INTRODUCTION
We are now living in the 21st century where most of our daily tasks are being carried out using digital electronic technology devices that assists us in delivering a high quality and accurate products within a shortest possible time. These digital electronic devices are called computer.
The Computing machine is any electronic device that can accept data as an input through the input medium such as keyboard, touch screen, or sensitive interface, process the input data by some components within the machine, and present the result of the processed data to the user or other devices through the output medium such as screen, printer, speaker, or software application.
The computer is made up of two basic constituents; the hardware components and the software. The physical parts of a computer that we can see, touch, or handle are called the computer hardware. The software are the set of instructions stored inside the computing machine that controls all the hardware components, and its also control some other installed software applications in the machine.
To fully understand exactly how computer operate, we have to study the different parts that made up the computer. Generally, computer hardware are classified into seven 7 different groups:
- Input devices
- Main memory devices
- Storage devices
- Processing devices
- Output devices
- Power management, and
- Computer casing
Since we are now only interested on knowing how our computer operate with data under this article, we will concentrate on the main parts that must be present in every computer before it can function. These main components are what made up a computer machine to be called computer. And these include the computer Main memory – Random Access Memory (computer heart), and Central Processing Unit (computer brain) components. These are the critical components of every digital electronic device that process and stores data or information in the device.
The computer random access memory stores the data and information that it’s going to be process by the processor at the moment. The central processing unit access the data directly from the random access memory, and execute the data. Let’s dive in and see how these mysterious devices hold and manipulates on the data we supplied to them.
RANDOM ACCESS MEMORY (RAM)
RAM stands for random access memory. It’s one of the most fundamental elements of computing. RAM is a temporary memory bank where the computer stores data it needs to retrieve quickly. RAM keeps data that is easily accessible to the processor so it can quickly find it without having to go into long-term storage such as hard disk drive or flash drive, to complete the immediate processing tasks.
RAM is present in all the computing devices, whether it’s a desktop computer, a tablet/smartphone, or even smart electronic devices (smart watch) and IoT computing devices. Nearly, all computers have a way of storing information for longer-term access, too. But the memory needed to run the application’s processes you’re currently working on is stored and accessed in the computer’s RAM.
Function of RAM in Device
RAM is a form of temporary storage that gets wiped off when the computer is turned off, or when the power from the computer goes down. RAM offers lightning-fast data access, which makes it ideal for the undergoing processes, apps, and programs the computer is actively working on, such as the data needed to stream the online video through web browser.
To understand how RAM operate, we have to first understand how it manipulates the ins and outs flow of electricity that represent all the data/information inside the computer system.
Putting Data to RAM
Software, in combination with the operating system, sends a burst of electricity along an address line. An address line is a microscopic strand of electrically conductive material etched onto a RAM chip – a very tiny integrated circuit. Each address line identifies the location of a spot in the chip where data can be stored. The burst of electricity identifies where to record data among the many address lines in a RAM chip.
The electrical pulse turns on (closes) a transistor that’s connected to a data line at each memory location in a RAM chip where data can be stored. A transistor is essentially a microscopic electronic switch.
NPN Transistor
While the transistors are turned on, the software sends bursts of electricity along the selected data lines. Each burst represents a 1 bit. The 1-bit and the 0-bit make up the native language of processors, commonly called the machine language. Bit is the most basic unit of information that a computer manipulates.
When the electrical pulse reaches an address line where a transistor has been turned on, the pulse flows through the closed transistor and charges a capacitor. Capacitor is an electronic device that stores electricity. This process repeats itself continuously to refresh the capacitor’s charge, which would otherwise leak out. When the computer’s power is turned off, all the capacitors lose their charges. Each charged capacitor along the address line represents a 1 bit. An uncharged capacitor represents a 0 bit. The computer uses 1 and 0 bits as binary numbers to store and manipulate all information, including the texts, graphics, and even the sound.
GETTING DATA FROM RAM
When software wants to get data stored in RAM, another electrical pulse is sent along the address line, once again closing the transistors connected to it. Everywhere along the address line that there is a capacitor holding a charge, the capacitor will be discharge through the circuit created by the closed transistors, sending electrical pulses along the data lines.
The software recognizes from which data lines the pulses come and interprets each pulse as a 1. Any line on which a pulse is lacking, indicates a 0. The combination of 1s and 0s from eight data lines forms a single byte of data. A byte can be used to hold the information of a single character or symbol, such as the letter X.
