Bolivarian Republic of Venezuela

Ministry of People's Power for Defense

National Experimental Polytechnic University of the Armed Forces
UNEFA

Computer Architecture

Prof.: Bachelor:
Eng. Maricela Materano Javier Vilchez
CI: 27,393,998

Puerto Cabello, August 2020

 Computer Architecture
Computer architecture is the conceptual design and fundamental operational structure of a
system that makes up a computer. That is, it is a model and functional description of the
design requirements and implementations for various parts of a computer, with special
interest in how the central processing unit (CPU) works internally and accesses memory
addresses.

What is a computer?
A computer is a complex synchronous sequential system that processes information, this
is binary information, using only the logical value digits '1' and '0'. These binary logic
values correspond to electrical voltage values, such that a logic '1' corresponds to a high
level of 5 volts and a logic '0' corresponds to a low voltage level close to 0 volts; these
voltages depend on the technology used by the computer devices.

Computer Architecture according to Von Neumann

Von Neumann architecture has its origins in the work of mathematician John Von
Neumann developed with John Mauchly and John P. Eckert and disclosed in 1945 at the
Moore School of the University of Pennsylvania, United States, where the EDVAC
(Electronic Discrete Variable Automatic Computer) was presented. From here arose the
architecture of the program stored in memory and sequential search/execution of
instructions. Generally speaking, a computer has to perform 3 functions:
 Data Processing
 Data Storage
 Data transfer
Just as a PC (Personal Computer) must process data, transforming the information
received, it must also store data as the final result of these. You must also perform data
transfer between your environment and it. The architecture of a computer refers to the
organization of its elements into modules with a defined functionality and the iteration
between them.
 Basic structure of a computer

 CPU (central processing unit): The central processing unit is the heart of the
computer. It controls the flow of data, processes it, and governs the sequencing
of actions throughout the system. To do this, it needs an external oscillator or
clock that synchronizes operations and marks the processing speed, this marks
the evolution of the CPU and measures its operating speed; unfortunately, the
CPU clock frequency is limited by the technology of the CPU and the entire
computer, depending on the peripherals, its graphic cards, memories, etc.
Therefore, excessive use of the computer's resources can result in overheating
that partially or totally damages the CPU.

 Memory: It is responsible for data storage.

 Input/Output: Transfers data between the outside environment and the

computer. It contains the peripheral controllers that form the interface between
the peripherals, memory and the processor.

 Interconnection system: Buses; it is the mechanism that allows the flow of data
between the CPU, the memory and the input/output modules. Here the electrical
signals are propagated and interpreted as logical ones and zeros.

 Peripherals: These devices are those that allow data to enter the computer, and
the information to be output once it has been processed. A group of peripherals
can be understood as a set of transducers between external physical information
and binary information interpretable by the computer. Examples of these devices
are the keyboard, monitor, mouse, hard drive, and network cards.

Central Processing Unit

Controls the operation of the elements of a computer. Since the system is powered by a
current, it does not stop processing information until the power supply is cut off. The CPU
is the most important part of the processor, because it is used to perform all the
computer's operations and calculations. The CPU has another internal structure that is
shown in the following image.

 Control Unit (UC): The control unit is responsible for reading the instructions to be
executed from memory and sequencing access to data and operations to be
performed by the processing unit. The UC generates the control signals that
establish the flow of data throughout the computer and internally in the CPU. An
instruction is nothing more than a combination of ones and zeros. It consists of a
binary operation code to execute the instruction, the UC stores it in a special
register, interprets its operation code and executes the appropriate sequence of
actions, in short, it decodes the instruction.

 Arithmetic Logic Unit or ALU: It is the part of the CPU responsible for performing
data transformations. Governed by the UC, the ALU consists of a series of
modules that perform arithmetic and logical operations. The UC is responsible for
selecting the operation to be performed by enabling the data paths between the
various ALU operators and between the internal registers.

The arithmetic-logic unit consists of the following elements:

 An operator with 4 control signals for selecting the operation to be performed

 2 multiplexers with 4 inputs and 1 output, which select between the possible
operands
 An accumulator register that allows storing intermediate results
 The output of the accumulator register can be transferred to the X multiplexer,
the data bus or the address bus, depending on the control signals.
  Internal Registers: Storing the results of the execution of instructions in main
memory could be slow and would have excessively large amounts of data in the
interconnection system with memory, which would decrease performance. In the
same way, the internal configuration of the CPU or the information during the last
operation of the ALU are also stored in internal registers. The main registers of a
CPU are:

1. Program counter: responsible for storing the address of the next

instruction to be executed.
2. Instruction Register: stores the instruction captured in memory and the
one being executed.
3. Status Register: composed of a series of bits that report the result
obtained in the last ALU operation.
4. Accumulator Register: some CPUs perform arithmetic operations in a
register called an accumulator, its function is to store the results of
arithmetic and logical operations.

The cycle to execute any instruction is divided into fetch cycle and instruction cycle. The
former causes the CPU to generate appropriate signals to access the memory and read
the instruction; the latter is similar; the difference between the two is the opcode of each
instruction.

Current complexity
The complexity of CPU design increased as various technologies made it easier to build
smaller, more reliable electronic devices. The first of these improvements came with the
advent of the transistor. With this improvement, more complex and more reliable CPUs
were built on one or more printed circuit boards containing discrete (individual)
components.

Microcontrollers
Microcontroller is an integrated circuit that is the main component of an embedded
application. It is like a small computer that includes systems to control input/output
elements. It also includes a processor and of course memory that can save the program
and its variables (flash and RAM). It works like a mini PC. Its function is to automate
processes and process information.
The microcontroller is used in all kinds of inventions and products where it is required to
follow an automatic process depending on the conditions of different inputs.

ELEMENTS OF A MICROCONTROLLER
 A microcontroller will at least have:
 Microprocessor.
 Peripherals (input/output units).
 Memory.

INTERNAL ELEMENTS IN A MICROCONTROLLER

Processor or Microprocessor.
A processor includes at least three elements, ALU, control unit and registers.
ALU. Also known as Arithmetic and Logic Unit. This unit is composed of digital electronic
circuits of the combinatorial type (gates, adders, multipliers), whose main function is to
perform operations. These operations are divided into three types:
Logical. Such as the basic operations of logic gates, such as logical addition (OR), logical
multiplication (AND), logical difference (XOR), and negation (NOT). A logical operation can
only have as inputs and as outputs a logical response (0 or 1). This depends on the
voltage levels of a digital signal.
Arithmetic. Arithmetic operations are addition, subtraction, multiplication and division.
Depending on the processor (8, 16, 32 or 64 bits) will be the speed with which these
operations can be performed.
Miscellaneous. All other operations such as bit transfer (<< >>) fall into these operations.
Control unit. The control unit is the set of sequential digital systems (those that have
memory) that allow the logic of the signals to be distributed.
Records. Registers are the main memories of processors, since they operate at the same
speed as the processor unlike other somewhat slower memories (such as RAM, FLASH or
CACHE). Registers are built by Flip-Flops. Flip-Flops are sequential digital circuits.
PERIPHERALS.
Peripherals are digital circuits that allow us to interact with the world "outside" the
microcontroller. Its function is to enable or disable digital outputs, read analog sensors,
communicate with digital terminals or extract analog signals from a digital conversion.
Parallel input/output ports. Ports are related to the size of the processor, meaning an 8-
bit port is because the processor is 8-bit. A 64-bit processor has the ability to have a 64-bit
port.
Serial ports. They allow us to transform parallel digital information (bytes of information)
into frames that can be transferred over one or more communication lines. There are for
example: serial port, i2c, SPI, USB, CAN, etc.
Analog peripherals. Such as those that convert analog signals to digital (ADC) or digital
signals to analog (DAC) or analog comparators.

Microprocessors
A microprocessor or simply a processor is the central integrated circuit of a computer
system, where logical and arithmetic operations (calculations) are carried out to allow the
execution of programs, from the Operating System to the application software.
A microprocessor can operate with one or more CPUs (Central Processing Units), each
consisting of registers, a control unit, an arithmetic-logic unit and a floating point
calculation unit (or mathematical coprocessor).
It is also generally connected via a socket to the motherboard or mainboard, along with a
heat dissipation system made up of certain heat dissipation materials and a fan cooler
(internal fan).
While a single microprocessor may have one or more physical or logical cores, in which all
the calculation work is carried out, a single computer system may have several processors
working in parallel.
The performance of such processors is not easy to measure, but clock speed (measured
in hertz) is often used to distinguish between the power of one and the other.

History of the microprocessor

Microprocessors emerged as a product of the technological evolution of two specific
branches: computing and semiconductors. Both had their beginnings in the mid-20th
century, in the context of the Second World War, with the invention of the transistor, which
replaced vacuum tubes.
From then on, silicon was used to generate simple electronic circuits, subsequently giving
rise (in the early 1960s) to the creation of the first digital circuits: Transistor-Resistor Logic
(RTL), Diode Transistor Logic (DTL), Transistor-Transistor Logic (TTL) and Emitter
Complemented Logic (ECL).
The next step towards microprocessors would be the invention of integrated circuits (SSI
and MSI), thus allowing the beginning of the aggregation and miniaturization of
components. The first calculators to use this technology, however, required between 75
and 100 integrated circuits, which was impractical. And so the next step in the reduction of
computing architecture was the development of the first microprocessors.
The first processor was the Intel 4004 manufactured in 1971. It contained 2300 transistors
and with its mere 4 bits of capacity it could perform 60,000 logical operations per second,
at a clock frequency of 700 Hz. From then on, the technological race invested in the
development of better and more powerful microchips: 8-bit, 16-bit, 32-bit and 64-bit,
currently reaching frequencies above 3 GHz.

Microprocessor Features
Cache prevents RAM from being used unnecessarily.
Microprocessors resemble a small miniature digital computer, so they have their own
architecture and perform operations under a control program. This architecture is
composed of:
 Encapsulation. A ceramic covering that encases the silicon and protects it from
the elements (such as oxygen in the air).
 Cache: A type of ultra-fast memory available to the processor, so that it does not
use RAM except when necessary, since data in use is stored in the various levels
of the cache memory for immediate recovery.
 Mathematical coprocessor: Called the floating point unit, it is the portion of the
processor that is responsible for logical and formal operations.
 Registers: A short working memory in the processor, designed to keep track of its
own operation and conditions.
 Ports: The conduits that allow the processor to communicate information with the
rest of the system components.
What is a microprocessor used for?
Microprocessors are the “brain” of the computer: its logical center for arithmetic and logical
operations, where all the system programs are executed, both those of the Operating
System itself and the applications executed by the user. The binary logic of the system
and memory accesses also occur there. That is to say: the processor is the information
engine of the computer.

