History of Intel processors. All time intel processor architectures
In 1995, Intel launched the Pentium Pro microprocessor. Despite the name, it had little in common with the regular Pentium. One of the main innovations in the Pentium Pro was that x86 instructions were not directly executed in it, but were decoded in a sequence of simple internal micro-ops. In other words, the Pentium Pro "inside" was more similar to contemporary RISC processors than to previous x86 family chips.
This architecture allowed Intel to implement many measures that led to performance improvements. In particular, the Pentium Pro was the first x86 processor to receive out-of-order execution. With out-of-order execution, micro-operations first enter the operations buffer, where they are sorted and sent to computational units not in the order of arrival, but in the order of readiness for execution. This approach made it possible to practically eliminate the idle time of the processing units of the processor. The address bus width was increased to 36 bits, which, in combination with PAE technology, made it possible to increase the maximum amount of RAM to 64 GB. (However, this functionality was implemented only in server chipsets, besides, the maximum amount of memory available to one process was still 4 GB.) MB that was running at full processor clock speed. As a result, the Pentium Pro became the world's fastest 32-bit microprocessor when it hit the market, beating the PowerPC chips developed by the AIM (Apple-IBM-Motorola) alliance.
It was originally planned that the Pentium Pro would completely replace the Pentium, but this did not happen just because of the already mentioned cache memory. It turned out that the output of suitable fast SRAM memory chips capable of operating at the full processor frequency is low, so the Pentium Pro had a very high cost price. As a result, the Pentium II, released in 1997, became the heir to the Pentium, which received the MMX instruction set and cache memory operating at half the processor frequency. In addition, the Pentium II improved performance when working with 16-bit code (which was important at the time, since Windows 95 and Windows 98 still contained a large amount of 16-bit code).
Pentium III Tualatin: the fastest Pentium III
In 1999, the Pentium II was replaced by the Pentium III, which was almost identical architecturally to it, but received a new set of additional instructions, known as SSE. The Pentium III went through several iterations, later chips in this family had clock speeds above 1 GHz and 512 KB of cache running at full processor speed.
"Network Explosion"
Despite the success of the P6 microarchitecture (underlying the Pentium Pro, Pentium II, and Pentium III), the Pentium 4 was built on a very different principle. Instead of a complex core with a high IPC (Instructions Per Clock) and a relatively low clock speed, it was decided to move to a simpler core with a long pipeline and a lower IPC, but a higher clock speed. If late Pentium III processors had a 10-stage pipeline, then in Pentium 4 the pipeline length ranged from 20 to 31 stages (depending on the chip version). To compensate for the low performance of the processor core, integer compute units (ALUs) within the processor ran at twice the clock speed. For example, on a 3 GHz Pentium 4 processor, the ALUs ran at 6 GHz. Initially, it was planned that processors with NetBurst microarchitecture would reach 4 GHz, but in reality the frequency of 3.8 GHz turned out to be the limit.
The NetBurst microarchitecture can be considered relatively unsuccessful, but processors based on it have several achievements at once: the Pentium 4 became the first x86 processor to reach a clock speed of 3 GHz, and the first 64-bit x86 processor from Intel. In addition, based on the Pentium 4, the Pentium D processor was created, which became the first dual-core Intel processor.
Pentium M and its descendants
Almost immediately after the appearance of mobile Pentium 4, it became clear that the NetBurst architecture, due to high heat dissipation and power consumption, is not suitable for laptops. Therefore, in 2003, the Pentium M processor appeared, which, in fact, was an improved and modernized version of the P6 core. This processor became the basis for the highly successful Intel Centrino mobile platform, which included the processor, chipset, and Intel wireless adapter. It was the Centrino platform that made the first thin and light laptops possible. At the same time, Intel made efforts to promote wireless networks, in particular, in Ukraine, under the auspices of the company, in the mid-2000s, projects were implemented to build Wi-Fi networks at the Kiev National University. T. G. Shevchenko and the international airport "Kyiv-Borispol".