DATA MOVEMENT FROM RAM CHIPS TO CPU AND VICE-VERSA
No matter how fast a processor can be, it’s limited by how fast memory feeds them data. Traditionally, the way to pump out more data to CPU was to increase its clock speed. The clock speed measures the number of cycles CPU executes per second. With each cycle, or tick, of the clock regulating operations in the processor and movement of memory data, synchronous dynamic random access memory (SDRAM) memory could store a value or move a value out and onto the data bus headed to the processor. But the increasing speeds of processors outstripped that of random access memory (RAM). Random Access Memory design narrowed the gap in two ways. One is double data rate (DDR). Previously, a bit was put or gotten on each cycle of the clock. It’s as if someone loaded cargo (putting data) onto a train traveling from X-city to Y-city, unloaded that cargo (getting data), and then had to send the empty train back to X-city again, despite having fresh cargo in Y-city that could hitch along for the return trip. With DDR, a handler could unload that same cargo when the train arrives in Y-city and then load it up with new cargo before the train makes its journey back to X-city. This way, the train is handling twice as much traffic (data) in the same amount of time. Substitute memory controller for the persons loading and unloading cargo and clock cycle for each round-trip of the train, and you have DDR. Another change to RAM, DDR2, which doubled the data rate in a different way. It cut the speed of memory’s internal clock to half the speed of the data bus. DDR2 quickly evolved into DDR3, and then to DDR4, each halving the clock rate of its predecessor. A significant impact to cutting memory speed is that the RAM uses less electricity, and it also pays off with cooler running, more reliable memory chips.
Example of personal computer random access memory (RAM).
Image credit: By An-d - Own work, CC BY-SA 3.0, https://commons.wikimedia.org/w/index.php?curid=27224495
FLASH MEMORY
When you are using a DESKTOP or LAPTOP computer, data in RAM that’s not saved to storage drive disappears when the computer is turned off. But computers that have evolved into smartphones, tablets, smartwatches, cameras, and other handhelds don’t have hard-disk drives for data storage. These devices also use memory chips for storing information, but yet when you turn off the smartphone, all the contacts, music, pictures, videos, and apps would still be there when the phone is turn on back again. All the recent computers that evolved latest technology don’t used same technology architecture as in ordinary RAM for the Desktop/Laptop computers. Instead, they used flash memory technology that freezes data in place.
Flash memory is a data-storage medium used with computers that have evolved into smartphones, cameras, and other electronic devices. Unlike previous forms of data storage, flash memory is an EEPROM (electronically erasable programmable read-only memory) form of computer memory and thus does not require a power source to retain the data.
HOW FLASH MEMORY HANDLES and MANAGES DATA
Flash memory is normally laid out along a grid of printed circuits running at right angles to each other. In one direction, the circuit traces are word addresses; circuits at a right angle to them represent the bit addresses. The two addresses combine to create a unique number address called a cell.
The cell contains two transistors that together determine if an intersection represents a 0 or a 1. One of the transistors called the control gate is linked to one of the passing circuits called the word line, which determines the word address.
A thin layer of metal oxide separates the control gate from the second transistor, called the floating gate. When an electrical charge runs from the source to the drain, the charge extends through the floating gate, on through the metal oxide, and through the control gate to the word line.
A bit sensor on the word line compares the strength of the charge in the control gate to the strength of the charge on the floating gate. If the control voltage is at least half of the floating gate charge, the gate is said to be open, and the cell represents a 1. Flash memory is sold with all cells open. Recording to it consists of changing the appropriate cells to zeros.
Flash memory uses Fowler Nordheim tunneling to change the value of the cell to a zero. In tunneling, a current from the bit line travels through the floating gate transistor, exiting through the source to a ground.
Energy from the current causes electrons to boil off the floating gate and through the metal oxide, where the electrons lose too much energy to make it back through the oxide. The electrons are trapped in the control gate, even when the current is turned off.
The electrons have become a wall that repels any charge coming from the floating gate. The bit sensor detects the difference in charges on the two transistors, and because the charge on the control gate is below 50 percent of the floating gate charge, it is considered to stand for a zero. When it comes time to reuse the flash memory, a current is sent through in-circuit wiring to apply a strong electrical field to the entire chip, or to predetermined sections of the chip called blocks. The field energizes the electrons, so they are once again evenly dispersed.
Flash memory comes in a variety of configurations, including SmartMedia, Compact Flash, and Memory Sticks (used mainly in cameras and other smart electronic devices). There’s also USB-based flash drives (also called thumb drives and flash drives), memory cards use flash memory technology to store and move data around. The form factors, storage capacities, and read/write speeds vary. Some include their own controllers for faster reads and writes.
