What is microprocessor

What is a Microprocessor

A microprocessor (or simply processor) is the most complex central integrated circuit in a computer system; by way of illustration, it is often called by analogy the “brain” of a computer.

It is in charge of executing all the programs, from the operating system to the user applications; It only executes instructions in binary language, performing simple arithmetic and logic operations, such as adding, subtracting, multiplying, dividing, binary logic, and memory accesses.

It may contain one or more central processing units (CPUs) consisting essentially of registers, a control unit, an arithmetic logic unit (ALU), and a floating-point computing unit (formerly known as a “math coprocessor”).

The microprocessor is generally connected through a specific socket on the computer’s motherboard; Normally for its correct and stable operation, it incorporates a cooling system that consists of a heat sink, made of a material with high thermal conductivity, such as copper or aluminum, and one or more fans that eliminate excess heat absorbed by the heat sink.

What is a Microprocessor
Intel microprocessor Pentium 4 HT 651 3.4 GHz – SL9KE-3369

Thermal paste is usually placed between the heat sink and the microprocessor capsule to improve heat conductivity.

There are other more efficient methods, such as liquid cooling or the use of Peltier cells for extreme cooling, but these techniques are used almost exclusively for special applications, such as overclocking.

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.

A performance metric is the clock frequency that allows comparing processors with cores of the same family, this being 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 can be equipped with several microprocessors working in parallel, and a microprocessor can, 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 single CPU, a logical core is the simulation of a physical core in order to more efficiently distribute the processing.

There is a tendency to integrate the largest number of elements within the processor itself, thus increasing energy efficiency and miniaturization.

Integrated elements include floating-point drives, RAM memory controllers, bus controllers, and dedicated video processors.

History of microprocessors

The history of the microprocessor

The microprocessor arose from the evolution of different predecessor technologies, basically computing, and semiconductor technology.

The beginning of the latter dates from the mid-1950s; these technologies merged in the early 1970s, producing the first microprocessor.

These technologies began their development from the Second World War; at this time scientists developed specific computers for military applications.

In the postwar period, in the mid-1940s, digital computing undertook strong growth for scientific and civil purposes as well.

Electronic technology advanced and scientists made great strides in the design of solid-state components (semiconductors). In 1948 Bell Labs created the transistor.

In the 1950s, the first general-purpose digital computers appeared. They were manufactured using vacuum tubes or bulbs as active electronic components.

Vacuum tube modules are made up of basic logic circuits, such as gates and flip-flops. Assembling them in modules was built the electronic computer (control logic, memory circuits, etc.).

Vacuum tubes were also part of the construction of machines for communication with computers.

For the construction of a simple adder circuit, some logic gates are required. Building a digital computer requires numerous electronic circuits or devices.

A momentous step in computer design was to have data stored in memory. And the idea of ​​storing programs in memory and then executing them was also of fundamental importance (von Neumann architecture).

Solid-state circuit technology evolved in the 1950s. The use of silicon (Si), inexpensive and with mass-production methods, made the transistor the most widely used component for the design of electronic circuits.

Hence the design of the digital computer was replaced from the vacuum tube by the transistor, in the late 1950s.

In the early 1960s, the state of the art in solid-state computer construction underwent a remarkable advance; Technologies emerged in digital circuits such as RTL (Transistor Resistor Logic), DTL (Transistor Diode Logic), TTL (Transistor-Transistor Logic), ECL (Complemented Emitter Logic).

In the mid-1960s the families of digital logic circuits were produced, integrated devices on the SSI and MSI scale that correspond to low and medium scale of component integration.

In the late 1960s and early 1970s, large-scale integration systems, or LSIs, emerged. LSI technology was making it possible to increase the number of components in integrated circuits.

However, few LSI circuits were produced, memory devices were a good example.

The first electronic calculators required between 75 and 100 integrated circuits.

Then an important step was taken in reducing the computer architecture to a simple integrated circuit, resulting in one that was called a microprocessor, a union of the words “Micro” from the Greek μικρο-, “pequeño”, and processor.

However, it is totally valid to use the generic term processor, since, over the years, the scale of integration has been reduced from micrometric to nanometric; and in addition, it is undoubtedly a processor.