Microprocessor function
The fetch is the sending of the specific instruction to the decoder.
A microprocessor operates based on a series of elementary instructions that are
preprogrammed and stored in the form of binary code. These instructions are going to be
organized into main memory, and are given according to several phases, which are:
 Prefetch: Pre-reading of the instruction from the system's main memory.
 Fetch: Sending the specific instruction to the decoder.
 Decoding: Translation of the instruction into a series of operations to be
performed, and reading of the operands necessary to do so.
 Execution: Carrying out the instruction by the system components.
 Writing: Recording the results back into main memory, or into registers.
These phases are carried out in several CPU cycles, and their duration depends on the
frequency at which the microprocessor operates.

Processor
The processor is the brain of the system, it processes everything that happens on the PC
and executes all the actions that exist. The faster the processor a computer has, the faster
the commands given to the machine will be executed. This component is part of the
hardware of many devices, not just your computer.
The processor is a silicon chip that is placed in the socket on the motherboard inside the
desktop computer case, the difference in a laptop is that it is directly soldered. The
processor is covered in something we call encapsulation, and of which there are 3 types:
PGA, LGA and BGA.
The processor is one of the computer components that has evolved the most, as
engineers are required to offer increasingly better processors so that computers run faster
and more efficiently. Its evolution has not only been internal, but its external form has also
been modified. The most popular PC processor manufacturers are Intel and AMD.
This component is the most important one, we could say, and generally the most
expensive, but without the rest of the components it could not serve or act.
Components of a processor
A processor is composed of:
Cores
Cache
Memory controller
Graphic card
Other auxiliary elements
How a processor works
The processor makes sure everything runs as programmed.
The operation of the processor is determined by a clock that synchronizes all the
functional blocks and ensures that everything works as it should or is programmed to be.
The operation has stages:
Read the memory instruction
Search for data
Perform the operation
Go to the next instruction

Speed
Operating speed (Clock Speed)
The operating speed, clock speed or clock speed is also known as frequency, since it is
measured in Hertz. This parameter represents the number of instructions per second that
the processor is capable of processing. For example, a processor running at 3 GHz means
it can complete up to 3,000,000,000 instructions per second on each of its cores. Thus, if a
job requires, say, 1,000,000,000 instructions, the processor would be able to complete it in
a third of a second, while one running at 1 GHz would do so in one second.
That's the theory, but given the differences in the architecture of different processors, two
processors with the same clock speed will not necessarily complete the task in the same
amount of time, which is why Intel and AMD processors have different performance even
though they have the same number of cores and the same speed.
Turbo Boost and Turbo Core
Intel uses Turbo Boost technology, while AMD calls it Turbo Core, but they basically do the
same thing: they are able to automatically overclock their cores to increase performance
under high workloads. In advanced versions of this technology, this overclock could be
performed on only one of the cores, but at a higher speed than when it is done on all of
them.
Bus type
A CPU's bus type is the way its cores communicate with the rest of the system. For the
average user this does not have much of an effect on the speed or performance of the
system, but on more modern computers it is already noticeable, and the new generation
buses are much more efficient and faster than the previous ones. At the moment, QPI
(Quick Path Interconnect) is the most common on Intel, while AMD's is called
Hypertransport, but both are moving towards Infiniband and Infinity Fabric.

TDP – Thermal Output

Thermal Output, also called TDP or Thermal Design Power, is the maximum amount of
dissipation power needed to keep the processor within the optimal operating temperature
range. The higher this value, the more heat the processor will generate, and as you know,
we will need a better heatsink to keep it at good temperatures.
Smart Cache
Smart Cache is essentially L3 but optimized by Intel to be more efficient at sharing
information across CPU cores. For practical purposes, it behaves the same as the L3
cache. This parameter may not be found in all processor specifications.
Graphics processor, iGPU or graphics processor
Many processors have a GPU inside, called an integrated graphics card or iGPU. It is
generally very low-power, but enough to perform basic tasks such as browsing the
Internet, watching videos, and even for some basic games, especially in the latest
generations since they are increasingly more powerful.
However, in the latest generations of processors, more and more powerful integrated
graphics are being introduced, and they are already capable of handling multiple monitors,
4K resolutions and are even capable of running some games at a decent FPS rate.
These are, as a general rule, all the technical specifications of Intel and AMD processors.
However, if you have any questions, feel free to ask them in the comments.

Costs
Best processors of the moment
Most interesting processors by price range, By power and price, they are very interesting
and allow you to optimize the investment in the processor for gaming or design equipment.

Freq.
Threads/Nuts base/turbo iGPU Price

Athlon 200GE 2/4 3.2 GHz, no Radeon 45.99 €

turbo Vega 3

Core i3-9100F 4/4 3.6/4.2 GHz 71.49 €

Ryzen 3 3100 4/8 3.6/3.9 GHz — 112.83

€

Ryzen 5 2600 6/12 3.4/3.9 GHz — 137.20

€

Ryzen 5 2600X 6/12 3.6/4.2 GHz — 149.89

€

Core i5-10400F 6/12 2.9/4.3 GHz 162.90

€

Ryzen 5 3600 6/12 3.6/4.2 GHz — 189.00

€

Ryzen 7 2700X 8/16 3.7/4.3 GHz — 200.00

€

Core i7-10700K 8/16 3.8/5.1 GHz UHD 407.00

Graphics €
 Freq.
Threads/Nuts base/turbo iGPU Price

630

Ryzen 7 3700X 8/16 3.6/4.4 GHz — 286.16

€

Ryzen 7 3800X 8/16 3.9/4.5 GHz — 346.88

€

Core i9-10900K 10/20 3.7/5.3 GHz UHD 609.90

Graphics €
630

Ryzen 9 3900X 12/24 3.8/4.6 GHz — 467.82

€

Ryzen 9 3950X 16/32 3.5/4.7 GHz — 718.34

€

Ryzen Thread- 32/64 3.7/4.5 GHz — 2099.00

ripper 3970X €
Internal memory
Primary memory (PM), main memory, central memory or internal memory is the computer
memory where both the data and the programs that the central processing unit (CPU) is
processing or is going to process at a given time are temporarily stored. Due to its
function, the MP must be inseparable from the microprocessor or CPU, with which it
communicates through the data bus and the address bus. The
https://es.wikipedia.org/wiki/Bus_(inform%C3%A1tica)bus width determines the
microprocessor's capacity for addressing addresses in memory.
It is sometimes called “internal memory” because unlike secondary memory devices, MP
cannot be removed so easily.
This type of memory is volatile, meaning that when the power is cut off, all the information
stored in it is erased.
MP is the core of the memory subsystem of a computer system, and has a smaller storage
capacity than secondary memory, but a speed millions of times greater. The larger the
amount of memory, the greater the data storage capacity.
When the CPU has to execute a program, it first places it in memory and then starts
executing it. The same thing happens when you need to process a set of data; before you
can process it, you have to get it into main memory.
Within the random access memory (RAM) there is a type of memory called cache memory,
which is characterized by being faster than the others, allowing the exchange of
information between the CPU and the MP to be faster.
The structure of main memory has changed over the history of computers. Since the
1980s, it has predominantly been a unit divided into cells that are identified by an address.
It is made up of blocks of integrated circuits or chips capable of storing, retaining or
"memorizing" digital information, that is, binary values; the computer's microprocessor has
access to said blocks.
Guys
Two types of memory are used in computers:
Read Only Memory (https://es.wikipedia.org/wiki/Read_Only_MemoryROM). It comes pre-set
with a series of programs. The ROM software is divided into two parts:
Boot routine or POSThttps://es.wikipedia.org/wiki/POSThttps://es.wikipedia.org/wiki/
Arranque_(inform%C3%A1tica) (Power On Self Test): performs a check of the computer
components; for example, video controller circuits, memory access, keyboard, disk drives,
etc. It is responsible for determining what hardware is present and setting up the
computer. Using a configuration program, the setup, it reads a memory called CMOS RAM
(Metal Oxide Semiconductor RAM). This can retain its content for several years, even if
the computer is turned off, with very little electrical power supplied by a battery, it saves
the date, time, available memory, hard disk capacity, and whether it has a floppy drive or
not. It is responsible for the next step of booting: it reads a BR (Boot Record) from the hard
drive or another drive (such as CD, USB, etc.), where there is a program that loads the
operating system into RAM. It then hands over control to the operating system and the
computer is ready to work.
BIOS (Basic Input-Output System) routine: remains active while the computer is in use.
Allows the activation of input/output peripherals: keyboard, monitor, mouse, etc.
Setup routinehttps://es.wikipedia.org/wiki/Setup: Primary stage in which basic options such
as the schedule can be modified. It is OS agnostic and starts before you log in.
Read-Write Memory (RWM): is the user memory that temporarily contains the program,
data, and results being used by the computer user. In general, it is volatile memory, it
loses its content when the computer is turned off, that is, it maintains the data and results
as long as the block receives electrical power, with the exception of CMOS RAM.
Both RAM and ROM are integrated circuits, called chipshttps://es.wikipedia.org/wiki/Chip.
A chip is a small chip of semiconductor material (silicon) that contains multiple integrated
circuits, such as transistors, among other electronic devices, with which numerous
functions are performed in computers and electronic devices; which allow, interrupt or
increase the flow of current. These chips are on a card or board.
It is common to mistakenly refer to read/write memory (RWM) as random access memory
(RAM), where the type of memory is confused with the way it is accessed.
The content of the memories is nothing other than binary digits or bits (binary digits), which
correspond to two logical states: 0 (zero) without electrical charge and 1 (one) with
electrical charge. Each of these states is called a bit, which is the smallest unit of data
storage.
The MP block is usually called
RAMhttps://es.wikipedia.org/wiki/Memoria_de_acceso_aleatorio, as this is the type of
memory chips that make up the block, but it is also associated with the CMOS
chiphttps://es.wikipedia.org/wiki/CMOS, which stores the system BIOS
programhttps://es.wikipedia.org/wiki/BIOS, and the peripheral devices of the secondary
memory (disks and other peripherals), to make up the computer's memory subsystem.
The CPU addresses the RAM locations in order to access the data stored in them and to
place the results of the operations.
RWM blocks, ROMs, and secondary memories make up the memory subsystem of a
computer.