THE CENTRAL PROCESSING UNIT
The Central Processing Unit (CPU), also called a processing unit, microprocessor, or a processor, is an integrated circuit. An integrated circuit that integrates billions of tiny electrical parts, arranging them into circuits and fitting them all into a small compact box.
The retrieval and execution of data and instructions for a computer is conducted inside the CPU. The processor is essentially the brain of a computing machine that processes and control everything of the computer, be it a software or hardware. The components inside the processor are classified into three groups; it consists of an arithmetic and logic unit (ALU), a control unit, and various registers.
The ALU performs arithmetic operations, logic operations, and related operations, according to the program set instructions. The control unit controls all the CPU operations, including the ALU operations, the movement of data within the CPU, and the exchange of data and control signals across external interfaces, through system bus. Registers, also called Accumulators are high-speed internal memory-storage units located within the CPU.
Simplified diagram of a microprocessor. Black lines indicate data flow, whereas red lines indicate control flow; arrows indicate flow directions.
DATA PROCESSING OPERATION INSIDE MICROPROCESSOR
The central processing unit interact with program instruction and other parts of a computer by using a specific language, similar to how we human use to communicate using a particular language among ourselves. The only language CPU understand is a binary language - language that deals with 1s and 0s only. Any type of data in computer, be it a texts, images, or sounds must be converted to 1s and 0s by the control unit before the processor will be able to execute it.
With this, we’re now ready to jump inside the central processing unit device and look out how it handles and executes data, and how it also controls other hardware parts and software inside the computer. But let start by observing how the data and information are been handles inside register-storage device located in the CPU.
Very few among us can do complex math operations in our heads. Even for something as simple as adding rows of numbers, we need a pencil and paper to keep track of which numbers go where. Microprocessors are not all that different in this regard. Although, they are capable of performing intricate math involving thousands or millions of numbers, processors, too, need notepads to keep track of their calculations. This CPU notepads are what we called registers, and their pencils are pulses of electricity.
The register of processor consists of several parts, and each part has a specific role. One of these parts of register is a reserved section for transistors in the faster memory inside the microprocessor.
Inside the CPU, there is the processor’s arithmetic logic unit (ALU), in charge of carrying out math instructions, and the control unit, which coordinates instructions and data through the processor, have quick access to the registers. The size of the registers determines how much data the processor can work with at one time. Most personal computers have registers with 32 or 64 bits for data.
The processor’s control unit directs the fetching and execution of program instructions. It uses an electrical signal to fetch each instruction, decodes it – translates the instruction into machine language i.e. 0s and 1s, and sends another control signal to the arithmetic logic unit telling the ALU what operation to carry out.
With each clock cycle - the small unit of time during which the different components of a computer can do one thing, mostly in fraction of a second, the processor writes or reads the values of the bits by sending or withholding a pulse of electricity to each. Each chunk of binary numbers is only that. They have no labels to identify them as instructions, data, values going into a computation, or the product of executing instructions. What the values represent depends on which registers the control unit uses to store them.
Address registers collect the contents of different addresses in RAM or in the processor’s on-board cache, where they have been prefetched in anticipation that they would be needed.
When the processor gets the contents of a location in memory, it tells the data bus to place those values into a memory data register. When the processor wants to put values of data to memory, it places the values in the memory data register, where the bus retrieves them to transfer to RAM.
A program counter register holds the memory address of the next value the processor will fetch. As soon as a value is retrieved, the processor increments the program counter’s contents by 1, so it points to the next program data location.
The processor puts the results of executing an operation into several accumulation registers, where they await the results of other executing operations, before assembling all the results and put it into yet another accumulator.
HANDLING MATH OPERATION INSIDE PROCESSOR
As we have already know, all the information inside processor - words and graphics as well as numbers and sounds - is stored in and manipulated by a processor in the form of binary numbers. In this binary numerical system, there exist only two digits – 0s and 1s. All numbers, words, and graphics are formed from different combinations of these digits.
| DECIMAL NUMBER | BINARY NUMBER |
|---|---|
| 0 | 0 |
| 1 | 1 |
| 2 | 10 |
| 3 | 11 |
| 4 | 100 |
| 5 | 101 |
| 6 | 110 |
| 7 | 111 |
| 8 | 1000 |
| 9 | 1001 |
| 10 | 1010 |
Equivalence of Decimal number in Binary number system
Transistor switches are used to manipulate binary numbers since there are only two possible states of a switch, it is either in open (off) state or in closed (on) state, which nicely matches the two binary digits. An open transistor, through which no current is flowing, represents a 0. A closed transistor, which allows a pulse of electricity regulated by the PC’s clock to pass through, represents a 1. The computer’s clock regulates how fast the computer works. The faster a clock ticks, causing pulses of electricity, the faster the computer works. Clock speeds are measured in megahertz, or millions of ticks per second. Current passing through one transistor can be used to control another transistor, in effect turning the switch on and off to change what the second transistor represents. Such an arrangement is called a gate because, like a fence gate, the transistor can be open or closed, allowing or stopping current flowing through it.