  • The first microprocessor was the Intel 4004 from Intel Corporation, produced in 1971. It was originally developed for a calculator and was revolutionary for its time. Containing 2300 transistors, it was a 4-bit architecture microprocessor that could perform up to 60,000 operations per second working at a clock frequency of around 700 kHz.
  • The first 8-bit microprocessor was the Intel 8008, developed in mid-1972 for use in computer terminals. The Intel 8008 integrated 3300 transistors and could process at maximum frequencies of 800 kHz.
  • The first truly general-purpose microprocessor, developed in 1974, was the 8-bit Intel 8080, which contained 4,500 transistors and could execute 200,000 instructions per second operating at around 2 MHz.
  • The first 16-bit microprocessor was the 8086, followed by the 8088. The 8086 was the beginning and the first member of the popular x86 architecture, currently used in most computers. The 8086 chip was introduced to the market in the summer of 1978, but since there were no applications on the market that operated with 16 bits, Intel released the 8088, which was released in 1979. They came to operate at frequencies greater than 4 MHz.
  • The microprocessor of choice to equip the IBM Personal Computer/AT, which caused it to be the most widely used in compatible PC-ATs in the mid to late 1980s was the Intel 80286 (also known simply as 286); it is a 16-bit microprocessor, of the x86 family, which was launched on the market in 1982. It had 134,000 transistors. The final versions reached speeds of up to 25 MHz.
  • One of the first 32-bit architecture processors was Intel’s 80386, manufactured in the mid-to-late 1980s; in its different versions it came to work at frequencies of the order of 40 MHz.
  • The DEC Alpha microprocessor was launched on the market in 1992, running at 200 MHz in its first version, while the Intel Pentium emerged in 1993 with a working frequency of 66 MHz. The Alpha processor, with RISC technology and 64-bit architecture , it marked a milestone, declaring itself as the fastest in the world, in its time. It reached 1 GHz frequency around 2001. Ironically, in mid-2003, when it was planned to be removed from circulation, the Alpha still topped the list of the fastest microprocessors in the United States.
  • Modern microprocessors have a much higher capacity and speed, work on 64-bit architectures, integrate more than 700 million transistors, as in the case of the Core i7 series, and can operate at normal frequencies somewhat higher than 3 GHz ( 3000 MHz).

Evolution of microprocessors

Until the early years of the 1970s the different electronic components that made up a processor could not be a single integrated circuit, it was necessary to use two or three “chips” to make a CPU (one was the “ALU” – Arithmetical Logic Unit, the other the “Control Unit”, the other the “Register Bank”, etc.).

In 1971, the Intel company managed for the first time to put all the transistors that made up a processor on a single integrated circuit, the 4004, the microprocessor was born.

Below is a chronologically ordered list of the most popular microprocessors that were emerging. In the USSR other systems were made that gave rise to the Elbrus microprocessor series.

  • 1971: The Intel 4004

The 4004 was the world’s first microprocessor, built on a single chip and developed by Intel. It was a 4-bit CPU and it was also the first commercially available.

This development fueled Busicom’s calculator and started the path to endowing inanimate objects with “intelligence” as well as the personal computer.

  • 1972: The Intel 8008

Initially coded 1201, it was ordered from Intel by Computer Terminal Corporation to be used in its Datapoint 2200 programmable terminal, but because Intel finished the project late and did not meet Computer Terminal Corporation’s expectations, it was ultimately not used in the Datapoint. .

Later Computer Terminal Corporation and Intel agreed that the i8008 could be sold to other customers.

  • 1974: The SC/MP

The SC / MP developed by National Semiconductor was one of the first microprocessors and was available since the beginning of 1974. The name SC / MP (popularly known as “Scamp”) is the acronym for Simple Cost-effective Micro Processor. and profitable).

It features a 16-bit address bus and an 8-bit data bus. One feature, ahead of its time, is the ability to free up the buses so that they can be shared by multiple processors.

This microprocessor was widely used, due to its low cost, and supplied in kits, for educational purposes, research, and for the development of various industrial controllers.

  • 1974: The Intel 8080

The 8080 became the CPU of the first personal computer, the MITS Altair 8800, allegedly named after a destination on the Starship spacecraft from the Star Trek television show, and the IMSAI 8080, forming the basis for machines running the CP / M-80 operating system.

Computer fanatics could buy an Altair for a price (at the time) of $ 395. In a period of a few months, tens of thousands of these PCs were sold.

  • 1975: Motorola 6800

The Motorola MC6800, better known as the 6800, is manufactured by Motorola. Its name comes from the fact that it contained approximately 6800 transistors.4 It was released shortly after the Intel 8080.

Several of the early microcomputers of the 1970s used the 6800 as a processor. Among them are the SWTPC 6800, which was the first to use it, and the well-known Altair 680.

