Computer System Components

This chapter introduces the lower-level components used in the construction of computer systems. We begin with the metal-oxide-semiconductor (MOS) transistor, which is employed extensively in memory circuits and across virtually all other modern digital devices. We will examine the design of MOS transistor-based computer memory and its interface with the processor. We’ll look at modern computer input/output interfaces, with a focus on the use of high-speed serial communication within the computer case, as well as data transfer over cable connections to external components. The functional requirements of system I/O devices including the graphics display, network interface, keyboard, and mouse will be discussed. The chapter ends with a descriptive example of the specifications for a modern computer motherboard.

After completing this chapter, you will have a solid understanding of the hardware components of modern computer systems, from technical...

Technical requirements

Files for this chapter, including answers to the exercises, are available at https://github.com/PacktPublishing/Modern-Computer-Architecture-and-Organization-Second-Edition.

Memory subsystem

The memory subsystem is an addressable sequence of storage locations containing instructions and data for use by the processor as it executes programs. Modern computer systems and digital devices often contain over a billion 8-bit storage locations in main memory, each of which can be independently read and written by the processor.

As we saw in Chapter 1, Introducing Computer Architecture, the design of the Babbage Analytical Engine included a collection of axes, each holding 40 decimal digit wheels, as the means of storing data during computations. Reading data from an axis was a destructive operation, resulting in zeros on each of an axis’s wheels after the read was complete. This was an entirely mechanical method of data storage.

From the 1950s to the 1970s, the preferred implementation technology for digital computer memory was the magnetic core. One bit of core memory is stored in a small toroidal (donut-shaped) ceramic permanent magnet. The set...

Introducing the MOSFET

Chapter 2, Digital Logic, described the NPN transistor, a type of bipolar junction transistor (BJT). The NPN transistor is called bipolar because it relies on both positive (P) and negative (N) charge carriers to function.

In semiconductors, electrons serve as the negative charge carriers. There are no physical particles with a positive charge involved in a semiconductor operation. Instead, the absence of a normally present electron in an atom exhibits the same properties as a positively charged particle. These missing electrons are referred to as holes. Holes function as the positive charge carriers in bipolar junction transistors.

The concept of holes is so fundamental to semiconductor operation that William Shockley, one of the inventors of the transistor, wrote a book entitled Electrons and Holes in Semiconductors, published in 1950. We’ll next examine the behavior of positive and negative charge carriers in unipolar transistors.

As an...

Constructing DRAM circuits with MOSFETs

A single bit in a standard dynamic random-access memory (DRAM) integrated circuit is composed of two circuit elements: a MOSFET and a capacitor. The following section presents a brief introduction to the electrical characteristics of capacitors.

The capacitor

A capacitor is a two-terminal passive circuit element capable of storing energy. Energy enters and leaves a capacitor as electrical current. The voltage across the capacitor terminals is proportional to the quantity of electrical energy contained in the capacitor.

To continue the hydraulic system analogy introduced in Chapter 2, Digital Logic, think of a capacitor as a balloon attached to the side of the pipe leading to a water tap. Water pressure in the pipe causes the balloon to inflate, filling it with some of the water from the pipe. Let’s assume this is a strong balloon, and that as it inflates, the balloon stretches, increasing the pressure within. The balloon fills...

I/O subsystem

Chapter 3, Processor Elements, introduced two broad categories of I/O architecture: memory-mapped I/O and port-mapped I/O. The pros and cons of each of these approaches were significant in the early days of PCs when the number of physical address lines limited the total processor memory space to a 1 MB range. Modern processor architectures can address a far larger memory range, typically in the tens of gigabytes. A consequence of this address space expansion is the ready availability of address regions for use in I/O interfaces. Because of this, modern 32-bit and 64-bit general-purpose processors employ memory-mapped I/O for most of their interface requirements.

Sophisticated modern processors usually implement a memory controller within the processor chip, communicating directly with DDR memory modules. Most other types of I/O performed by these processors are offloaded to one or more external integrated circuits, typically referred to as a chipset. The term chipset...