Samsung X10: one of the first thin and light laptops based on Centrino
In 2004-2005, it became clear that Pentium M processors provide higher performance than desktop processors based on the NetBurst microarchitecture. That is why the architectural solutions used in them formed the basis of the Core microarchitecture, which was used in both desktop and mobile processors. In 2006, the first desktop 4-core Intel processor was released - it was the Core 2 Extreme QX6700 with a clock speed of 2.67 GHz and 8 MB of L2 cache.
From Core"ki to Core"ki
In 2008, Intel introduced the Core i7 brand, under which top processors based on the new Nehalem microarchitecture were sold. These processors received a new system bus, integrated graphics, as well as integrated memory and PCIe bus controllers. In 2009-2010, the Core i5 and Core i3 brands were also introduced, and the Core 2 processors and their derivatives were forced out of all price segments.
In 2011, processors based on the Sandy Bridge architecture entered the market, in 2012 an improved version of Sandy Bridge called Ivy Bridge was introduced, which became the first Intel processor using the 22 nm process technology and 3D processors. In 2013, Haswell processors were introduced, followed by Broadwell in 2014 and 2015. Broadwell processors are manufactured using the 14 nm process technology. These include, among other things, the Core M processor, which has an estimated heat dissipation of only 4.5 W, which allows it to be used in devices with passive cooling.
It can be noted that the growth rate of pure processor performance has somewhat slowed down recently: in principle, even Core 2 processors (not to mention the first-generation Core i7 / i5) are enough for almost any task. This is due to the fact that manufacturers are paying more attention to improving the energy efficiency of processors and such a parameter as “performance per watt”. As a result, today's laptops built on energy-efficient Intel processors can last up to 9-12 hours on battery power and still provide enough performance for almost any task. Even 3-4 years ago this was impossible.
Atom: netbooks, tablets, smartphones...
In parallel with high-performance Core processors, Intel is also developing a line of energy-efficient Atom processors. They first appeared in 2008 as processors for netbooks (that is, low-end and cheap laptops), but have since found use as chips for smartphones and tablets based on Android and Windows operating systems. In fact, Atom, today, is the only competitor to various chips based on the ARM architecture. In 2014, 46 million Atom-based tablets were released.
Quark: smaller than Atom
Intel Galileo Quark Processor Development Board
Intel's newest family of processors is the Quark line. These are very simple processors, architecturally close to the original Pentium. Each processor also includes all the controllers needed to build the complete device. These processors are intended, first of all, to create embedded solutions combined in the "Internet of Things". For enthusiasts and developers, Intel releases Intel Galileo boards with Quark processors, these boards are compatible with Arduino and can be used to create your own projects and perform various automation tasks.
Today we are so accustomed to modern realities that we take them for granted. A smartphone in our pocket or a laptop in a bag does not seem to us a miracle of technology, but something ordinary. But it all started with a tiny chip containing 2,300 transistors and running at a clock speed of 740kHz. Sometimes it pays to look back to appreciate the extent of the journey that has been made.
This article will take a detailed look at the latest generations of Intel processors based on the Core architecture. This company occupies a leading position in the computer systems market. Most modern computers are assembled on the chips of this particular company.
Intel: development strategy
Previous generations of processors from Intel were subject to a two-year cycle. This strategy for the release of new processors of this company was called "Tick-Tock". The first stage, called "tic", is the transfer of the processor to a new technological process. So, for example, the Evey Bridge (2nd generation) and Sandy Bridge (3rd generation) generations were identical in terms of architecture. However, the production technology of the first was based on the norm of 22 nm, and the second - on 32 nm. The same can be said about Broad Well (5th generation) and Has Well (4th generation). The “so” stage, in turn, implies a fundamental change in the architecture of semiconductor crystals and a significant increase in performance. The following transitions can be given as an example:
- 1st generation West merre and 2nd generation "Sandy Bridge". In this case, the technological process was identical (32 nm), but the architecture has undergone significant changes. The northbridge of the motherboard and the built-in graphics amplifier were transferred to the central processor;
- 4th generation "Has Well" and 3rd generation "Evie Bridge". The power consumption level of the computer system was optimized, as well as the clock frequencies of the chips were increased.