This microprocessor was widely used as part of a kit for the development of controller systems in the industry. Starting from the 6800, several derivative processors were created, one of the most powerful being the Motorola 6809

  • 1976: The Z80

The Zilog Inc. company creates the Zilog Z80. It is an 8-bit microprocessor built-in NMOS technology, and it was based on the Intel 8080. Basically, it is an extension of this, which supports all its instructions.

The first computer to use the Z80 was the Cromemco Z-1 launched in 1976. A year later the Cromemco Z-2 and the Tandy TRS-80 Model 1 came on the market with a 1.77 MHz Z80 and 4 KB of RAM..

It is one of the most successful processors on the market, of which numerous clone versions have been produced, and is still widely used today in a multitude of embedded systems.

The Zilog company was founded in 1974 by Federico Faggin, who was the chief designer of the Intel 4004 microprocessor and later the Intel 8080.

  • 1978: The Intel 8086 and 8088

A sale by Intel to IBM’s new personal computer division caused IBM’s PCs to take a major commercial hit with the new product with the 8088, the so-called IBM PC.

The success of the 8088 propelled Intel to the list of the 500 best companies, in the prestigious magazine Fortune, and the same-named the company as one of the commercial triumphs of the sixties.

  • 1982: The Intel 80286

The 80286, popularly known as the 286, was the first Intel processor that could run all the software written for its predecessor.

This software compatibility remains a hallmark of the Intel microprocessor family. Within six years of its introduction, there were an estimated 15 million 286-based PCs installed around the world.

  • 1985: The Intel 80386

This Intel processor, popularly called 386, was integrated with 275,000 transistors, more than 100 times as many as in the original 4004.

The 386 added a 32-bit architecture, with multitasking capability and a page translation unit, which made it much easier to implement operating systems that used virtual memory.

  • 1985: VAX 78032

The VAX 78032 microprocessor (also known as DC333), is a 32-bit single chip, and was developed and manufactured by Digital Equipment Corporation (DEC); installed in the MicroVAX II equipment, together with its separate floating-point coprocessor chip, the 78132, they had a power close to 90% of what the VAX 11/780 minicomputer introduced in 1977 could deliver.

This microprocessor contained 125,000 transistors, it was manufactured with DEC’s ZMOS technology. VAX systems and those based on this processor were preferred by the scientific and engineering community during the 1980s.

  • 1989: Intel 80486

The 486 generations really meant having a personal computer with advanced features, including an optimized instruction set, a floating-point unit or FPU, an improved bus interface unit, and a unified cache memory, all built into the computer itself. microprocessor chip.

These improvements made the i486s twice as fast as the i386-i387 pair operating at the same clock frequency.

The Intel 486 processor was the first to offer an integrated math coprocessor or FPU; with him that the calculation operations were accelerated remarkably.

Using an FPU unit the most complex mathematical operations are performed by the coprocessor practically independently of the function of the main processor.

  • 1991: AMD AMx86

Processors manufactured by AMD were 100% compatible with the Intel codes of that time.

Called Intel “clones”, they even went beyond the clock frequency of Intel processors and at significantly lower prices. This includes the Am286, Am386, Am486, and Am586 series.

  • 1993: PowerPC 601
IBM PowerPC 601 Microprocessor
IBM PowerPC 601 Microprocessor

It is a 32-bit RISC technology processor, in 50 and 66 MHz. In its design, they used the bus interface of the Motorola 88110.

In 1991, IBM seeks an alliance with Apple and Motorola to promote the creation of this microprocessor, the AIM alliance (Apple, IBM, and Motorola) arises whose objective was to remove the dominance that Microsoft and Intel had in systems based on the 80386 and 80486.

PowerPC (abbreviated PPC or MPC) is the original name of the family of RISC-type architecture processors, which was developed by the AIM alliance.

The processors of this family are used mainly in Apple Computer’s Macintosh computers and their high performance is strongly due to their RISC-like architecture.

  • 1993: Intel Pentium

The Pentium microprocessor had an architecture capable of executing two operations at the same time, thanks to its two data pipes of 32 bits each, one equivalent to the 486DX (u) and the other equivalent to 486SX (u).

In addition, it was equipped with a 64-bit data bus, and allowed 64-bit memory access (although the processor still maintained 32-bit support for internal operations, and the registers were also 32-bit).

Versions that included MMX instructions not only provided the user with more efficient handling of multimedia applications but were also offered at speeds of up to 233 MHz.

A 200 MHz version was included and the most basic one ran at around 166 MHz clock frequency.

The name Pentium, mentioned in comics and on daily television talk, actually became a very popular word shortly after its introduction.