Number Systems
Binary coding
Binary code is the coding system used for the representation of texts, or computer
instruction processors, using the binary system (two-digit numerical system, or bit: "0" and
"1"). In computing and telecommunications, binary code is used to encode data such as
character strings, or bit strings. For example, in the case of a CD, the signals that will
reflect the "laser" will bounce off the CD and be received by a sensor in a different way,
thus indicating whether it is a zero or a one.
In a fixed-width binary code, each letter, digit, or other symbol is represented by a bit string
of the same length, such as a binary number that typically appears in tables in octal,
decimal, or hexadecimal notation. According to Anton Glaser, in his History of Binary and
other Nondecimal Numeration, he comments that the first binary codes were used in the
year 1932: CE Wynn-Williams ("Scale of Two"), later in 1938: Atanasoff-Berry Computer,
and in 1939: Stibitz ("excess three") the code in Complex Computer.
It is also common to see the word bit referring either to the absence of a signal, expressed
with the digit "0", or to its existence, expressed with the digit "1". The byte is a group of 8
bits, that is, in it we have 256 possible binary states.
The binary system, also called the dyadic system in computer science, is a numbering
system in which numbers are represented using only two digits: zero (0) and one (1). It is
one of the systems used in computers, because they work internally with two voltage
levels, which is why their natural numbering system is the binary system.
Other Mathematical Systems of Enumeration:
Octal system
The positional numbering system whose base is 8 is called octal and uses the Indian-
Arabic digits: 0,1,2,3,4,5,6,7. In computing, octal numbering is sometimes used instead of
hexadecimal. It has the advantage that it does not require using symbols other than digits.
However, to work with bytes or sets of bytes, assuming that a byte is an 8-bit word, the
hexadecimal system is usually more convenient, since every byte thus defined is
completely representable by two hexadecimal digits.
Octal number system is a number system in base 8, a base that is an exact power of 2 or
binary numbering. This feature makes converting to binary or octal number system quite
simple. The octal system uses 8 digits (0, 1, 2, 3, 4, 5, 6, 7) and each digit has the same
value as in the decimal numbering system.
Since the octal number system uses positional notation then
for the number 3452.32 we have:
2*80 + 5*81 + 4*82 + 3*83 + 3*8-1 + 2*8-2 =
2 + 40 + 4*64 + 3*512 + 3*0,125 + 2*0,015625 =
2 + 40 + 256 + 1536 +0,375 + 0,03125 =
1834 + 0.40625d
So, 3452.32q = 1834.40625d; better yet: 3452.32(8).
The subscript "q" indicates an octal number, the letter q is used to avoid confusion
between the letter 'o' and the number 0. In computing, octal numbering is sometimes used
instead of hexadecimal. It has the advantage that it does not require using symbols other
than digits. Octal numbering may have been used in the past instead of decimal, for
example to count spaces between fingers or fingers other than thumbs.
It is used as a shorthand way of representing binary numbers that use six-bit characters.
Every three bits (half a character) is converted into a single octal digit (from the Greek oktō
'eight'). This is very important because of that.

Hexadecimal system
Positional numbering system based on 16. The hexadecimal system (abbreviated as 'Hex',
not to be confused with the sexagesimal system) is the positional numbering system that
has 16 as its base. Its current use is closely linked to computer science and computing
where CPU operations often use the byte or octet as the basic unit of memory, because a
byte represents28 possible values, and this can be represented as
8 4 4 2 1 0
2 =2 +2 =16∗16=0∗16 + 0∗16 + 0∗16
Which is equivalent to the number in base 1610016 , two hexadecimal digits correspond to
exactly one byte. In principle, since the usual numbering system is based on decimal
numbers and, therefore, only ten digits are available, the convention of using the first six
letters of the Latin alphabet to supply the missing digits was adopted. The set of symbols
is as follows:
S={0,1,2,3,4,5,6,7,8,9,A,B,C,D,F}
It should be noted that the letters correspond to the following decimal numerical values:
A = 10, B = 11, C = 12, D = 13, E = 14 and
F = 15.
Lowercase letters are sometimes used instead of capital letters. As in any positional
numbering system, the numerical value of each digit is altered depending on its position in
the string of digits, being multiplied by a certain power of the base of the system, which in
this case is 16. For example:
3E0A16 = 3×163 + E×162 + 0×161 + A×160 = 3×4096 + 14×256 + 0×16 + 10×1 = 15882.
The current hexadecimal system was introduced into the computing field for the first time
by IBM in 1963. An earlier representation, with 0-9 and uz, was used in 1956 by the
Bendix G-15 computer.

Type of Operations:
Binary arithmetic
The arithmetic operations of a computer or calculator are processed in the ALU (Arithmetic
Logic Unit) of the microprocessor. There, logic gates and other devices are combined so
that they can perform arithmetic operations at high speeds (in less than a microsecond).
The fundamental binary arithmetic operations are similar to those performed with decimal
numbers, but taking into account that the carry in addition or what is borrowed in
subtraction corresponds to the base of the number.
Binary addition
Binary addition uses the same algorithm as decimal numbers; but you should take into
account the note made in the section about carrying.

I made:
Binary subtraction
Binary addition is performed in the same way as with decimal numbers; but you should
take into account the note made in the section on binary borrowing. To subtract binary
numbers there can be two cases:
Case 1: Unsigned subtraction.

Perform the following operations

Case 2: Signed subtraction. In this case you can convert subtraction to addition, like this:
ab=a+(-b). Then apply the concept of the maximum number of bits that numbers should
have, which is usually achieved by adding and calculating the range of values.
Binary complements. There are two known types of binary complement: 1's complement
and 2's complement, which are very useful for performing signed binary operations.
In digital systems, Flip-Flops are frequently used. These are 1-bit (binary digit) memory
elements made up of combinatorial logic components connected in a feedback
configuration. Flip-Flops are electronic circuits that alternate between two states; they can
maintain an output state after the input condition that produced that state disappears.
Digital systems, in particular computers, store binary numbers in Flip-Flops to store the
sign of these numbers; in effect, an additional bit is introduced to the left of the binary, like
this:
It will store 0, if the number is positive.
It will store 1, if the number is negative.
1's complement. It allows you to change bits or binary numbers, like this: 0 for 1 and 1 for
0. Example 2.4: The 1's complement of 10111001012 is 01000110102
2's complement. It consists of adding 1 to the most significant digit of the binary number
obtained in 1's complement.
The 2's complement of 10111001012 is

The 2's complement is the most commonly used technique for representing signed binary
numbers, allowing subtraction to be converted into addition. The above means that the
digital system uses only one circuit to perform both operations.
Example 2.6: Express the following numbers with 6 bits plus the sign
a) +25
b) –25
Solution:

Note: the range of values that the binary number of the result of a signed operation must
have goes from –2n to +(2n -1) with n= number of bits. It is such that those who create
compilers have to take into account such an interval.
What is the number of bits that integers in a programming language have, given that their
range is –32768 to 32767?

Therefore, in some programming languages signed integers are written with a maximum of
16 bits.
Express with sign +49, -49, +25 and -25
Solution:
The maximum integer that can be formed with these numbers is 74, which is between
64=26 and 128=67. Then, 7 bits are needed to express said integers.

If you want to obtain –49, you would in fact have to find the 2's complement to the result
(1):

Similarly, to find the binary of –25 the result (3) is taken. By

the consequent result would be:
25=111001112 (4)
Binary multiplication
To multiply binary numbers, the same procedure is followed as with whole numbers. For
this purpose, the multiplication table that will serve as a basis for carrying out this
operation is presented below.

Perform a)101112x1012 b)11.0012x10102

Binary division
To divide binary numbers, the same procedure is followed as with whole numbers. To do
this, perform the division using the subtraction method, in which the product of the divisor
times the remainder part is subtracted from the dividend part.
Perform a)11100112 ÷ 1012 b) 10011.012 ÷ 10102

Solution:

The part in parentheses (1001)2 indicates that it is repeated; that is, it corresponds to a
repeating decimal.

Logical Operations
Binary logic is the one that works with binary variables and logical operations of Boolean
Algebra. Thus, variables only take two discrete values, T (true) and F (false), although
these two logical values can also be denoted as yes and no, or as 1 and 0 respectively.
It is the basis of Digital Systems and this also implies the basis of the structure of
computers.
Overview
What is commonly false or true in logic, in binary logic we see it represented by digits
using exclusively the values 0 and 1, numbers that in themselves do not have a numeric
value of the Real type, but rather of a discrete type; that is, 0 and 1 represent different
states of the object of study, for example, when it comes to developing a circuit.
digital.
Digital circuits generally operate under voltages of 5 volts in direct current (for example,
TTL technology) although there are exceptions such as the CMOS series, which works in
different ranges that can go from 4 to 18 volts.
Generally, the logical state 0 represents an absence of voltage, or a low level; and the
logical state 1 represents an existence of voltage, or a high level. By combining these
values, it is possible to generate a series of data convertible to any code using the
regulations applicable in each case.

Principle of duality
All boolean expressions remain valid if the '+' and '·' operators, and the '0' and '1'
elements, are interchanged. Thus, to obtain a dual algebraic expression, the operators
"AND" and "OR" are exchanged, and ones are replaced by zeros and vice versa.

Truth tables of fundamental binary operations

Logical negation or complement
It is a unary function that inverts the logical value of its argument, which is why it is also
called the NOT function, from the English "not". It can also be interpreted as subtracting
the value of the argument from 1, which is why it is sometimes known as logical
subtraction.
Notation
It is usually symbolized by a horizontal bar over its argument a () or by adding an
apostrophe after it (a ´ ).
0=1

1=0

Logical union or sum

It is a function of several arguments that equals 0 only if all its arguments equal 0. In the
rest of the cases it is equal to 1. It is equivalent to addition because there are only two
possible values, 0 and 1, so 1+1 still results in 1.

Notation
It is usually represented as a binary operator between its arguments,
symbolized by +. Or by OR (from the English conjunction "or").
Assessment
0+ 0=0
0+1=1
1+0=1
1+1=1

Logical intersection, product or multiplication

It is a function of several arguments that equals {\displaystyle 1}1 only if all of its
arguments equal 1 {\displaystyle 1}. In the rest of the cases it equals 1{\displaystyle
0}11111111 . It is equivalent to multiplication.
Notation
Just like the logical union, it is usually represented as a binary operator between its
arguments, symbolized by {\displaystyle \cdot \;} ∙∙or by AND{\displaystyle AND\;} (from the
English conjunction "and").
Assessment
0 · 0=0
0 · 1=0
1 ·0=0
1 ·1=1

Compound Logical Operations

Following Boolean Algebra, these operations can be combined using several variables and
obtaining more complex results. Below is a truth table for a compound logic operation.

Example: A · (B + C)
TO B C A·(B+C)
1 1 1 1
1 1 0 1
1 0 1 1
1 0 0 0
0 1 1 0
0 1 0 0
0 0 1 0
0 0 0 0
Binary to text encoding
Binary to text encoding is a type of data transport encoding, which is intended to protect
data being sent to other computers from being damaged due to certain restrictions in the
network transmission layer responsible for transporting the data.
An example of this limitation might be the inability of some older machines to transmit data
larger than 7 bits. Trying to send 8-bit data (i.e. data made up of 8-bit bytes) without
special encoding can cause unwanted effects when it reaches an older computer (the one
the data passes through) that may take certain octets as control characters and perform
an action on them instead of transferring them as is.
Examples of Systems:
Base64
Base 64 is a positional number system that uses 64 as the base. It is the highest power
that can be represented using only printable ASCII characters. This has led to its use for
email encryption, PGP, and other applications. All of the famous variants known as
Base64 use the character range AZ, az and 0-9 in this order for the first 62 digits, but the
symbols chosen for the last two digits vary considerably from one to another. Other
encoding methods such as UUEncode and later versions of binhex use a different set of
64 characters to represent 6 binary digits, but these are never called Base64.
UUEncode
UUEncode comes from UNIX to Unix Encoding. It is an encoding algorithm that transforms
binary code into text. Specifically, the input is a block of 8-bit bytes (usually a binary file),
and the output is a multi-line text block with LF or CR-LF line breaks and standard text
characters belonging to the UUEncode alphabet.
BinHex
BinHex is the name of a character alphabet derived from the name given to the Binary to
Hexadecimal conversion algorithm, which uses 7-bit ASCII characters.
BinHex is an algorithm that encodes files using this method that can be much larger than
the original, so BinHex encoding is usually subjected to additional compression using an
external utility such as Stuffit which is standard equipment on Macintosh computers.
 Microprocessors
A microprocessor, or central processing unit (CPU), is an internal hardware component
that performs the mathematical calculations necessary for computers to execute programs
and commands. Processors are generally made of silicon material that contains tiny
electrical components embedded in the surface. Typical computer programs that must be
processed by CPUs are Internet browsers, games, and video editing software.