- 6th Generation Sky Like and 5th Generation Broad Well: Clock speeds have also been increased and power consumption has been improved. Several new instructions have been added to improve performance.
Core architecture processors: segmentation
CPUs from Intel are positioned in the market as follows:
— Celeron are the most affordable solutions. Suitable for use in office computers designed to solve the most simple tasks.
- Pentium - almost completely identical to the Celeron processors in terms of architecture. However, higher frequencies and increased L3 cache give these processor solutions a definite advantage in terms of performance. This CPU belongs to the entry-level gaming PC segment.
- Corei3 - occupy the middle segment of the CPU from Intel. The two previous types of processors, as a rule, have two computing units. The same can be said about Corei3. However, for the first two families of chips, there is no support for Hyper Trading technology. Corei3 processors have it. Thus, at the program level, two physical modules can be converted into four program processing threads. This allows for a significant increase in performance. Based on such products, you can build your own mid-range gaming personal computer, entry-level server or even a graphics station.
- Corei5 - occupy a niche of solutions above the average level, but below the premium segment. These semiconductor crystals boast the presence of four physical cores at once. This architectural feature provides them with a performance advantage. The newer generation of Corei5 processors has high clock speeds, which allows you to constantly get a performance boost.
- Corei7 - occupy a niche in the premium segment. In them, the number of computing units is the same as in Corei5. However, they, like Corei3, have support for the Hyper Trading technology. For this reason, four cores at the software level are converted into eight processing threads. It is this feature that allows you to provide a phenomenal level of performance that any personal computer built on the basis of Intel Corei7 can boast of. These chips are priced accordingly.
Processor sockets
Generations of Intel Core processors can be installed in various types of sockets. For this reason, it will not be possible to install the first chips based on this architecture in the 6th generation CPU motherboard. And the chip codenamed "SkyLike" cannot be installed on the motherboard for the second and first generation of processors. The first processor socket is called Socket H or LGA 1156. The number 1156 here indicates the number of pins. This connector was released in 2009 for the first CPUs manufactured in 45nm and 32nm process standards. To date, this socket is already considered morally and physically obsolete. The LGA 1156 was replaced in 2010 by the LGA 1155 or Socket H1. This series motherboards support 2nd and 3rd generation Core chips. Their code names are "Sandy Bridge" and "Evie Bridge" respectively. 2013 was marked by the release of the third socket for chips, created on the basis of the Core architecture - LGA 1150 or Socket H2. It was possible to install the processor of the fourth and fifth generations in this processor socket. In 2015, the LGA 1150 socket was replaced by the current LGA 1151 socket.
First generation chips
The most affordable processors were Celeron G1101 (operating at 2.27 GHz), Pentium G6950 (2.8 GHz), Pentium G6990 (2.9 GHz). All of these solutions had two cores. The mid-range segment was occupied by Corei 3 processors with the designation 5XX (two cores / four threads for processing information). Above one step were processors with the designation 6XX. They had identical parameters with Corei3, but the frequency was higher. At the same stage was the 7XX processor with four real cores. The most productive computer systems were assembled based on the Corei7 processor. These models were designated as 8XX. In this case, the fastest chip was marked 875 K. Such a processor could be overclocked due to an unlocked multiplier. However, it also had a price to match. For these processors, you can get a significant increase in performance. The presence of the K prefix in the designation of the central processing unit means that the processor multiplier is unlocked and this model can be overclocked. The prefix S was added to the designation of energy-efficient chips.
"Sandy Bridge" and planned renovation of the architecture
The first generation of chips based on the Core architecture was replaced in 2010 by a new solution codenamed Sandy Bridge. The key feature of this device was the transfer of the integrated graphics accelerator and the north bridge to the silicon chip of the processor.