  • 1994: PowerPC 620

In this year, IBM and Motorola developed the first prototype of the 64-bit PowerPC processor, the most advanced implementation of the PowerPC architecture, which was available next year.

The 620 was designed for use in servers and specially optimized for use in configurations of four and up to eight processors in the database and video application servers.

This processor incorporates seven million transistors and runs at 133 MHz.

It is offered as a migration bridge for those users who want to use 64-bit applications, without having to give up running 32-bit applications.

  • 1995: Intel Pentium Pro

Released in the fall of 1995, the Pentium Pro (professional) processor was designed with 32-bit architecture. It was used on servers, and programs and applications for (network) workstations quickly pushed its integration into computers.

32-bit code performance was excellent, but the Pentium Pro was often slower than a Pentium when running 16-bit code or operating systems. The Pentium Pro processor was made up of around 5.5 million transistors.

  • 1996: AMD K5

Having abandoned the clones, AMD manufactured with technologies analogous to Intel. AMD launched its first proprietary processor, the K5, a rival to the Pentium.

The RISC86 architecture of the AMD K5 was more similar to the architecture of the Intel Pentium Pro than the Pentium.

The K5 is internally a RISC processor with an x86 Decoder Unit, it transforms all x86 commands (from the current application) into RISC commands. This principle is used to this day in all x86 CPUs.

In most aspects the K5 was superior to the Pentium, even lower in price, however, AMD had little experience in the development of microprocessors, and the different production milestones set were surpassed with little success, it was delayed 1 year from its departure to the market, as a result, its operating frequencies were lower than those of the competition, and therefore, PC manufacturers assumed it was lower.

  • 1996: AMD K6 and AMD K6-2

With the K6, AMD not only managed to seriously compete with Intel’s Pentium MMXs, but also embittered what would otherwise have been placid market dominance, offering a processor almost on par with the Pentium II but for a price very inferior.

In floating-point calculations, the K6 also fell below the Pentium II, but above the Pentium MMX and the Pro. The K6 had a range that goes from 166 to more than 500 MHz and with the MMX instruction set. , which have already become standards.

Later an improvement of the K6 was launched, the 250 nanometer K6-2, to continue competing with the Pentium III, the latter being superior in floating-point tasks, but inferior in tasks of general use. A SIMD instruction set called 3DNow! Is introduced.

  • 1997: Intel Pentium II
Intel Pentium Pro
Intel Pentium Pro

A 7.5 million transistor processor, it is sought among the fundamental changes with respect to its predecessor, improving performance in the execution of 16-bit code, adding the MMX instruction set, and eliminating the second-level cache memory of the kernel. of the processor, placing it on a printed circuit board next to it.

Thanks to the new design of this processor, PC users can capture, review and share digital photos with friends and family via the Internet; review and add text, music, and others; with a phone line; sending video over normal telephone lines over the Internet becomes an everyday occurrence.

  • 1998: Intel Pentium II Xeon

Pentium II Xeon processors are designed to meet the performance requirements of mid-range computers, more powerful servers, and workstations.

Consistent with Intel’s strategy to design processor products to fill specific market segments, the Pentium II Xeon processor offers technical innovations designed for workstations and servers that use demanding business applications such as Internet services, storage of corporate data, digital creations, and others.

Systems based on this processor can be configured to integrate four or eight processors working in parallel, also beyond that number.

  • 1999: Intel Celeron

Continuing the strategy, Intel, in the development of processors for the specific market segment, the Celeron processor is the name that carries the low-cost line of Intel.

The objective was to be able, by means of this second brand, to penetrate in the markets prevented from the Pentium, of higher performance and price.

It is designed to add value to the PC market segment. It provided consumers with great performance at a low cost, and it delivered outstanding performance for uses such as games and educational software.

  • 1999: AMD Athlon K7 (Classic and Thunderbird)

The processor was fully compatible with the x86 architecture. Internally the Athlon is a redesign of its predecessor, but the floating-point system has been substantially improved (now with three floating point units that can work simultaneously) and the first level cache memory (L1) has been increased to 128 KB ( 64 Kb for data and 64 Kb for instructions).

It also includes 512 Kb of second-level cache (L2). The result was the most powerful x86 processor around.

The Athlon processor with a Thunderbird core appeared as the evolution of the Athlon Classic. Like its predecessor, it is also based on the x86 architecture and uses the EV6 bus. The manufacturing process used for all these microprocessors is 180 nanometers.