Transistor: Open (OFF)
Transistor: Closed (ON)
The simplest operation that can be performed with a transistor is called a NOT logic gate, made up of only a single transistor. This NOT gate is designed to take one input from the clock and one from another transistor. The NOT gate produces a single output - one that’s always the opposite of the input from the transistor. When current from another transistor representing a 1 is sent to a NOT gate, the gate’s own transistor switches open so that a pulse, or current, from the clock can’t flow through it, which makes the NOT gate’s output 0. A 0 input closes the NOT gate’s transistor so that the clock pulse passes through it to produce the opposite of the input, as its output which is 1.
| INPUT FROM OTHER TRANSISTOR | NOT LOGIC GATE OUTPUT |
|---|---|
| 1 | 0 |
| 0 | 1 |
NOT logic gate
NOT gates strung together in different combinations to create other logic gates, all of which have a line to receive pulses from the clock and two other input lines for pulses from other logic gates. With different combinations of logic gates, a computer performs the math that is the foundation of all its operations. This is accomplished with gate designs called half-adders and full–adders. A half-adder consists of an XOR gate and an AND gate, both of which receive the same input representing a one-digit binary number. A full-adder consists of half-adders and other switches.
XOR logic gate
AND logic gate
A combination of a half-adder and a full-adder handles larger binary numbers and can generate results that involve carrying over numbers. To add the decimal numbers 6 and 3 (110 and 11 in the binary system), first the half-adder processes the digits on the right side through both XOR and AND gates.
MOVEMENT OF DATA IN THE CPU
There exist more than 100 billion transistors inside the microprocessor of a computer. Taking a walk through one of the transistors could easily get a person hopelessly lost. Old or new, however, how a processor performs its most basic functions hasn’t changed. Microprocessor could have as many several processor chips bundled in it, called multi-cores, and multiple caches but, like old Pentium III processor, they all face the same problem of how to move data quickly and without any hitch (Ron White, 2015).
A processor and its integrated cache share the same interface to the computer’s information. Program code or data manipulated by that program code move in and out of the processor chip at the computer’s maximum bus speed. Much of a computer’s architecture is structured to alleviate the bus bottleneck by minimizing the times a clock cycle - the smallest time in which a computer can do anything - ticks away without the processor completing an operation.
When information enters the processor through the bus interface unit (BIU), the BIU duplicates the information and sends one copy to the CPU’s closest data caches that are housed directly within the processor core. The BIU sends program code to the Level 1 instruction cache, or I-cache, and sends data to be used by the code to another part of L1 cache called, the data cache (D-cache).
While the fetch/decode unit – the unit responsible for getting data and instructions from RAM or from cache memory, is pulling in instructions from the I-cache, the branch target buffer (BTB) compares each instruction with a record in a separate set-aside buffer to see whether any instruction has been used before. The BTB is looking in particular for instructions that involve branching, a situation in which the program’s execution could follow one of two paths. If the BTB finds a branch instruction, it predicts, based on past experience, which path the program will take.
As the fetch/decode unit pulls instructions in the order predicted by the BTB, three decoders working in parallel break up the more complex instructions into mops, which are smaller micro-operations. Then, another unit called the dispatch/execution unit processes several mops faster than it processes a single higher-level instruction.
The decode unit sends all mops to the instruction pool, also called the reorder buffer. This contains two arithmetic logic units (ALUs) that handle all calculations involving integers. The ALUs use a circular buffer, with a head and tail, that contains the mops in the order in which the BTB predicted they would be needed.
The dispatch/execute unit checks each mop in the buffer to see whether it has all the information needed to process it, and when it finds a mop ready to process, the unit executes it, stores the result in the micro-op itself, and marks it as done. If a mop needs data from memory, the execute unit skips it, and the processor looks for the information first in the nearby L1 cache. If the data isn’t there, the processor checks the next cache level, L2 in this case.
