Computer Hardware

  • The CPU (Central Processing Unit)

    Computers perform calculations using a Central Processing Unit (CPU). The computer runs at a pace determined by the CPU's clock speed, sometimes referred to as processor speed. A CPU has an internal clock that runs at a constant speed. A computer’s clock speed is measured in cycles per second, known as Hertz (Hz). A computer’s clock cycles billions of times per second (gigahertz or GHz). For instance, a 3.4 GHz CPU clock cycles 3.4 billion times per second.

    All else equal, the faster the clock speed, the faster the CPU will perform. But because other factors limit computer performance, increasing the clock speed a certain percentage does not mean that computer performance will increase by that same percentage. Other factors like the amount of cache, amount of RAM, and the number of CPU cores also impact computing speed.

    Today, high-performance computers have multiple physical cores. Each physical core on a processor has the same clock speed. The number of physical cores is the number of independent physical processors on the CPU. Having multiple cores means that the computer can actually do more than one thing at a time. Each physical core in can also be optimized using Multithreading and/or Hyperthreading.

    Multithreading

    This occurs when a computer with a single core can appear to run multiple processes at once by switching back and forth among processes many times per second. Multithreading occurs when more than one process is broken into slices and then slices from different processes are interwoven so quickly that it appears that multiple programs are running at the same time. In reality, only one process is executed at any given point, but it switches so fast that it creates the illusion of two processes occuring simultaneously.

    in Multithreading, the operating system (OS) sets up a timer, which interrupts the system at a fixed interval. A single interval is known as a time slice. Every time this interruption occurs, the OS:

    1. prioritizes among waiting programs to select the next thread that needs to be executed
    2. The state of the partially completed task is saved
    3. the processor switches to the new thread, and execution continues

    Hyperthreading

    Hyper-threading (a.k.a. "symmetrical multi-threading") is different than multithreading. Whereas multithreading interleaves time slices from different programs, hyper-threading allows a single processor to perform two separate threads of computer instructions from the same computer application

    With hyper-threading, each physical processor core is assigned two virtual cores. These virtual cores are called logical cores (i.e. they mean the same thing):

    1. Each logical core is assigned a thread

    2. A single processing task is broken down into parts and the parts are assigned to threads

    3. The parts are scheduled for processing in a sequence that will keep the physical processor working as efficiently as possible

    It should be noted that the term "hyper-threading" is actually a term trademarked by Intel to describe what is better termed as "symmetrical multi-threading". Even though AMD processors do not use "hyper-threading," an Intel product, they have recently developed their own version of symmetrical multi-threading in their line of Zen processors.

  • Storage Mediums

    Primary Storage

    often referred to simply as memory, is memory that the CPU can draw from. The CPU reads instructions and data stored in primary memory as required. Primary memory is volatile, meaning that if power is turned off, everything stored in primary memory is lost. Primary storage is made up of random-access memory (RAM) and cache . The term random-access refers to the ability to access any specific location in RAM about as quickly as any other location.

    Functional Areas

    RAM

    Sticks of memory separate from the CPU. Blocks of data and instructions that require less frequent access remain in RAM until transferred to cache memory. Cache is much more expensive to manufacture than RAM, so the storage capacity of cache on the CPU is much smaller than the storage capacity of RAM. For example, an i5 CPU has 6MB of cache. But it is common for a system with this CPU to have 4-8GB of RAM.

    Cache

    Very fast memory chips located on the central processing unit. In addition to the physical cores, cache is built right into the CPU. Cache operates at much higher frequencies than RAM. Data can be read more quickly from the cache than from RAM. Cache speeds up processing by allowing blocks of data and instructions that need to be accessed frequently to be stored in locations where they can be accessed faster than they could be accessed in RAM.

    Cache is so beneficial is because CPUs Hz is much higher than RAM. In addition to a slower frequency, RAM also has latency. Latency refers to how much delay there is between when a request is made for something in RAM and when the response is available. Given the combination of slower frequencies and latency delays, a CPU typically operates about six times faster than RAM.

    If nothing could be done about the mismatch between the speed of a CPU and RAM, there would be no reason to develop faster CPUs. The CPU would sit idle five sixths of the time waiting for a transfer of content from RAM. Cache mitigates this problem by acting as a high-speed intermediate store of recent and frequently used instructions and data.

    Secondary Storage

    Secondary storage differs from primary storage in three ways:

    1. content must be loaded into primary memory first before the CPU can access it. The computer uses its input/output channels to access secondary storage and transfers the desired data into primary memory so that the CPU can access it.

    2. It is non-volatile, meaning that content is not lost if the power is turned off.

    3. it takes much longer to access content from secondary storage than primary storage.

    There are two types of secondary storage in popular use: Hard-disk drives (HDDs) and Solid-state Drives (SSDs)

    Functional Areas

    HDDs

    Data is stored on multiple magnetic platters. These platters are coated with magnetic media and are mounted on a spindle that rotates the platters. While the platters spin, an arm that contains a read/write head moves in and out across the platter to find data or to write new data. This mechanical process takes a lot of time.

    SSDs

    SSDs use a form of flash memory. Flash memory is more compact, faster, and more tolerant of shaking. Small mobile devices(such as tablets and smartphones) use SSDs instead of HDDs because they are less sensitive. They have only recently become popular in larger electronics due to their rapidly decreasing cost.