The Athlon Thunderbird consolidated AMD as the second-largest microprocessor manufacturing company, thanks to its excellent performance (always surpassing the Pentium III and the first Intel Pentium IVs at the same clock frequency) and low price, they made it very popular with both those in the know and computer literate.

  • 1999: Intel Pentium III

Intel Pentium III

The Pentium III processor offers 70 new Internet Streaming instructions, SIMD extensions that dramatically boost performance with advanced, 3D imaging, adding better quality audio, video, and performance in speech recognition applications.

It was designed to strengthen the performance area on the Internet, allowing users to do things such as, browse through heavy pages (with lots of graphics), virtual stores, and stream high-quality video files.

The processor is integrated with 9.5 million transistors and was introduced using 250-nanometer technology.

1999: Intel Pentium III Xeon

The Pentium III Xeon processor extends Intel’s strengths in the workstation and server market segments and adds enhanced performance in advanced business computing and e-commerce applications.

The processors incorporate enhancements that strengthen multimedia processing, particularly in video applications.

Processor III Xeon technology accelerates the transmission of information through the system bus to the processor, significantly improving performance. It is designed primarily with systems with multiprocessor configurations in mind.

  • 2000: Intel Pentium 4

This is a 7th generation microprocessor based on the x86 architecture and manufactured by Intel. It is the first with a completely new design since the Pentium Pro.

The NetBurst architecture was released, which did not give considerable improvements over the previous P6. Intel sacrificed the performance of each cycle for more cycles per second and improved SSE instructions in return.

  • 2001: AMD Athlon XP

When Intel released the 1.7 GHz Pentium 4 in April 2001, the Athlon Thunderbird was found to be out of step.

It was also impractical for overclocking, so to stay ahead of the x86 processor performance, AMD had to design a new core and release the Athlon XP.

It made compatible with the SSE instructions and the 3DNow! Among the improvements over Thunderbird, we can mention the hardware data recovery, known in English as prefetch, and the increase of the TLB entries, from 24 to 32.

  • 2004: Intel Pentium 4 (Prescott)

In early February 2004, Intel introduced a new version of Pentium 4 called ‘Prescott’. First, a 90 nm manufacturing process was used in its manufacture and then it was changed to 65 nm.

Their difference from the previous ones is that they have 1 MiB or 2 MiB of L2 cache and 16 Kb of L1 cache (double that of Northwoods), execution prevention, SpeedStep, C1E State, improved HyperThreading, SSE3 instructions, instruction handling AMD64, 64-bit created by AMD, but called EM64T by Intel, however, due to serious temperature and consumption problems, they were a failure compared to the Athlon 64.

  • 2004: AMD Athlon 64

The AMD Athlon 64 is an 8th generation x86 microprocessor that implements the AMD64 instruction set, which was introduced with the Opteron processor.

The Athlon 64 features a memory controller on the microprocessor’s own integrated circuit and other architectural enhancements that give it better performance than previous Athlon and Athlon XP running at the same speed, even running legacy 32-bit code.

The Athlon 64 also features a processor speed reduction technology called Cool’n’Quiet: when the user is running applications that require little use of the processor, it slows down and its stress is reduced.

  • 2006: Intel Core Duo

Intel released this range of dual-core processors and quad-core 2×2 MCM (Multi-Chip Module) CPUs with the x86-64 instruction set, based on Intel’s new Core architecture.

The Core microarchitecture returned to low CPU speeds and improved processor utilization of both speed and power cycles compared to previous NetBurst Pentium 4 / D2 CPUs.

The Core microarchitecture provides more efficient decoding stages, execution units, cache, and buses, reducing the power consumption of Core 2 CPUs while increasing processing capacity.

Intel CPUs have varied very sharply in power consumption according to processor speed, architecture, and semiconductor processes, shown in the CPU power dissipation tables. This range of processors was manufactured from 65 to 45 nanometers.

  • 2007: AMD Phenom

Phenom was the name given by Advanced Micro Devices (AMD) to the first generation of three and four-core processors based on the K10 microarchitecture.

As a common feature, all Phenom have 65nm technology achieved through Silicon on insulator (SOI) manufacturing technology.

However, Intel was already manufacturing using the most advanced 45nm process technology in 2008.

Phenom processors are designed to facilitate intelligent use of power and system resources, ready for virtualization, generating optimal performance per watt.

All Phenom CPUs have features such as an integrated DDR2 memory controller, HyperTransport technology, and 128-bit floating-point units, to increase the speed and performance of floating-point calculations.

The Direct Connect architecture ensures that all four cores have optimal access to the integrated memory controller, achieving 16 Gb/s bandwidth for intercom of the microprocessor cores and HyperTransport technology so that performance scales improve with the number of nuclei.