Physical Connection and Encapsulation: chips and socket

Measuring the performance of a microprocessor is a complex task, since there are
different types of "loads" that can be processed with different effectiveness by processors
of the same range. One performance metric is clock speed, which allows comparing
processors with cores from the same family, but this is a very limited indicator given the
wide variety of designs with which processors of the same brand and reference are
marketed. A high-performance computer system may be equipped with several
microprocessors working in parallel, and a microprocessor may, in turn, be made up of
several physical or logical cores. A physical core refers to a quasi-independent internal
portion of the microprocessor that performs all the activities of a solitary CPU, a logical
core is the simulation of a physical core in order to more efficiently distribute processing.
There is a trend to integrate the largest number of elements within the processor itself,
thus increasing energy efficiency and miniaturization. Integrated elements include floating
point units, RAM controllers, bus controllers, and dedicated video processors.
Encapsulated:
In integrated circuit manufacturing and electronic engineering, encapsulation is the result
of the final stage of the semiconductor device manufacturing process, in which a
semiconductor or integrated circuit is placed in a casing to protect it from physical damage,
corrosion, evacuate the heat generated and in turn allow communication with the outside
through electrical contacts. The term encapsulation is commonly understood as something
to protect the semiconductor wafer piece with which integrated circuits such as
microprocessors, microcontrollers and DSPs are built; but they also protect other
electronic components such as TO-92 (Examples: Transistors 2N3904 and 2N3906,
Temperature sensor IC LM35), TO-3 (Transistor 2N3055), TO-220 (Regulators IC 78xx
and 79xx, Transistors TIP31 and TIP32), DO-41 (Diodes of the 1N4000 series), DO-41G
(5.1V Zener Diode 1N4733).

The first integrated circuits had flat ceramic packages. They were used by the military for
many years because of their reliability and small size. Commercial integrated circuits
adopted the (DIP) form, initially in ceramic and later in plastic. In the 1980s, the number of
pins in VLSI integrated circuits exceeded the practical limit for DIP packaging, leading to
new formats such as pin grid array (PGA), (LCC) (QFP). Surface mount components
appeared in the 1980s and became popular. These new surface mount packaging formats
further reduce the size of the electronic equipment they are part of.
Among the most common we have:
 DIP (Dual in-line package).
 PGA (Pin grid array).
 QFP (Quad Flat Package).
 LQFP (Low-profile Quad Flat Package).
 PLCC (Plastic Leaded Chip Carrier).

UC, Mathematical Coprocessor (Floating PU)Floating point

unit. It is the part of the micro specialized in this kind of mathematical calculations,
previously it was outside the processor in another chip. This part is considered a "logical"
part along with the registers, control unit, memory and data bus.
A Floating Point Unit (FPU) or, more commonly known as a math coprocessor, is a CPU
component specialized in floating point operations. The basic operations that every FPU
can perform are arithmetic (addition and multiplication), although some more complex
systems are also capable of performing trigonometric or exponential calculations. Not all
CPUs have a dedicated FPU. In the absence of an FPU, the CPU can use microcode
programs to emulate a floating-point function through the arithmetic logic unit (ALU), which
reduces hardware cost in exchange for a noticeable loss of speed. The objective of this
article is to show how a mathematical coprocessor can be implemented using VHDL, for
its implementation in any FPGA.

Cache Memory: Levels L1, L2 and L3

It is an ultra-fast memory that the processor uses to have direct access to certain data that
will "predictably" be used in subsequent operations, without having to go to RAM, thus
reducing the waiting time for data acquisition. All PC-compatible micros have the so-called
internal first level cache or L1; that is, the one that is inside the micro, encapsulated with it.
The most modern micros (Core i3, Core i5, Core i7, etc.) also include another level of
cache inside, larger, although somewhat slower, it is the second level cache or L2 and
there are even some with level 3 cache memory, or L3.
Level 1 (L1) Cache Memory:
This cache memory is extremely fast but relatively small and today it is integrated into the
CPU (years ago it may or may not have been integrated into the CPU). All instructions are
searched for here first, if they are not present then proceed to the next level.
Level 2 (L2) Cache Memory:
This cache is considerably larger than L1 and is also inside the CPU (years ago it was
not). If the instructions were not found in Level L1 then they are searched in this Level L2,
this type of memory is not as fast as the one used in L1 so some latency (delay) is to be
expected.
Level 3 (L3) Cache Memory:
This is a specialized memory level that helps improve the performance of the L1 and L2
Cache Levels. It is much slower than L1 or L2 memory, but much faster than System
RAM. In the case of Processors with many Cores, each of them has its own L1 Cache and
L2 Cache, but, they all share the same L3 Cache. When an instruction is fetched in L3 it is
pushed up to a higher level cache.

Rest: Records, Buses

Registers: They are basically a type of small, special-purpose memory that the micro has
available for some particular uses. There are several groups of registers in each
processor. One group of registers is designed for programmer control and there are others
that are not designed to be controlled by the processor but are used by the CPU in some
operations, in total there are thirty-two registers.
Buses: In computer architecture, a bus (or channel) is a digital system that transfers data
between the components of a computer. It is made up of wires or tracks on a printed
circuit, devices such as resistors and capacitors, and integrated circuits.
There are two types of transfer on buses:
Serial: The bus is only capable of transferring data bit by bit. That is, the bus has a single
cable that transmits information.
Parallel: The bus allows several bits to be transferred simultaneously, for example 8 bits.
Memory: is the place where the processor finds the program instructions and their data.
Both data and instructions are stored in memory, and the processor accesses them from
there. Memory is an internal part of the computer and its essential function is to provide
storage space for ongoing work.
Ports: This is how the processor communicates with the outside world. A port is
analogous to a telephone line. Any part of the computer circuitry with which the processor
needs to communicate is assigned a "port number" that the processor uses like a
telephone number to call special circuits or parts.
Transistors: Basically, transistors are semiconductor devices that are used to switch
electronic signals. In microprocessors, a higher number of transistors means better CPU
performance. For example, Intel Pentium 4 processors have between 40 and 50 million
transistors, while older Pentium 3 CPUs have 9.5 million. More transistors enable multiple
instruction decoders and pipelines, allowing multiple processes to complete during each
clock cycle.
Control Signals: Control signals are electronic signals that control the processor
components used to perform an operation or execute an instruction. An element called a
"sequencer" sends control signals to tell the specific unit what it needs to do next. For
example, a read or write signal may be sent to the cache letting it know that the processor
is preparing to read or write data to the processor memory.

Arithmetic logic unit

Arithmetic logic units (ALUs) in microprocessors enable computers to add, subtract,
multiply, divide, and perform other logical operations at high speeds. Thanks to advanced
ALUs, modern microprocessors and GPUs (Graphics Processing Units) are capable of
performing very complicated operations on large floating point numbers.
Speed: Internal and External (FSB)
The internal frequency of the processor is usually based on the FSB frequency. To
calculate the internal frequency of the CPU, the bus frequency is multiplied by a number
called the clock multiplier. It is important to note that for the calculation, the CPU uses the
actual bus frequency and not the effective frequency. To determine the actual bus
frequency for processors that use dual-data rate (DDR) buses (AMD Athlon and Duron)
and quad data rate (all Intel processors starting with the Pentium 4) the effective bus
speed should be divided by 2 for AMD and by 4 for Intel.

Motherboards alternate between high and low voltages, creating a constant pulse. The
external frequency is the number of times per second that this pulse completes a cycle.
For example, a motherboard with a pulse that cycles 100 million times per second has an
external clock speed of 100 megahertz. Although this value technically describes the
motherboard, users and manufacturers often refer to it as the external CPU frequency.
The CPU is the fastest component present in most computers, operating at a rate much
higher than the external frequency. According to DSL Reports, CPUs achieve this by using
an internal core multiplier. This component multiplies the external frequency so that the
CPU operates at a faster rate, known as the internal frequency, while remaining
synchronized with the motherboard. For example, a CPU with a frequency of 100 MHz and
a multiplier of 24x has an internal frequency of 2400 MHz, or 2.4 gigahertz.
Other computer components use multipliers to increase their operating speeds beyond
the external frequency of the motherboard. These include random access memory (RAM)
and graphics processing units. Although these components can operate faster than the
motherboard, they generally cannot operate as fast as the CPU. External frequency
consistency allows manufacturers to develop RAM and GPU units that operate as fast as
possible without having to make other components operate at the exact same rates.
Frequency boosting is a method of altering the external frequency to improve the
performance of a computer. The process involves entering the Basic Input/Output System
(BIOS) and increasing the external frequency from its default setting. Doing this requires
detailed knowledge of all of your computer's components, as this often pushes them
beyond their factory recommended limits. Increasing the frequency without understanding
and accounting for the precise limitations of your system can damage or destroy your
computer.

Multiprocessor (SMP) and Multicore (core) systems

A CPU, or Central Processing Unit, is what is commonly referred to as a processor. A

processor contains many discrete parts within it, such as one or more instruction and data
caches, instruction decoders, and various types of execution units for performing
arithmetic or logical operations.
A multiprocessor system contains more than one CPU, allowing them to work in parallel.
This is called SMP, or Simultaneous Multiprocessing.
A multi*core* CPU has multiple cores executing on a CPU. Now, this can mean different
things depending on the architecture, but it basically means that a certain subset of the
CPU components is duplicated, so multiple "cores" can work in parallel on separate
operations. This is called CMP, Chip Level Multiprocessing.
For example, a multi-core processor may have an L1 and execution unit cache for each
core, while having an L2 cache shared by the entire processor. That means that while the
processor has a large pool of slowdown cache, it has separate fast memory and
artithmetic/logical units for each of several cores. This would allow each core to perform
operations at the same time as the others.
There is even a new division, called SMT, Simultaneous Multithreading. This is where an
even smaller subset of a processor or component core is duplicated. For example, an SMT
kernel might have duplicated thread scheduling resources, so that the kernel looks like two
separate "processors" to the operating system, even though it only has one set of
execution units. A common application of this is Intel Hyperthreading.
So you could have multiple multi-core, multi-processor system. Something like two quad-
core, hyperthreaded processors gives you 2x4x2 = 16 logical processors from the OS
perspective.
Different workloads benefit from different configurations. A single-threaded workload that is
mostly performed on a single-purpose machine benefits from a very fast,
single-core/system CPU. Workloads that benefit from highly parallel systems, such as
SMP/CMP/SMT configurations, include those that have a lot of small pieces that can be
worked on simultaneously, or systems that are used for many things at once, such as a
desktop used to browse the web, play a Flash game, and watch a video all at once. In
general, hardware these days is trending more and more towards highly parallel
architectures, as most single CPU/core raw speeds are "fast enough" for common
workloads on most models.