Graphics displays

In the domains of gaming, video editing, graphic design, and animation, video processing performance is critical. Generating and displaying high-resolution graphics requires an enormous number of mathematical calculations.

While general-purpose processors can perform the necessary computations, such processors lack the performance users of these applications have come to expect.

High-performance graphics cards, called graphics processing units (GPUs), are essentially miniature supercomputers, heavily optimized to perform graphical computational tasks such as 3D scene rendering. Because the computations involved in scene rendering are highly repetitive, substantial performance gains can be achieved with the use of hardware parallelization. Graphics processors contain a large number of relatively simple computational units, each performing a small portion of the overall task.

A GPU may contain thousands of individual processing units that each function...

Network interface

A computer network is a collection of digital devices interacting over a shared communication medium. A local area network (LAN) consists of a limited number of computers that might reside in a single physical location such as a home or office building. The connected computers, phones, and other digital devices in your home represent a LAN. Device connections within a LAN environment may use a wired interface, usually Ethernet, or a wireless interface, typically Wi-Fi.

Geographically separated computers and LANs communicate using a wide area network (WAN). WAN services are often provided by a telecommunication company such as a cable television provider or telephone company. Your home LAN most likely connects to the Internet via WAN services provided by your telephone or cable company.

Home and business network interface devices (called routers) provided by WAN service providers usually offer Ethernet and Wi-Fi options for connecting local devices to the...

Keyboard and mouse

Compared to the high-speed interfaces discussed earlier in this chapter, the bandwidth requirements for a keyboard and mouse are quite modest. These devices are the sole input methods used by the human operator in most computer configurations, and thus are only required to operate at the speed of human actions. Even the fastest typist can only press one or two dozen keys per second.

Keyboard

A mechanical computer keyboard consists of a collection of keys, each of which activates an electrical momentary switch. A standard full-size keyboard typically contains 104 keys including the arrow keys, control keys (Home, Scroll Lock, and so on), and the numeric keypad. Modern keyboards commonly use a USB cable for connection to the computer or connect wirelessly.

Because the bandwidth requirements for human interaction are so low, some computer motherboards provide a slower USB 2.0 port for keyboard connection while offering higher-performance USB 3.2 or faster...

Modern computer system specifications

With the information in this chapter, you should be able to interpret most of the technical features of a modern computer motherboard, processor, and chipset. This section provides an example of the specifications for a current (2021) motherboard with some explanation of the individual features.

The designers of a computer motherboard must make several decisions such as the number of PCIe expansion ports, number of DIMM slots, number of USB ports, and the number of SATA ports to be provided in a particular motherboard model. These decisions are guided by the target customer demographic, whether it be gamers, business users, or cost-conscious home users.

The example motherboard presented here is the ASUS Prime X570-Pro, a higher-performance board intended for gaming applications that supports gaming-related features such as overclocking. Overclocking refers to increasing the clock frequencies for the processor and other system components...

Summary

This chapter began with an introduction to the computer memory subsystem, the MOSFET, and the capacitor. We examined the circuitry that implements the DRAM bit cell. We reviewed the architecture of DDR5 memory modules and the operation of multichannel memory controllers. Other types of I/O devices were introduced, with a focus on high-speed differential serial connections and their ubiquitous use in technologies such as PCIe, SATA, USB, and video interfaces.

Popular video standards were presented, including VGA, DVI, HDMI, and DisplayPort. We looked at the Ethernet and Wi-Fi networking technologies as well. We continued with a discussion of standard computer peripheral interfaces including the keyboard and mouse. The chapter concluded with a description of an example modern motherboard, highlighting some of its interesting features.

With the information presented in this chapter, you should have a solid understanding of modern computer components from the level of...

Exercises

1. Create a circuit implementation of a NAND gate using two CMOS transistor pairs. Unlike NPN transistor gate circuits, no resistors are required for this circuit.
2. A 16-gigabit DRAM integrated circuit has two bank group selection inputs, two bank selection inputs, and 17 row address inputs. How many bits are in each row of a bank in this device?

Join our community Discord space

Join the book’s Discord workspace for a monthly Ask me Anything session with the author: https://discord.gg/7h8aNRhRuY