It has a shared L3 cache for faster data access (and thus less dependent on RAM latency time), plus AM2, AM2 +, and AM3 socket infrastructure support to allow for a smooth upgrade path.

Regardless, they fell short of matching the performance of the Core 2 Duo series.

  • 2008: Intel Core i7 Nehalem

Intel Core i7 is a family of quad-core processors of the Intel x86-64 architecture.

The Core i7s are the first processors to use Intel’s Nehalem microarchitecture and are the successor to the Intel Core 2 family.

FSB is replaced by the QuickPath interface on i7 and i5 (socket 1366) and replaced in turn on i7, i5, and i3 (socket 1156) by the DMI eliminating the northbridge and implementing PCI Express ports directly.

Three-channel memory (192-bit data width) – Each channel can support one or two DDR3 DIMMs.

Core i7-compatible motherboards have four (3 + 1) or six DIMM slots instead of two or four, and DIMMs must be installed in groups of three, not two.

Hyperthreading was reimplemented by creating logical cores. It is manufactured with 45 nm and 32 nm architectures and has 731 million transistors, its most powerful version.

High frequencies were used again, although consumption skyrocketed in return.

  • 2008: AMD Phenom II and Athlon II

Phenom II is the name given by AMD to a family of microprocessors or multicore CPUs (multicore) manufactured in 45 nm, which succeeds the original Phenom and supported DDR3.

One of the advantages of the change from 65 nm to 45 nm is that it allowed increasing the amount of L3 cache. In fact, this was increased in a generous way, going from 2 MiB of the original Phenom to 6 MiB.

Among them, the AMD Phenom II X2 BE 555 dual-core emerges as the bin-core processor on the market. Three Athlon IIs are also launched with only L2 Cache, but with a good price/performance ratio.

The Amd Athlon II X4 630 runs at 2.8 GHz. The AMD Athlon II X4 635 continues the same line.

AMD also launches a triple-core, called Athlon II X3 440, as well as an Athlon II X2 255 dual-core. Also comes out the Phenom X4 995, quad-core, which runs at more than 3.2 GHz. Also, AMD launches the Thurban family with six physical cores inside the package

  • 2011: Intel Core Sandy Bridge

They arrive to replace the Nehalem chips, with Intel Core i3, Intel Core i5 and Intel Core i7 2000 series, and Pentium G.

Intel released its processors which are codenamed Sandy Bridge. These Intel Core processors do not have substantial changes in architecture with respect to Nehalem, but the necessary ones to make them more efficient and faster than previous models.

It is the second generation of Intel Core with new 256-bit instructions, doubling the performance, improving the performance in 3D, and everything related to the multimedia operation.

They arrived the first week of January 2011. Includes a new set of instructions called AVX and an integrated GPU with up to 12 execution units

  • 2011: AMD Fusion

AMD Fusion is the code name for the Turion microprocessors, a product of the fusion between AMD and ATI, combining the process of 3D geometry and other functions of current GPUs. The GPU is integrated into the microprocessor itself.

The first models came out between the last months of 2010 and the first of 2011 called Ontario and Zacate (low consumption), Llano, Brazos, and Bulldozer (medium and high ranges) came out between mid and late 2011.

  • 2012: Intel Core Ivy Bridge

Ivy Bridge is the code name for the processors known as 3rd generation Intel Core. They are therefore successors to the micros that appeared at the beginning of 2011, whose code name is Sandy Bridge.

We went from the 32-nanometer wide transistor at Sandy Bridge to the 22 nanometers wide at the Ivy Bridge. This allows you to put twice as many of them in the same area.

A greater number of transistors means that you can put more functional blocks inside the chip. That is, it will be able to do a greater number of tasks at the same time.

  • 2013: Intel Core Haswell

Haswell is the codename for 4th generation Intel Core processors. They are the correction of errors of the third generation and implement new graphic technologies for gaming and graphic design, operating with lower consumption and having a better performance at a good price.

It continues like its predecessor in 22 nanometers but works with a new socket with an 1150 key. They have a high cost compared to the APUs and FX from AMD but have a higher performance.

  • 2017: Intel Core i7-7920HQ

This processor is in the line of the seventh generation, incorporating power and responsiveness never seen before.

Specially manufactured for demanding users who want to increase their productivity, without neglecting those who also want to think about entertainment and sensational games, with high data transfer and much more, it is now available on the market.