    Current notebook HDDs have sequential read and write speeds of about 60 to 90 MBs per second. Current SSDs have sequential read and write speeds of about 200 to 400 MBs per second. So SSDs are 4 to 5 times faster than HDDs.

  • How Information Is Stored

    All content processed or stored on a computer, including textual characters, numbers, and computer instructions, are stored as sequences of zeroes and ones. The figure below shows some basic units of data used by computers. A bit is the smallest discrete unit of storage that can be stored. Each bit can hold only two states: 0 (zero) or 1 (one). Eight bits make up one byte. Each byte can represent 2 8 = 256 unique combinations of zeroes and ones. In addition, each computer has a computer word size. In this context, word size refers to the number of bits a computer’s CPU can process at one time. So computers with larger word sizes have a processing advantage. Today’s computers typically use word sizes of 32 or 64 bits. A 32-bit processor uses a 32-bit (four byte) word size and a 64-bit processor uses a 64-bit (eight byte) word size.

    Functional Areas

    Inside the computer, each character is represented by a unique sequence of eight zeroes and ones, stored as a byte. For example, the capital characters of “I” and “S” are represented by [01001001] and [01010011], respectively.

    A computer with a 32-bit word size would have to process two computer words in sequence because four bytes represents 32 bits. Conversely, a 64-bit computer could process all of these characters at once because 8 bytes represents 64 bits, a single computer word.

    An ERP system is a set of integrated programs that manage a company’s vital business operations for an entire organization across each functional area—even for a complex, multisite, global organization. This is the system that fills in all of the spaces in the triangle figures above. When you hear ERP, think, the “system that does everything.”

    64-bit computers also have an advantage over 32-bit computers when processing numbers because larger computer word sizes mean that larger numbers can be processed at once.

    For example, the computer word size indicates the largest integer a computer can process in a single instruction. The word size also determines how much memory the processor can address. A 32-bit processor is limited to 232 ≈ 4 billion memory addresses, each of which holds one byte. Hence, 32-bit PCs and Macs are limited to a maximum of 4 gigabytes of RAM (see table below).

    Theoretically, a 64-bit computer can access 264 ≈ 16.8 terabytes of memory, but computer manufacturers limit the amount of RAM to 128 gigabytes(though few computers ever have that much).

    In Summary, a 64-bit computer has significant advantages over a 32-bit computer because it is possible to use more than four gigabytes of RAM. Some applications like databases, data mining applications, and sophisticated video games perform much better with more than 4 gigabytes of RAM. The table below offers a guide to understanding and comparing the different levels of storage capacity.

    Functional Areas
  • The Motherboard

    A computer’s motherboard is the board in a computer that everything else connects to. The motherboard contains an internal bus, which serves as the communication highway within the computer. The CPU and RAM snap into the motherboard. The bus transfers information between the CPU, RAM, and secondary storage devices. Other input and output devices like video cards, computer mice, and keyboards plug into the motherboard.

    CPU cache is a small amount of blazing fast memory that is used only by and for the processor. More cache is better for handling larger processing loads. Furthermore, some CPUs use hyper-threading and some do not. With so many different options to choose from, it can be difficult to know how much performance improvement you will get by paying more for an incrementally better CPU. The table below provides a comparison of a few CPUs with different characteristics. The tables below includes processors with various attributes along with their Passmark CPU performance benchmark, which is a general "overall" score of how well the processor performs across a variety of applications. The next image summarizes the same processors, but sorted based on their 3DMark Physics score--a text for video gaming and video processing.

  • Computer Buying Guide

    The useful life of a computer typically averages about 3-4 years. So you will likely choose many computers during your lifetime. This chapter is designed to help you understand the fundamental factors that determine computer performance so that you can be a knowledgeable consumer. Rather than focus on a specific brand and model, we focus on principles that do not change. That way, the concepts will be useful to you for many years.

    CPU cache is a small amount of blazing fast memory that is used only by and for the processor. More cache is better for handling larger processing loads. Furthermore, some CPUs use hyper-threading and some do not. With so many different options to choose from, it can be difficult to know how much performance improvement you will get by paying more for an incrementally better CPU. The table below provides a comparison of a few CPUs with different characteristics. The table below includes processors with various attributes along with their Passmark CPU performance benchmark, which is a general "overall" score of how well the processor performs across a variety of applications.

    Functional Areas

    The next image summarizes the same processors, but sorted based on their 3DMark Physics score--a text for video gaming and video processing.

    Functional Areas

    Notice that the "best" processor is based on a combination of speed, cache size, cores, hyperthreading and more. Additionall, processors vary in terms of the amount of power required (TDP in the table). More power means that the processor will run hotter at max usage. Heat slows down the processor well below the theoretical limits. Notice that in the 3DMark benchmark table, the i7-6850K outperforms the i7-6900K and costs half the price. Similarly, notice that the i3-6100 outperforms the i3-4170 even though they have the same speed, cores, and cache. Why? Perhaps because the i3-6100 requires less power and, therefore, generates less heat to slow down processing.

    In summary, you don't always get what you pay for. If you are willing to put in the effor to research processors, you can likely spend far less money by examining benchmark results and selecting the perofrmance for the money for your particular type of usage.