In the niche of more budget processor solutions was the Celeron processors of the G5XX and G4XX series. In the first case, two computing units were used at once, and in the second, the third-level cache was cut and only one core was present. One step higher are the Pentium G6XX and G8XX processors. In this case, the difference in performance was provided by higher frequencies. G8XX, precisely because of this important characteristic, looked much more preferable in the eyes of the user. The Corei3 processor line was represented by the 21XX models. For some designations, the T index appeared at the end. It denoted the most energy-efficient solutions with reduced performance. Corei5 solutions were designated 25XX, 24XX, 23XX. The higher the model is marked, the higher the performance level of the CPU. If the letter “S” is added at the end of the name, then this means an intermediate option in terms of power consumption between the “T” version and the standard crystal. Index "P" means that the graphics accelerator is disabled in the device. Chips with the "K" index had an unlocked multiplier. This marking remains relevant for the third generation of this architecture.
New progressive technological process
In 2013, the third generation of processors based on this architecture was released. The key innovation was a new technological process. Otherwise, there were no significant innovations. All of them are physically compatible with the previous generation of the processor. They could be installed in the same motherboards. The notation structure remains the same. Celerons were designated G12XX, while Pentiums were designated G22XX. At the beginning, instead of "2" was "3". This indicated belonging to the third generation. The Corei3 line had 32XX indexes. More advanced Corei5 processors were designated 33XX, 34XX, and 35XX. The flagship Core i7 devices were labeled 37XX.
Fourth Generation Core Architecture
The fourth generation of Intel processors was the next step. In this case, the following marking was used. Central processing units of the economy class were designated as G18XX. Pentium processors - 41XX and 43XX had the same indexes. Corei5 processors could be recognized by the abbreviations 46XX, 45XX and 44XX. The designation 47XX was used to designate Corei7 processors. The fifth generation of Intel processors based on this architecture focused mainly on use in mobile devices. For desktop personal computers, only chips related to the i7 and i5 lines were released, and only a limited number of models. The first of them were designated as 57XX, and the second - 56XX.
Promising Solutions
In early autumn 2015, the sixth generation of Intel processors debuted. At the moment, this is the most relevant processor architecture. In this case, entry-level chips are referred to as G39XX for Celeron, G44XX and G45XX for Pentium. Corei3 processors are designated 61XX and 63XX. Corei5, in turn, are designated as 64XX, 65XX and 66XX. Only one 67XX solution has been allocated for flagship models. The new generation of processor solutions from Intel is only at the beginning of development, so such solutions will remain relevant for a long time to come.
Overclocking Features
All chips based on this architecture have a locked multiplier. For this reason, overclocking of the device can only be done by increasing the frequency of the system bus. In the latest sixth generation, motherboard manufacturers will have to disable this ability to increase system performance in the BIOS. In this regard, the processors of the Corei7 and Corei5 series with the K index are an exception. These devices have an unlocked multiplier. This allows you to significantly increase the performance of computer systems built on the basis of such semiconductor products.
User opinion
All generations of Intel processors listed in this material have a high degree of energy efficiency and a phenomenal level of performance. Their only drawback is that they are too expensive. The reason here is only that Intel's direct competitor, AMD, cannot oppose worthwhile solutions. For this reason, Intel sets the price tag on its products based on its own considerations.
Conclusion
In this article, the generations of Intel processors for desktop personal computers were considered in detail. Such a list will be quite enough to understand the designations and names of processors. There are also options for computer enthusiasts and various mobile sockets. This is all done to ensure that the end user can get the most optimal processor solution. To date, the sixth generation chips are the most relevant. When assembling a new PC, you should pay attention to these models.
Intel has come a very long way from a small chip manufacturer to a world leader in processor manufacturing. During this time, many technologies for the production of processors have been developed, the technological process and device characteristics have been greatly optimized.