Encapsulation (Sockets)

The CPU socket is a type of electronic socket (electromechanical support and electrical
connection system) installed on the motherboard, which is used to fix and connect the
microprocessor, without soldering it, which allows it to be removed later. For this reason, it
is used in open architecture equipment, where modularity is sought in the variety of
components, allowing the card or integrated circuit to be changed. In proprietary
architecture equipment, the integrated circuits are added to the motherboard by soldering
them, as happens in video game consoles.
There are variants from 40 connections for small integrated circuits to more than 1300 for
microprocessors. The integrated circuit retention and connection mechanisms depend on
each type of socket, although currently the use of sockets with pins (Zero Insertion Force,
ZIF) or LGA with contacts predominates.

Slots and Sockets

The microprocessor socket is the location on the motherboard where the microprocessor
is inserted. The most common types of skirting boards are those listed below.
Types of skirting boards
PGA (Pin Grid Array): It consists of a square of connectors where the microprocessor
pins are inserted and it is fixed by means of a screw on the side of the socket and by
pressure with the heatsink. Currently obsolete.
ZIF (Zero Insertion Force): This is an evolution of the PGA, where the pins are carried by
the microprocessor and are inserted into the socket connectors, but with a major variation,
the lever next to the socket allows it to be inserted without force, preventing the pins from
being bent, and locking it so that it does not come out even if the heatsink is not fitted.

BGA (Ball Grid Array): In this case, instead of pins, there are copper balls that are
soldered directly to the motherboard. Eliminates any possibility of expanding or replacing
the microprocessor.
LGA (Land Grid Array): This socket follows the ZIF philosophy, but this time the pins are
located on the motherboard instead of on the microprocessor, while the microprocessor
contains holes on its bottom where the pins will be inserted. This prevents damage to the
microprocessor pins. It is the most modern of all the plinths.
Slot: Totally different from the previous ones. It is a rectangle similar to the expansion
slots. It uses contacts (and not pins like the previous ones) for the transmission of data
and electricity. It is aided by side holding tabs to prevent it from moving. It is obsolete, only
used by Intel in the Celeron, Pentium II and Pentium III microprocessors.

Cooling: Air (Heatsinks and Fans) and Liquid

Heatsink: The unit's heatsink acts as a thermal conductor. This means that it collects all
the heat emitted by the processor, transporting it to areas with a lower temperature
gradient to try to achieve thermal equilibrium. Heat sinks use aluminum or copper.
Aluminum ones, despite being able to conduct a good amount of heat, have a low thermal
conductivity compared to copper ones, specifically we are talking about 237 W/(K m)
On the other hand, copper is heavier than aluminum, which puts extra weight on the
motherboard, but the conductivity is much higher, reaching up to 385 W/(K m)
Liquid cooling: Liquid cooling or water cooling is one of the best options available for
cooling your computer's components, but these are the high-end systems that come with a
slightly higher price, although there are currently quite affordable models.
Nowadays, liquid cooling offers many benefits that we must take into account. Light weight
and larger exchange surface.
Liquid cooling is an alternative to traditional air fans. It involves the circulation of coolants
through specific components of our computer to collect all the heat using a liquid element.
Every computer generates heat, but certain circumstances, such as unfavorable external
environment or using too much processing resources, can result in overheating. Excessive
heat can damage internal components and that is why a cooling system is a critical
component.
Advantages of liquid cooling over air cooling

Less clutter and more space: Traditional air cooling uses fans. The most basic heatsink
can contain a single fan or multiple fans installed. Another important point is the
incorporation of several fans inside the box and the creation of a forced air flow that will
allow greater movement of air through the exchange surfaces.
Liquids have much lower conductivity than metals but have a greater capacity to store it.
And that is why we created a closed circuit where a liquid collects the heat from the CPU.
This is taken to an exchanger or radiator through which this liquid circulates, distributing
the heat through the metal fins, which in turn expel the heat into the air in contact with
them.
Many people make the mistake of believing that the more powerful the computer is, the
more fans it needs to cool it. Really all we need is good flow and the right fans to facilitate
air regeneration.
One of the advantages of liquid cooling is that it does not take up much space due to its
smaller volume and a system different from air cooling. This allows the construction of a
computer with a thinner profile than the traditional one and is becoming more and more
standard in high-performance ITX cases.

Overclocking techniques: hardware hacking

The term overclock literally means "over the clock", referring to the act of increasing the
clock speed of components such as the processor or graphics card. These two are the
main components that are usually overclocked (with integrated graphics cards this is not
possible), although RAM is another objective when it comes to improving the overall
performance of the computer, and it can be applied to any component with an internal
clock or frequency marker.
By default, all graphics cards and processors come from the factory at the same level of
GHz, voltage and other parameters, although this is not done because this is the limit that
they can reach. For example, a processor that comes with multiple 3.2 GHz cores from the
factory can be overclocked via BIOS settings or special software, although how far it can
go depends solely on the individual unit in question.
Even if we buy two identical processors or graphics cards, we will most likely not be able
to overclock both to the same level since the quality of each motherboard and processing
chip also influences. In addition, to achieve high levels of overclocking we will need to
have a good cooling system, such as liquid cooling.
Overclocking is a technique primarily intended for gamers, as other users will not need to
use it for the applications they may run, since the main objective is to obtain more FPS to
display spectacular graphics without performance problems. However, not everything is an
advantage when it comes to overclocking a processor, graphics card or RAM.
Doing this puts more stress on these components, generating more energy consumption,
and therefore causing them to produce more heat, which can reduce their useful life.
Additionally, manufacturers will not process warranty on overclocked products unless
otherwise stated, although graphics cards that are already factory overclocked will logically
continue to be warranted as long as we do not make additional changes.
IA-32 (Intel Architecture, 32-bit), known generically as x86, x86-32 or i386, is the
instruction set architecture of Intel's most commercially successful processor. It is a 32-bit
extension, first implemented in the Intel 80386, derived from the older 16-bit Intel 8086,
80186 and 80286 processors and the common denominator of all subsequent x86
designs. This architecture defines the instruction set for the family of microprocessors
installed in the vast majority of personal computers in the world.
The longevity is due in part to complete backwards compatibility and the fact that the
architecture has also been extended to 64-bit, without breaking compatibility. This
extension is known as Intel 64 by Intel or AMD64 by AMD (and referred to generically as
x86-64 or x64) and is not related in any way to the 64-bit IA-64 architecture implemented
by Intel's Itanium series.

Hardware Architectures

Intel 8086

 Introduced June 8, 1978

 Clock rates:
o 5 MHz, 0.33 MIPS
o 8 MHz, 0.66 MIPS
o 10 MHz, 0.75 MIPS

 Memory is divided into even and odd banks. Simultaneously accesses both banks to
read 16 bits of data in one clock cycle
 Data bus width: 16 bits, address bus: 20 bits
 29,000 transistors at 3 μm
 Addressable memory 1 megabyte (1024^2 B)
 Up to 10 times the performance of 8080
 First used on the IBM Compaq Deskpro PC compatible computers. Later used in por-
table computing, and in the IBM PS/2 Model 25 and Model 30. Also used in the AT&T
PC6300 / Olivetti M24, a popular IBM PC compatible (predating the IBM PS/2 line)
 Segment registers were used to access more than 64 KB of data at a time, which
many programmers complained made their jobs excessively difficult.
 The first x86 CPU
 It was later renamed iAPX 86

Intel 80286

 Introduced on February 2, 1982

 Clock rates:
 o 6 MHz, 0.9 MIPS
o 8 MHz, 10 MHz, 1.5 MIPS
o 12.5 MHz, 2.66 MIPS
o 16 MHz, 20 MHz and 25 MHz available.

 Data bus width: 16 bits, address bus: 24 bits

 Includes memory protection hardware to support multitasking operating systems with
per-process address space.
 134,000 transistors at 1.5 μm
 Addressable memory 16 MB
 Protected mode features were added to 8086 with essentially the same instruction set
 3–6 × the performance of the 8086
 Widely used in contemporary IBM PC AT and AT clones.

Intel 386
The Intel 80386 (i386, 386) is a CISC microprocessor with x86 architecture. During its de-
sign it was called 'P3', because it was the prototype of the third generation x86. The i386
was used as the central processing unit of many personal computers from the mid-1980s
to the early 1990s.
Manufactured and designed by Intel, the i386 processor was launched on the market on
October 16, 1985. Intel was against manufacturing it before that date because the produc-
tion costs would have made it unprofitable. The first processors were shipped to cus-
tomers in 1986. Similarly, motherboards for i386-based computers were initially very elab-
orate and expensive, but over time their design was streamlined.
In May 2006, Intel announced that production of the 386 would end at the end of Septem -
ber 2007.1 Although it has become obsolete as a personal computer CPU, Intel has con-
tinued to manufacture the chip for embedded systems and aerospace technology.
The i386 processor was a major evolution in the world of the processor line that dates
back to the Intel 8008. The predecessor of the i386 was the Intel 80286, a 16-bit processor
with a segmented memory system. The i386 added a 32-bit architecture and a page trans-
lation unit, making it much easier to implement operating systems that used virtual mem-
ory.
Intel later introduced the i486, but neither it nor its successors introduced as many
changes to the x86 architecture as the i386 did with its 32-bit flat addressing system. Other
microprocessors, such as the Motorola 68000, had planar addressing much earlier.
Due to the high degree of compatibility, the architecture of the i386-compatible processor
set is often referred to as the i386 architecture. The instruction set for this architecture is
currently known as IA-32.
The Intel 80486 (i486, 486) is a family of 32-bit microprocessors with x86 architecture
designed and manufactured by Intel Corporation and also manufactured through license or
reverse engineering by other companies such as IBM, Texas Instruments, AMD, Cyrix and
Chips and Technologies with different or cloned designs.
The i486 is very similar to its predecessor, the Intel 80386. The main differences are that
the i486 has an optimized instruction set, a floating point unit and a unified cache
integrated into the microprocessor chip itself, and an improved bus interface unit. These
improvements make the i486 twice as fast as an i386 and i387 at the same clock speed.
However, some low-end i486s are slower than the fastest i386s.
Typical clock speeds for the i486 were:

 16 MHz (not very common)

 20 MHz (also not common)
 25 MHz
 33 MHz
 40 MHz
 50 MHz (typically a 25 MHz motherboard with clock doubling within the microproces-
sor)
 66 MHz (33 MHz with clock doubling)
 75 MHz (25 MHz with clock tripling)
 80 MHz (AMD 40 MHz version with clock doubling)
 100 MHz (33 MHz with clock tripling)
 120 MHz (40 MHz with clock tripling, AMD exclusive).