Instead of sitting idle while that information is fetched, the execute unit continues inspecting each mop in the buffer for those it can execute. This is called speculative execution because the order of mops in the circular buffer is based on the BTB’s branch predictions. The unit executes up to five mops simultaneously. When the execution unit reaches the end of the buffer, it starts at the head again, re-checking all the mops to see whether, any have finally received the data they need to be executed.
If an operation involves decimal-point numbers, such as 7.34 or .23333, the ALUs hand off the job to the floating-point math unit, which contains processing tools designed to manipulate decimal-point numbers quickly.
When a mop that had been delayed is finally processed, the execute unit compares the results with those predicted by the BTB. Where the prediction fails, a component called the jump execution unit (JEU) moves the end marker from the last mop in line to the mop that was predicted incorrectly. This signals that all mops behind the end marker should be ignored and can be overwritten by new mops. The BTB is told that its prediction was incorrect, and that information becomes part of its future predictions.
Meanwhile, the retirement unit is also inspecting the circular buffer. It first checks to see whether the mop at the head of the buffer has been executed. If it hasn’t, the retirement unit keeps checking it until it has been processed. Then, the retirement unit checks the second and third mops. If they’re already executed, the unit sends all three results - its maximum, to the store buffer. There, the prediction unit checks them out one last time before they’re sent to their proper place in system RAM.
Multi-Core Processors and How it Works
There are billion or so transistors in any recently produced microprocessors, we may think that these number of transistors would more than satisfy the requirement of the most intense software operations we can push through the processor chips. But in reality, the word enough doesn’t exist in computing world. So, if it’s too hard to put more transistors on the processors, there’s another solution: Put more processors as single device in the computer, the term called multi-core processors. Multi-core processors are like putting a couple of computers together into a single unit and having them share the same memory, power, and input/output devices. Machines with between two and more processor cores are standard issue, and that number will only get bigger. (Ron White, 2015)
Let’s use a quad-core processor as an example, the one that has four executable processor cores all assembled onto a single chip, for the explanation that follows on how processor works.
Specific designs vary, but a typical quad-core processor combines four execution cores onto a single die, or silicon chip. Other designs spread their cores across two dies. Regardless, these identical cores are the heart of any microprocessor and the part that does the heavy work of executing instructions from software.
To gain the speed and other advantages possible with a multi-processor computer, the operating system (OS) running on it must be designed to recognize that the computing machine has multi-core processors, be able to distinguish them, and know how to handle operations such as hyperthreading. Similarly, software applications, games, and utilities need to be rewritten to use the multiple cores. Such software is referred to as threaded, or multi-threaded. The software app adding, say, a column of four-place numbers could divide the job into four threads: Adding the 1s-place numbers, the 2nd-place numbers, the 3rd-place numbers, and the 4th-place numbers. Each of those subtasks is directed to a different core.
When the subtasks exit the cores, the operating system combines the threads into a single number, and sends that operation to one of the cores for execution.
If the application software isn’t equipped to work in multiple cores, the operating system can still take advantage of them. It picks one of the cores to run the software and creates an affinity between that core and the program. It then creates affinities between the remaining cores and various tasks. A second core may handle background operations, such as disk optimizing; a third core might supervise a download; and the fourth could render a video that’s streaming from the Internet. Neither the operations nor their finish times are affected by the processing going on in the other cores.
The OS puts that operation into a time-staggered queue along with requests that are going to other cores. Each of the operations enters its respective core on different clicks of the computer’s clock so they are less likely to run into each other or cause a traffic jam in the areas they have mutual access to. Each processor core isn’t completely distinct. They might share access to resources like an on-die graphics processor, memory caches, and more. The operating system can determine how each core shares access to these resources. If only one core is active, for example, the OS dynamically allocates more of the shared cache to that core.
Random access memory (RAM) and the central processor unit are very crucial components in every computing machine, in which without them there will be no data manipulation and execution could have ever happened. Knowing how these devices manipulates information increases our awareness on how computer or any digital electronics handles and executes data, which have a great impact to us in selecting and using the appropriate computing machines for our tasks.
The recent operating systems in our computing machine takes the advantages of multi-core processors in handling and executing tasks within a very short period of time – about a fraction of second, by dividing these tasks into groups and handling them to the available cores within the processor chip for execution.
Kernel, the core service that all the operating systems built on, is responsible for controlling how the OS interact with all the number of cores present inside the CPU chip, and also how the data is been accessed from the RAM. If you want to understand more details on the main layer between operating system and the underlying computer hardware, I advise you to read a detail texts on the operating system kernel.