Many performance indicators of processors depend on the location of transistors on a silicon chip. The transistor arrangement technology is called microarchitecture or simply architecture. In this article, we will look at which Intel processor architectures have been used throughout the development of the company and how they differ from each other. Let's start with the most ancient microarchitectures and look all the way to new processors and plans for the future.
As I said, in this article we will not consider the capacity of processors. Under the word architecture, we will understand the microarchitecture of the microcircuit, the location of transistors on the printed circuit board, their size, distance, technological process, all this is covered by this concept. We will not touch the RISC and CISC instruction sets either.
The second thing to pay attention to is the generations of the Intel processor. Probably, you have already heard many times - this fifth generation processor, that fourth, and this seventh. Many people think that this is indicated by i3, i5, i7. But in fact, there is no i3, and so on - these are processor brands. And the generation depends on the architecture used.
With each new generation, the architecture improved, processors became faster, more economical and smaller, they generated less heat, but at the same time they cost more. There are few articles on the Internet that would describe all this in full. Now let's look at how it all began.
Intel processor architectures
I say right away that you should not expect technical details from the article, we will consider only the basic differences that will be of interest to ordinary users.
First processors
First, let's briefly dive into history to understand how it all began. Let's not go too far and start with 32-bit processors. The first was Intel 80386, it appeared in 1986 and could operate at a frequency of up to 40 MHz. Old processors also had a generational countdown. This processor belongs to the third generation, and the 1500 nm process technology was used here.
The next, fourth generation was 80486. The architecture used in it was called 486. The processor ran at a frequency of 50 MHz and could execute 40 million instructions per second. The processor had 8 KB of the first level cache, and the 1000 nm manufacturing process was used for manufacturing.
The next architecture was P5 or Pentium. These processors appeared in 1993, here the cache was increased to 32 kb, the frequency was up to 60 MHz, and the technical process was reduced to 800 nm. In the sixth generation P6, the cache size was 32 KB, and the frequency reached 450 MHz. The process was reduced to 180 nm.
Then the company began to produce processors based on the NetBurst architecture. Here we used 16 KB of the first level cache per core, and up to 2 MB of the second level cache. The frequency increased to 3 GHz, while the technical process remained at the same level - 180 nm. 64-bit processors appeared already here, which supported addressing more memory. Many command extensions were also made, and Hyper-Threading technology was added, which allowed two threads to be created from a single core, which improved performance.
Naturally, each architecture improved over time, the frequency increased and the process technology decreased. There were also intermediate architectures, but here everything has been simplified a bit, since this is not our main topic.
Intel Core
NetBurst was replaced in 2006 by the Intel Core architecture. One of the reasons for the development of this architecture was the impossibility of increasing the frequency in NetBrust, as well as its very large heat dissipation. This architecture was designed for the development of multi-core processors, the size of the first level cache was increased to 64 KB. The frequency remained at the level of 3 GHz, but the power consumption was greatly reduced, as well as the process technology, to 60 nm.
Core architecture processors supported Intel-VT hardware virtualization as well as some command extensions, but did not support Hyper-Threading, as they were designed based on the P6 architecture, where this capability was not yet available.
First generation - Nehalem
Further, the numbering of generations was started from the beginning, because all the following architectures are improved versions of Intel Core. The Nehalem architecture replaced the Core, which had some limitations, such as the inability to increase the clock speed. She appeared in 2007. It uses a 45 nm process and has added support for Hyper-Therading technology.
Nehalem processors have 64 KB L1 cache, 4 MB L2 cache and 12 MB L3 cache. The cache is available to all processor cores. It also became possible to integrate a graphics accelerator into the processor. The frequency has not changed, but the performance and the size of the printed circuit board have increased.
Second generation - Sandy Bridge
Sandy Bridge appeared in 2011 to replace Nehalem. The 32 nm process technology is already used here, the same amount of first-level cache, 256 MB of second-level cache and 8 MB of third-level cache are used here. Experimental models used up to 15 MB of shared cache.