There was a 133 MHz 80486 manufactured by AMD called Am5x86-P75 that had 16 KB of
L1 cache, 0.35 micron architecture (compared to 0.6 microns of previous models), a 4x
multiplier and 33 MT/s FSB, of which different versions were also manufactured with
different voltages and different encapsulation, 3.3 V and 3.45 V, enabling an enormous
capacity for overclocking that allowed it to go up to 160 MHz, matching its performance
with that of a Pentium-90. AMD later designed the Am5x86-P75+ with 150 MHz, 16 KB L1
cache, 3x multiplier and 55 MT/s FSB at 3.45 V, making it the most powerful 80486
processor ever manufactured, of which only a few units were sold and it is a prized
collector's item among enthusiasts.
The successor to the Intel 80486 microprocessor is the Intel Pentium.
Pentium Pro

 Introduced on November 1, 1995

 Multichip module (2 dice)
 Precursor of Pentium II and III
 Mainly used in server systems
 Socket 8 (387-pin) processor package (Dual SPGA)
 5.5 million transistors
 Family 6 model 1
 0.6 μm process technology
o 16KB L1 cache
o 256KB integrated L2 cache
o System bus clock speed of 60 MHz
o Variants
 150 Mhz
 0.35μm process technology (dual die, one 0.35μm CPU with 0.6μm L2 cache)
 o 5.5 million transistors
o 512 KB or 256 KB integrated L2 cache
o System bus clock speed of 60 or 66 MHz
o Variants:

Pentium II

 Introduced May 7, 1997

 Pentium Pro with MMX and improved 16-bit performance
 242-pin processor package Slot 1 (SEC)
 Voltage Identification Pins
 7.5 million transistors
 32KB L1 cache
 512 KB 1⁄2 bandwidth external L2 cache
 The only Pentium II that did not have the L2 cache at 1⁄2 core bandwidth was the Pen-
tium II 450 PE.
 Klamath: 0.35 μm process technology (233, 266, 300 MHz)
o System bus clock speed of 66 MHz
o Family 6 model 3

Pentium III

 Katmai: 0.25 μm process technology

o Introduced on February 26, 1999
o Improved PII, i.e. P6-based core, now including Streaming SIMD Extensions
(SSE)
o 9.5 million transistors
o 512 KB (512 x 1024 B) 1⁄2 bandwidth L2 External Cache
o 242-pin processor package Slot 1 SECC2 (single edge contact cartridge 2)
o System bus clock speed 100 MHz, 133 MHz (B models)
o Slot 1
o Family 6 model 7

Pentium 4

 0.18 μm process technology (1.40 and 1.50 GHz)

o Introduced on November 20, 2000
o L2 cache was 256KB Advanced Transfer Cache (integrated)
o The processor package style was PGA423, PGA478
o System bus clock speed 400 MHz
o SIME SSE2 Extensions
o 42 million transistors
o It is used in entry-level desktops and workstations.
 0.18 μm process technology (1.7 GHz)
o Introduced on April 23, 2001
o See chips 1.4 and 1.5 for more details.
 0.18 μm process technology (1.6 and 1.8 GHz)
o Introduced on July 2, 2001
o See 1.4 and 1.5 chips for more details.
o Core voltage is 1.15 volts in maximum performance mode; 1.05 volts in optimized
battery mode
o Power <1 watt in optimized battery mode
o It is used in full-size and then lightweight mobile PCs
 0.18 μm Willamette process technology (1.9 and 2.0 GHz)
o Introduced on August 27, 2001
o See 1.4 and 1.5 chips for more details.
 Family 15 model 1
 Pentium 4 (2 GHz, 2.20 GHz)
o Introduced on January 7, 2002
 Pentium 4 (2.4 GHz)
o Introduced on April 2, 2002
 0.13 μm Northwood A process technology (1.7, 1.8, 1.9, 2, 2.2, 2.4, 2.5, 2.6, 2.8
(OEM), 3.0 (OEM) GHz)
o Improved branch prediction and other microcode tweaks
o 512KB integrated L2 cache
o 55 million transistors
o 400 MHz system bus
 Family 15 model 2
 0.13 μm Northwood B process technology (2.26, 2.4, 2.53, 2.66, 2.8, 3.06 GHz)
o 533 MHz system bus. (3.06 includes Intel Hyper-Threading Technology)
 0.13 μm Northwood C process technology (2.4, 2.6, 2.8, 3.0, 3.2, 3.4 GHz)
o 800 MHz system bus (all versions include Hyper-Threading)
o 6500 to 10,000 MIPS
The Centrino brand has represented Intel Wi-Fi and WiMAX adapters such as Intel
Centrino Wireless since January 7, 2010.1
Previously the Centrino or Centrino Duo brand (also known as Centrino Mobile
Technology in English/Centrino Mobile Technology in Spanish) was a technology
developed by Intel to promote in the design of a portable personal computer a certain
combination of:

 CPU Intel Pentium M or later Intel Core Duo or Intel Core 2 Duo
 Motherboard with Intel 855 or Intel 915 or Intel 945 chipsets
 Wireless network interface of type Intel PRO/Wireless 2100 (IEEE 802.11a/b) or
PRO/Wireless 2200 (IEEE 802.11b/g) or later
Intel ran a powerful advertising campaign to promote the Centrino brand. Due to the
ubiquity of the advertising campaign, many consumers mistakenly refer to the Pentium M
processor as the Centrino processor, when Centrino is a technology that encompasses the
processor, the chipset, and the integrated Wi-Fi wireless network card.
Different generations are designated with geographical names (cities or rivers) adding
Refresh to revisions considered minor.

Pentium M

 Banias: 0.13 μm process technology

o Introduced in March 2003
o 64KB L1 cache
o 1 MB L2 cache (integrated)
o Based on the Pentium III core, with SIME SSE2 instructions and a deeper pipeline
o 77 million transistors
o Micro-FCPGA, Micro-FCBGA processor package
o Heart of the Intel Centrino mobile system
o 400 MHz Netburst style system bus
o Family 6 model 9

Pentium D

 Dual-core microprocessor
 No Hyper-Threading
 800 (4 × 200) MHz Front Side Bus
 LGA 775 (T socket)

 Smithfield (Pentium D): 90nm process technology (2.66–3.2GHz)

o Introduced on May 26, 2005
o 2.66–3.2 GHz (model numbers 805–840)
o 230 million transistors
o 1MB × 2 (not shared, 2MB total) L2 cache
o Cache coherence between cores requires communication over the FSB
o 60% performance increase over Prescott with similar clock
o 2.66 GHz (533 MHz FSB) Pentium D 805 introduced December 2005
o Contains 2x Prescott dice in a pack
o Family 15 Model 4

 Presler (Pentium D): 65nm process technology (2.8–3.6 GHz)

o Introduced on January 16, 2006
o 2.8–3.6 GHz (model numbers 915–960)
o 376 million transistors
o 2 × 2 MB (not shared, 4 MB total) L2 cache
o Contains 2 cedar molds in one package
o Variants
  Pentium D 945
 Intel Core Duo is a sixth generation microprocessor launched in January 2006 by
Intel, after the Pentium D and predecessor to the Core 2 Duo. It has two execution
cores which makes this processor special for multi-threaded and multitasking appli-
cations. It can run multiple demanding applications simultaneously, such as graph-
ics-intensive games or calculation-intensive programs, while allowing you to down-
load music or scan your PC with an antivirus in the background, for example.
 This microprocessor implements 2 MiB of shared cache for both cores plus a 667
or 553 MHz front-side bus; it also implements the SSE3 instruction set and im-
provements to the SSE and SSE2 execution units. However, integer performance
is slightly lower due to its higher latency cache, and it does not support EM64T so it
only works at 32 bits.
 The Core Duo contains 151 million transistors, including the 2MiB cache. The pro-
cessor execution core contains a 12-stage pipeline with expected execution
speeds between 1.06 and 2.50 GHz. Communication between the L2 cache and
the two execution cores is controlled by a bus arbiter module that eliminates coher-
ence traffic across the front-side bus (FSB), at the cost of raising the core-to-L2
communication latency from 10 clock cycles (on the Pentium M) to 14 clock cycles.
Increasing the clock speed counterbalances the impact of increasing latency.
 Intel Core Duo was Intel's first microprocessor used in Apple Macintosh computers.
 There is also a version with only one core called Core Solo.
 The Intel Core 2 brand refers to a range of Intel's commercial 64-bit dual-core and
quad-core 2x2 MCM (Multi-Chip Module) CPUs with the x86-64 instruction set,
based on Intel's Core microarchitecture, derived from the 32-bit dual-core Yonah
mobile processor. Note 1 The quad-core 2x2 MCM CPU1 had two separate dual-core
dies (CPUs) - side-by-side - in a quad-core MCM package. The Core 2 relegated
the Pentium brand to the lower-midrange market, and reunited the desktop and
notebook lines, which had previously been split into the Pentium 4, D, and M
brands.
 Intel returned to lower CPU speeds and improved processor utilization of both
clock speed and power cycles compared to previous NetBurst Pentium 4/ D2 CPUs.
The Core microarchitecture provides more efficient decode stages, execution units,
cache, and buses, reducing the power consumption of Core 2 CPUs while increas-
ing processing capability. Intel CPUs have varied very sharply in power consump-
tion depending on processor speed, architecture, and semiconductor processes, as
shown in the CPU power dissipation tables.
 The Core 2 brand was introduced on July 27, 2006,3 encompassing Solo (single-
core), Duo (dual-core), Quad (quad-core), and Extreme (enthusiast dual- or quad-
core CPUs) lines through 2007.4 Intel Core 2 processors with vPro technology (de-
signed for business) include dual-core and quad-core branches.5
 Xeon is a family of Intel microprocessors for PC and Macintosh servers. The first
Xeon processor appeared in 1998 under the name Pentium II Xeon.
 The Pentium II Xeon used either the Bilingoo 440GX or 450NX chipsets. In 2000,
the Pentium II Xeon was replaced by the Pentium III Xeon.
 In 2001, the Pentium III Xeon was replaced by the Intel Xeon processor. The Xeon
is based on Intel's NetBurst architecture, the same one used by the Pentium 4
CPU.
 In 2002, Intel added the Xeon MP processor to the Xeon family, which combined
HyperThreading technology with NetBurst. Their chipsets use the 603 socket and
 have GC-LE (2 processors, 16 GiB of addressable memory) and GC-HE (4 proces-
sors or more, 64 GiB addressable) versions, all using a 400 MHz bus.
 Like Intel's standard x86/IA-32 family of desktop PC processors, the Xeon line of
processors was 32-bit, with 64-bit AMD64-based versions, such as the Xeon No-
cona, later emerging. And later the desktop processor version with this technology,
the EM64T.
 On May 9, 2004, Intel announced that future Xeon processors would be based on
the company's Pentium M architecture. Interestingly, the Pentium M is largely
based on the Pentium III architecture, so the "new" Xeon may be more similar to
the Pentium III Xeon than the NetBurst-based Xeons.
 On June 26, 2006, Intel announced the next generation Xeon Dual Core with dual-
core technology. Intel claims the new processor delivers 80 percent more perfor-
mance per watt and is 60 percent faster than the AMD competition. The new gen-
eration also offers more than twice the performance of the previous generation of
Intel Xeon processor-based servers and is capable of running 32- and 64-bit appli-
cations.
 The latter processor also replaced the aging PowerPC in MacPro workstations and
also its new 2013 model and Apple's XServe servers as the Power PC transitioned
to x86, improving its efficiency with EFI boot technology.
 Currently, it is also used by many servers that offer hosting on the Internet, given
its performance and speed of modern and current Intel Xeon processors.

