Even though these types of memory are numbered DDR and DDR2 and DDR3, you cannot use different types of memory on different motherboards. You can’t take a DDR2 memory and use it on a motherboard that’s designed for DDR3. Even if you try to do that, you’ll notice that it doesn’t fit onto the motherboard.

So if you’re moving memory around and you notice it’s not quite going into the slot, there are these sections on the memory that are designed to fit right into a key that is on the memory slot itself. This ensures that when we are plugging memory in, that we’re not using the wrong memory type on that motherboard. It should fit perfectly. And if it doesn’t, you may want to look at the tiny little slot that is in your memory and make sure it lines up with the plastic key that is inside the memory slot on your motherboard.

Tags: , , , , , , , , , , , , , , ,

Category: CompTIA A+ 220-801

Comments are closed.