  • 2017: AMD Ryzen

It is a brand of processors developed by AMD launched in February 2017, it uses the Zen microarchitecture in the manufacturing process of 14 nm and has 4800 million transistors, they offer great multi-wire performance but a lower one using a single wire than those of your competition Intel.

These require the AM4 socket and all motherboards for this type of processor incorporate unlocked multipliers for overclocking, in addition, all products support automatic overclocking, although these processors do not have an integrated GPU, so they depend on a dedicated solution.

Ryzen processors returned AMD to high-end desktop CPUs, capable of competing in performance against Intel’s Core i7 processors at lower and competitive prices; Since its launch, AMD’s market share has increased.

  • 2019: AMD Ryzen 3rd. Generation

These processors are manufactured in the new 7nm Zen2 architecture manufacturing process and were released on July 7, 2019.

They have had a great acceptance that they have made the market share of AMD has increased and surpassed in many countries to its direct competitor intel.

The most powerful processor of this generation is the AMD Ryzen 9 3950X, a processor that has 16 cores and 32 threads inside and will be released in 4Q 2019.

  • 2020: Intel Core S 10th. Generation

These tenth-generation processors are geared toward desktop gaming computers, reaching a maximum processing frequency of 5.3 GHz in their top-of-the-range model, the i9-10900K.

  • 2020: AMD Ryzen 5000

These processors feature the 7-nanometer Zen 3 architecture for gamers, content creators, designers, and other heavy workers.

They are proficient in multi-process tasks such as 3D or video rendering and software compilation.

They have a set of technologies to increase their power such as Precision Boost 2, Precision Boost Overdrive, and PCIe 4.0.

  • 2020: Apple M1

After leaving Intel, Apple introduces its first proprietary processor for computers. It has been designed with 5-nanometer arm technology to offer high performance with low power consumption.

Various functionalities such as I / O ports, memory, and security have been integrated into the processor.

It uses a unified memory architecture (UMA) that incorporates low-latency, high-bandwidth memory in a single set of resources to increase performance and efficiency.

It has four high-performance cores with four low-power cores for less demanding tasks. It also includes high-performance integrated graphics and Neural Engine, which is an engine designed to accelerate machine learning and Artificial Intelligence.

  • 2021: 11th Generation Intel Core

These processors feature the new SuperFin transistors, combine new technologies such as WiFi 6, Thunderbolt 4, AV1 media decoding, PCI Express Gen 4 interface attached to the processor, and hardware-enhanced security features.

Supports speeds up to 4.8Ghz and Intel Optane H10 with solid-state storage for the fastest drives. They also offer artificial intelligence acceleration with their Intel Deep Learning Boost engine and dedicated card-quality Intel Iris X graphics with billions of colors, HDR 10, Dolby Atmos sound, and Dolby Vision with hardware acceleration.

  • 2021: Apple M1 Pro and M1 Max

Continuing with the development of its own platform, Apple launches its second generation of its own processors for computers. The M1 Pro processor has triple the bandwidth and is 70% faster than the original M1 processor.

While the M1 Max boasts twice the bandwidth of the M1 Pro processor and even more power. Thanks to 5-nanometer technology, the M1 Pro processor offers 33.7 billion transistors, more than double the original M1.

It has 10 cores, 8 high-performance, and 2 high-efficiency. It features a 16-core GPU that’s twice as fast as the original M1, plus it supports up to 32GB of unified memory with up to 200GB/s of bandwidth, allowing creative professionals to do much more.

On the other hand, the M1 Max processor incorporates the 10 cores of the M1 Pro processor and a 32-core GPU that allows a performance 4 times faster than the original M1.

It is Apple’s largest processor ever made with 57 billion transistors, offering up to 400GB / s memory bandwidth to support up to 64GB of unified fast memory.

M1 Pro and Max processors offer ProRes multimedia engine to speed up video processing and save power, 16-core Neural Engine to enhance machine learning and artificial intelligence, display driver for multiple external monitors, dedicated video processor and features Security features such as Secure Enclave, hardware verified secure boot, and runtime protection technologies.

Functioning of Microprocessor

From a logical, singular, and functional point of view, the microprocessor is basically composed of several registers: a control unit, a logical arithmetic unit, and, depending on the processor, it can contain a floating-point unit.

The microprocessor executes instructions stored as sequentially organized binary numbers in the main memory. The execution of the instructions can be carried out in several phases:

  • Prefetch, pre-reading of the instruction from the main memory.
  • Fetch, sending the instruction to the decoder
  • Decoding of the instruction, that is, determining what instruction it is and therefore what should be done.
  • Read operands (if any).
  • Writing of results to main memory or registers.