AMD
The Am386 microprocessor was launched by AMD in 1991. It was a processor with
characteristics similar to the Intel 80386 and 100% compatible with the latter, millions of
units of it were sold, and this positioned AMD as a legitimate competitor to Intel, being
more than just the second source of x86 microprocessors (at that time called the 8086
family).
Although the processor was essentially ready for market by 1991, Intel kept its launch
blocked by a lawsuit. AMD had been the second largest manufacturer of Intel's designs,
and AMD's interpretation of the contract was that it covered all of its processors. Intel,
however, claimed that the contract only covered models 80286 and earlier. After a few
years, AMD finally won the case and the right to sell its Am386. This opened a line of
competition also in the market for 32-bit processors compatible with the 80386, thus
reducing the cost of purchasing a PC.
While Intel's 386 design topped out at 33 MHz, AMD released a 40 MHz version for both
its 386DX and 386SX microprocessors, extending the architecture's lifespan.
It was AMD's first solo appearance in the mass market for processors and was a decent
success considering that the market was already completely occupied by Intel.

AMD 486
The Am486 was a computer microprocessor compatible with the Intel 80486, produced by
AMD in the 1993s.
At market, Intel beat AMD by almost four years, but AMD offered its 40MHz 486 at or
below Intel's 33MHz chip, offering, for the same price, about a 20% performance
improvement.
The first AMD 486 chips were replacements for their socketed Intel counterparts, but AMD
later doubled the clock speed, and they ran at 3.3 volts instead of the 5 V of Intel's
processors, limiting their ability to upgrade Intel chips until third-party voltage adapters
became available on the market.
While competing chips to the 486 underperformed the equivalent Intel chip (such as those
from Cyrix), AMD's chips matched Intel's performance on a cycle-by-cycle basis.
While the Am386 was primarily used by small computer manufacturers, by 1994, the
Am486DX, DX2, and SX2 chips gained acceptance among large computer manufacturers,
especially Acer and Compaq.
The higher clock speeds of AMD's 486 chips provided superior performance to many early
Pentiums, especially the Pentium 60 and 66 MHz chips. While the equivalent Intel
80486DX4 chips were priced higher and required minor socket modification, AMD's was
priced lower. Initially, Intel's DX4 chips had twice the cache of AMD's, giving them slightly
higher performance, but AMD's DX4-100 cost less than Intel's DX2-66.
The Enhanced Am486 series supported new features such as extended power saving
modes and L1 Write-Back Cache, later versions had an improvement to 16 KB L1 Write-
Back Cache.
The 133 MHz AMD Am586 processor was an improved Am486.

AMD K5
The AMD K5 is an x86 type microprocessor, a direct rival to the Intel Pentium. It was the
first proprietary processor developed by AMD.
The AMD K5's RISC86 architecture was more similar to the Intel Pentium Pro architecture
than to the Pentium. The K5 is internally a RISC processor with an x86 Decoder Unit that
transforms all x86 commands from the application into RISC commands. This principle is
still used in all x86 CPUs today.
In every respect, the K5 was superior to the Pentium, however AMD had little experience
in developing microprocessors and the various production milestones set were
unsuccessfully exceeded. For this reason, it was necessary to wait a year after planned to
bring it to market. It was released on March 27, 1996. This version was still of a
"provisional" type, and was known as SSA/5, with the bugs in the L1 hidden. In the next
phase it was marketed as 5K86 and later renamed K5.
Due to late market entry and slow production as well as low production quantities, the
fastest K5 was a PR166 at 116 MHz. AMD was thus unable to convince PC manufacturers
to mount the K5. The press and the trade also assumed that the K5 was worse. The K5
can be considered a failure for AMD: "Too late". The K6 processor, successor to the K5,
changed things.

AMD K6
The K6 was a microprocessor launched in 1997 by AMD. It was designed to run on Socket
7 motherboards, compatible with those using Intel's Pentium processors. It was presented
as a competition for the Pentium II processors. The main competitive advantage of the K6
over Intel's offering was its price, which was considerably cheaper than the Pentium II and
even the Pentium MMX, which it surpassed in performance. It had significant market
acceptance, presenting AMD as a strong rival to Intel's monopoly on the market. Its
successor was the K6-2 microprocessor.
In terms of raw power, comparing its performance when running 16-bit software, there
were no major differences with its competitors. As for 32-bit applications, the Pentium Pro
and Pentium II were more powerful, with the K6 being close to these two and far above the
Pentium MMX, and even the Cyrix 6x86. In floating point calculations, essential for 3D
video games of the time, the K6 was below the Pentium II, but above the Pentium MMX.
The K6 had a range from 166 to 300 MHz, and included the MMX instruction set, which
had already become standard.

AMD K7-Athlon
Classic Core
The Athlon processor was launched on August 21, 1999. The first Athlon core, codenamed
"K7" (in homage to its predecessor, the K6), was initially available in 500 to 650 MHz ver-
sions, but later reached speeds of up to 1 GHz, making it the first processor to break the
GHz barrier. The processor supports x86 architecture and must be plugged into mother-
boards with Slot A, which are mechanically, but not electrically, compatible with Intel Slot
1.
Internally, the Athlon is a redesign of its predecessor, which has substantially improved the
floating point system (there are now 3 floating point units that can work simultaneously)
and increased the first level cache memory (L1) to 128 KiB (64 KiB for data and 64 KiB for
instructions). It also includes 512 KiB of second level (L2) cache external to the processor
integrated circuit and generally operating at half its speed (In the highest frequency models
the cache operated at 2/5 [In the 750, 800 and 850 MHz] or 1/3 [In the 900, 950 and 1,000
MHz] of the processor frequency). The communication bus is compatible with the EV6 pro-
tocol used in Alpha's DEC 21264 processors, operating at a frequency of 100 MHz DDR
(Dual Data Rate, 200 MHz effective).
The result was the most powerful x86 processor at the time. The Athlon Classic was sold
until January 2002.
In economic terms the Athlon Classic was a success, not only due to its own merits and its
low price compared to the competition, but also due to Intel's production problems.
AMD K8
The AMD K8 Hammer , also called SledgeHammer, is a computer processor
microarchitecture designed by AMD as the successor to the AMD K7 Athlon
microarchitecture . The K8 was the first implementation of the 64-bit AMD64 extension to
the x86 instruction set architecture. The K8 core is very similar to the K7. The most radical
change is the integration of AMD64 instructions and an on-chip memory controller. The
memory controller dramatically reduces memory latency and is largely responsible for
most of the performance gains from K7 to K8.

Cyrix
Cyrix was a microprocessor manufacturing company that began operations in 1988 as a
supplier of high-performance math coprocessors for 286 and 386 systems. The company
was founded by former employees of Texas Instruments, with which it had a long but
difficult relationship throughout its history. Cyrix founder Jerry Rogers aggressively
recruited several engineers and put them to work together, eventually achieving a small
but efficient design team of 30 people.
Cyrix merged with National Semiconductor on November 11, 1997, and was later sold to
VIA Technologies.
Its first products included the 486SLC and 486DLC processors, launched in 1992 and
which, despite their names, were pin-compatible with the 386SX and DX architectures,
respectively. Although they included an on-chip L1 cache as well as the 486 instruction
set, they were somewhere between a 386 and a 486 in performance. These processors
were generally used as upgrades by users looking to improve the performance of their
older 386 systems, and especially by integrators, who would swap out the processor to
convert older, hard-to-sell 386 boards into low-priced 486 systems. Later that year Cyrix
released its most famous processor, the 6x86, which was the first Cyrix CPU to surpass
the performance of the Intel processor it was intended to compete against. Cyrix initially
attempted to charge extra for this increased performance, but the 6x86's math coprocessor
was not as fast as the one in Intel's Pentium. Due to the growing popularity of 3D first-
person shooter games of the time, which made heavy use of floating point operations,
Cyrix was forced to reduce its prices. While the 6x86 quickly gained ground among
enthusiasts and independent shops, unlike AMD its processors remained unused by major
integrators. In 1996 Cyrix launched the MediaGX processor, which integrated all the
important individual components of a PC, including sound and video, into a single chip.
Initially based on the older 5x86 technology and running at 120 or 133 MHz, its
performance was widely criticized but its low price made it a success. MediaGX scored
Cyrix's first big win when Compaq used it in its lower-end Presario 2100 and 2200
computers. This led to more sales of the MediaGX to Packard Bell and also seemed to
give more prestige to Cyrix, as sales of the 6x86 to Packard Bell and eMachines followed.
Later versions of MediaGX ran at speeds up to 333 MHz and added MMX support. A
second chip was added to improve graphics capabilities.

Transmeta
Transmeta was an American company founded in 1995 whose main market is the design
of x86 compatible microprocessors with very low consumption and small size.
The main exponent of this company is the Crusoe series of microprocessors, although the
company develops various products.
Crusoe1 is a family of x86-compatible microprocessors manufactured by Transmeta. They
use a software abstraction layer or virtual machine, known as Code Morphing Software
(CMS) and which runs on a VLIW core. CMS is the only application written for the native
VLIW architecture, and translates the incoming stream of x86 instructions into VLIW
instructions.
In theory, it is possible to modify CMS to handle other instruction streams (for example, to
emulate other microprocessors), although this does not seem likely to happen since the
hardware to be used is likely to have already been optimized for x86.
Adding an abstraction layer between the x86 instruction stream and the hardware allows
the architecture to change without breaking x86 compatibility, simply by modifying CMS.
For example, Efficeon, the second generation of Crusoe, has a 256-bit VLIW core, while
the first generation had a 128-bit VLIW core.
Crusoes perform through software some of the functions classically implemented through
hardware, such as instruction reordering. This results in simpler hardware with fewer
transistors. The relative simplicity of the hardware means that the Crusoe consumes less
power (and therefore dissipates less heat) than other x86-compatible processors running
at the same clock speed.

Itanium

 Code name Merced

 Family 7
 Released on May 29, 2001
 733 MHz and 800 MHz
 2 MB cache
 All retired and replaced by Itanium 2

Itanium 2

 0x1F Family
 Released in July 2002
 900 MHz - 1.6 GHz
 McKinley 900 MHz, 1.5 MB cache, Model 0x0
 McKinley 1 GHz, 3 MB cache, Model 0x0
 Deerfield 1 GHz, 1.5 MB cache, Model 0x1
 Madison 1.3 GHz, 3 MB cache, Model 0x1
 Madison 1.4 GHz, 4 MB cache, Model 0x1
 Madison 1.5 GHz, 6 MB cache, Model 0x1
 Madison 1.67 GHz, 9 MB cache, Model 0x1
 Deep 1.4 GHz, 4 MB cache, Dual-core MCM, Model 0x1

IBM
POWER (Performance Optimization With Enhanced RISC) is a microarchitecture with a
RISC instruction set designed by IBM. POWER is also the name of a family of IBM
processors with the instruction set of this architecture and which are used as the main
CPU in IBM servers, as well as minicomputers, workstations and supercomputers.
However, there are many microprocessors that are derivatives or variants of this one that
are found in a wide variety of equipment ranging from car computers to video game
consoles.
AMD64 Opteron
Opteron is a line of x86 microprocessors from AMD for servers and workstations, and was
the first x86-based microprocessor to use the AMD64 instruction set, also known as x86-
64. It was released on April 22, 2003 with the SledgeHammer (K8) core and was aimed at
competing in the server and workstation market, particularly in the Intel Xeon processor
segment. Processors based on the AMD K10 (Barcelona) architecture were announced on
September 10, 2007, incorporating a new quad-core configuration. The most recent launch
of the Opteron processors is the Opteron 4300 and 6300 series processors ("Seoul" and
"Abu Dhabi" respectively) based on the Piledriver architecture. Opteron combines two im-
portant capabilities into a single processor:

1. Native execution of x86 32-bit applications without performance loss

2. Native execution of x86-64 64-bit applications
The first feature is notable because at the time of the Opteron's introduction, the only
commercially available 64-bit architecture with 32-bit x86 compatibility (Intel's Itanium) ran
native x86 applications only with a significant speed penalty. The second capability, by
itself, is less important, because older RISC architectures (such as SPARC, Alpha, PA-
RISC, PowerPC, MIPS) have been 64-bit for many years. By combining these two
capabilities, however, the Opteron gained recognition for its ability to cost-effectively run
the large installed base of x86 applications while also offering a path to upgrade systems
to 64-bit.
Opteron processors have an integrated memory controller supporting DDR SDRAM,
DDR2 SDRAM or DDR3 SDRAM (depending on the processor generation). This
eliminates the latency to access the main RAM and eliminates the need for a separate
northbridge IC.

AMD Athlon 64
The AMD Athlon 64 is an eighth-generation x86 microprocessor that implements the
AMD64 instruction set, which was introduced with the Opteron processor.
For the first time in computing history, the x86 instruction set has not been extended by
Intel. In fact, Intel has used this same instruction set for its later processors, such as the
Xeon Nocona. Intel calls its implementation Extended Memory Technology (EM64T), and it
is fully compatible with the AMD64 architecture. The AMD64 architecture looks set to be
the dominant computing architecture of the 64-bit generation, beating out alternatives such
as Intel's IA-64 architecture in this market.
The Athlon 64 features an on-chip memory controller and other architectural improvements
that give it better performance than the previous Athlon and Athlon XP while running at the
same speed, even when running legacy 32-bit code. AMD has chosen to measure
processor performance based on the megahertz at which a hypothetical Athlon
Thunderbird would have to run to give the same performance as an Athlon 64, rather than
stating the megahertz at which it actually runs.
There are two variants of the Athlon 64: The Athlon 64 and the Athlon 64 FX. The Athlon
64-FX is similar to the Opteron and more powerful than the regular Athlon 64. Both can run
16-bit, 32-bit, and AMD's own 64-bit assembler code. Currently, Windows NT 6.x,
GNU/Linux, OpenBSD, FreeBSD, and NetBSD support the Athlon 64's 64-bit mode, while
Microsoft has released a version of Windows XP for 64-bit computers. NOTE: The Athlon
64 also features a processor speed reduction technology called Cool'n'Quiet. When the
user is running low processor intensive applications, the processor speed and voltage are
reduced. This causes the maximum consumption to drop from 89 W to 22 W.
The Athlon 64 comes in three CPU sockets: One has 754 pins, one has 939 pins, and the
remaining one has 940 pins. The lower pinout supports a single memory channel. The 939
socket has support for memory in Dual Channel configuration, of the DDR RAM type. In
mid-2006, the 940-pin AM2 socket was introduced, leading to the implementation of DDR2
memory on Athlon 64 platforms.

POWER PC
It is a reduced instruction set architecture (RISC) created by the 1991 Apple-IBM-Motorola
alliance, known as AIM. PowerPC, as an evolving instruction set, has since 2006 been
named Power ISA, while the old name remains a trademark for some implementations of
Power Architecture-based processors.
PowerPC was the cornerstone of AIM's PReP and Common Hardware Reference Platform
initiatives in the 1990s. Originally intended for personal computers, the architecture is best
known for being used by Apple's Power Macintosh, PowerBook, iMac, iBook, and Xserve
lines from 1994 until 2006, when Apple migrated to Intel x86. It has since become a niche
in personal computers, but remains popular for high-performance and embedded
processors. Its use in the seventh generation of video game consoles and embedded
applications provided a variety of uses. Additionally, PowerPC CPUs are still used in the
AmigaOne and third-party AmigaOS 4 personal computers.
PowerPC is largely based on IBM's older POWER instruction set architecture, and retains
a high level of compatibility with it; the architectures have been kept close enough that the
same programs and operating systems will run on both if care is taken in preparation;
newer POWER series chips use the Power ISA.

Current Microprocessors
Intel
The Intel Core i9 is a computer processor designed and manufactured by Intel, which was
introduced in May 2017, and is considered to be of very high quality in the field of multi-
core microprocessors.
This is based on the Skylake microarchitecture (from the Kaby Lake processor family),
engraved in 14nm. They are designed for professional use, where a lot of computing
power is needed.
They were introduced in May 2017. With their high core count, high power consumption,
high thermal output, high performance, and unique desktop connector, LGA 2066, they are
intended to be used by enthusiasts. A mobile version based on the standard BGA1440
socket was launched in 2018, featuring six hyperlinked cores and 12MB cache. It has
been shown to reach 5 gigahertz under ideal conditions.

AMD
Ryzen is a line of microprocessors from AMD. The brand was introduced in 2017 with
products incorporating the Zen microarchitecture.
The first products in the line were officially announced during AMD's New Horizon
conference on December 13, 2016.
Zen Microarchitecture
Summit Ridge

 All models fit into the AM4 socket.

 All models support Single Rank DDR4-2666 ×2, Dual Rank DDR4-2400 ×2, Single
Rank DDR4-2133 ×4, and Dual Rank DDR4-1866 ×4.
 All models support: x86, MMX, SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, AES,
CLMUL, AVX, AVX2, FMA, CVT16/F16C, ABM, BMI1, BMI2, SHA.
 Transistors: 4.8 billion.
 Pickup size: 192 mm
 Revision: B1.
 Motherboards for Ryzen processors incorporate unlocked multipliers for overclocking.
All products support automatic overclocking, called XFR, with the X products allowing
double the boost over the others, although AMD does not mention non-X Ryzen prod-
ucts that include XFR.
 AMD officially revealed the Summit Ridge line of Ryzen processors on February 22,
2017. Ryzen desktop processors differ from APUs with Zen microarchitecture in that
they do not include an integrated GPU and rely on a dedicated one.
 Some Ryzen processors include updated versions of the Wraith cooler, called Wraith
Spire, Wraith Stealth, and Wraith Max. The original Wraith cooler had been positively
received when it was released in mid-2016. The new models incorporate RGB lighting,
except for the Wraith Stealth. The Wraith Max model is included as an option with the
X models, while the Wraith Spire is included with the Ryzen 7 1700, Ryzen 5 1600,
and Ryzen 5 1500X (the Ryzen 7 1700's cooler has RGB lighting, while the Ryzen 5
does not), and the Wraith Stealth is included with the Ryzen 5 1400.
 All models support AMD SenseMI technology, which uses Infinity Control Fabric to de-
liver the following features:
o Pure Power: Optimizes energy usage by collecting temperatures from hundreds
of sensors and distributing the workload without sacrificing performance.
o Precision Boost: Increases voltage and speed precisely (25 MHz steps) to opti-
mize power consumption.
o XFR (eXtended Frequency Range): Increases voltage and speed beyond the
maximum supported by Precision Boost, provided sufficient cooling is provided.15
o Neural Net Prediction and Smart Prefetch: Using AI prediction within the pro-
cessor to optimize workflow and cache management with intelligent prefetching of
information, it prefetches data that it predicts will be needed, optimizing RAM ac-
cess.

History
Intel was founded in 1968 but it was not until 1971 that it released its first processor, the
Intel 4004. It was the world's first microprocessor and also the first commercially available.
It was created on a simple chip, it was a 4-bit CPU with a speed of 740 KHz and it was
developed entirely by Intel. Thanks to this first microprocessor, the Busicom calculator, the
first intelligent object in history, was developed.

The following year, the brand launched the 8-bit Intel 8008, which was initially supposed to
be part of the Datapoint 2200 computer, but ultimately was not. This processor went a step
further and was up to three to four times more powerful than the 4004.
In 1974, the Intel 8080 was launched, considered the first truly usable microprocessor that
was marketed to the general public. It had 8 bits and ran at a speed of 2 MHz.

Although Intel had already marked a before and after in the technological world, with the
launch of the 8086 and 8088 processors, between 1978 and 1979, they went much further
and created the basis for what we have today. These were the first 16-bit processors and
with them they inaugurated the x86 architecture.
Back in 1985, Intel created the 80686, the first 32-bit processor. The Compaq Deskpro 386
computer used this processor and inaugurated the era of PC clones.

Years later, in 2004, Intel launched its first 64-bit processor for the Xeon line (Nocona).
With this release came the 64-bit x86 architecture, which is the one used in today's
computers.
Two years later, Intel ushered in the era of dual- and quad-core chips with the Intel Core 2
Duo. These processors significantly increased the performance of the earlier Pentium 4s.

Evolution demanded much more. In 2008, there were a lot of devices that were crying out
to be smarter. Intel then launched the first ATOM processors, specially designed for ultra-
portables and smartphones.

And as Intel's last major innovation, in 2010, it launched the first chips with integrated
graphics.