Each of these phases is carried out in one or more CPU cycles, depending on the structure of the processor, and specifically on its degree of segmentation.

The duration of these cycles is determined by the clock frequency, and can never be less than the time required to perform the individual task (performed in a single cycle) with the highest time cost.

The microprocessor is connected to a PLL circuit, usually based on a quartz crystal capable of generating pulses at a constant rate, so that it generates several cycles (or pulses) in a second. This watch currently generates thousands of megahertz.

Performance

Processor performance can be measured in different ways, until a few years ago it was believed that the clock frequency was an accurate measure.

And today many people believe that the number of cores is due to the incorporation of several of them within the same encapsulated to increase performance through parallel computing, but these myths, known as the “megahertz myth” and the “core myth” have been debunked by the fact that processors have not always required higher frequencies and higher frequencies.

larger numbers of cores to increase your computing power.

During the last years, the frequency has remained in the range of 1.5 GHz to 4 GHz and the number of cores has reached 16 at the moment (2021), resulting in processors with higher processing capacities compared to the first that achieved those values.

Anyway, a reliable way to measure the power of a processor is by obtaining the Instructions per cycle.

Measuring performance with the frequency and the number of cores is valid only between processors with very similar or the same architectures so that their internal operation is the same: in that case, the frequency and the number of cores is a valid comparison index.

Within a family of processors, it is common to find different options in terms of clock frequencies, because not all silicon chips have the same operating limits: they are tested at different frequencies until they show signs of instability, then they are classified according to the results of the tests.

This could be reduced in that the processors are manufactured in batches with different internal structures attending to ranges and extras such as a cache of different sizes, although this is not always the case and the high ranges differ much more from the low ranges than simply their cache memory, using different architectures in several cases.

After obtaining the batches according to their range, they are subjected to processes in a test bench and depending on their support at temperatures or that they are showing signs of instability, they are assigned a frequency, with which it will be programmed as standard, but with overclocking practices can be increased.

The capacity of a processor depends heavily on the remaining components of the system, especially the chipset, RAM, and software.

But ignoring these characteristics, an approximate measure of the performance of a processor can be obtained through indicators such as the number of floating-point operations per unit of FLOPS time, or the number of instructions per unit of MIPS time.

The exact measurement of the performance of a processor or a system is very complicated due to the multiple factors involved in the computation of a problem, generally, the tests are not conclusive between systems of the same generation.

Architecture

The microprocessor has an architecture similar to the digital computer. In other words, the microprocessor is like the digital computer in that they both perform calculations under a control program.

Consequently, the history of the digital computer helps to understand the microprocessor. It made possible the manufacture of powerful calculators and many other products.

The microprocessor uses the same type of logic that is used in the central processing unit (CPU) of a digital computer.

The microprocessor is sometimes called a microprocessor unit (MPU). In other words, the microprocessor is a data processing unit. In a microprocessor different parts can be differentiated:

Encapsulated:

It is what surrounds the silicon wafer itself, to give it consistency, prevent its deterioration (for example, by oxidation by air) and allow the connection with the external connectors that will connect it to its socket on the motherboard.

Cache memory:

It is an ultra-fast memory that the processor uses to have direct reach to certain data that “predictably” will be used in the following operations, without having to go to the RAM memory, thus reducing the waiting time for data acquisition.

All PC-compatible mics have the so-called internal first-level or L1 cache; that is, the one inside the micro, encapsulated next to it.

The most modern microphones (Core i3, Core i5, Core i7, etc.) also include another level of cache, larger, although somewhat less fast, is the second level cache or L2 and there are even those with cache memory level 3, or L3.

Math coprocessor:

Floating point unit. It is the part of the microphone specialized in that kind of mathematical calculations, formerly it was outside the processor on another chip.

This part is considered as a “logical” part together with the registers, the control unit, memory, and the data bus.

Registers:

They are basically a type of small memory with special purposes that the micro has available for some particular uses. There are several groups of registers in each processor.

A group of registers is designed for control of the programmer and there are others that are not designed to be controlled by the processor but that the CPU uses in some operations, in total there are thirty-two registers.

Memory:

It is the place where the processor finds the instructions of the programs 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 work in progress.

Ports:

It is the way in which the processor communicates with the external world. A port is analogous to a telephone line. Any part of the computer’s circuitry that the processor needs to communicate with is assigned a “port number” that the processor uses like a telephone number to call circuits or special parts.

Leave a Reply

Your email address will not be published. Required fields are marked *