Also, now all devices are available with a built-in graphics accelerator. The maximum frequency has been increased, as well as the overall performance.
Third generation - Ivy Bridge
Ivy Bridge processors are faster than Sandy Bridge processors and use the 22nm process technology. They consume 50% less energy than previous models, and also provide 25-60% higher performance. The processors also support Intel Quick Sync technology, which allows you to encode video several times faster.
Fourth generation - Haswell
The Intel Haswell processor generation was developed in 2012. The same manufacturing process was used here - 22 nm, the cache design was changed, power consumption mechanisms were improved and performance was slightly improved. But on the other hand, the processor supports many new connectors: LGA 1150, BGA 1364, LGA 2011-3, DDR4 technologies, and so on. The main advantage of Haswell is that it can be used in portable devices due to its very low power consumption.
Fifth generation - Broadwell
This is an improved version of the Haswell architecture that uses the 14nm process technology. In addition, several improvements were made to the architecture, which resulted in an average performance increase of 5%.
Sixth generation - Skylake
The next architecture of intel core processors - the sixth generation of Skylake was released in 2015. This is one of the most significant updates to the Core architecture. To install the processor on the motherboard, an LGA 1151 socket is used, DDR4 memory is now supported, but DDR3 support has been preserved. Thunderbolt 3.0 is supported, as well as the DMI 3.0 bus, which gives twice the speed. And already by tradition there was increased productivity, as well as reduced power consumption.
Seventh generation - Kaby Lake
The new, seventh generation Core - Kaby Lake was released this year, the first processors appeared in mid-January. There haven't been many changes here. The 14 nm process technology has been retained, as well as the same LGA 1151 socket. DDR3L SDRAM and DDR4 SDRAM memory sticks, PCI Express 3.0, USB 3.1 buses are supported. In addition, the frequency was slightly increased, and the density of the transistors was also reduced. The maximum frequency is 4.2 GHz.
conclusions
In this article, we looked at the Intel processor architectures that were used in the past, as well as those that are used now. Further, the company plans to switch to the 10 nm process technology and this generation of Intel processors will be called CanonLake. But so far, Intel is not ready for this.
Therefore, in 2017 it is planned to release an improved version of SkyLake under the code name Coffe Lake. It is also possible that there will be other microarchitectures of the Intel processor until the company fully masters the new process technology. But we will learn about all this in time. I hope this information was useful to you.
about the author
Founder and administrator of the site site, I am fond of open source software and the Linux operating system. I currently use Ubuntu as my main OS. In addition to Linux, I am interested in everything related to information technology and modern science.
Understand the company Intel and its three founders is possible only when you understand Silicon Valley and its origins. And in order to do this, you need to penetrate into the history of the company. Shockley Transistor, Treacherous Eight and Fairchild Semiconductor. Without their understanding, Intel will remain the same to you as it is to most people - a mystery.
The invention of computers did not mean that a revolution immediately began. The first computers, based on large, expensive, rapidly breaking vacuum tubes, were expensive monsters that only corporations, research universities, and the military could maintain. The advent of transistors, and later new technologies to etch millions of transistors on a tiny microchip, meant that the processing power of many thousands of ENIAC devices could be concentrated in a rocket head, in a lap computer, and in handheld devices.
In 1947, Bell Laboratory engineers John Bardeen and Walter Brattain invented the transistor, which was introduced to the general public in 1948. A few months later, William Shockley, one of the employees of Bell, developed a model of a bipolar transistor. The transistor, which is essentially a solid-state electronic switch, has replaced the bulky vacuum tube. The move from vacuum tubes to transistors started a miniaturization trend that continues today. The transistor was one of the most important discoveries of the 20th century.
In 1956, Nobel laureate in physics William Shockley formed the Shockley Semiconductor Laboratory to work on four-layer diodes. Shockley failed to bring in his former employees from Bell Labs; instead, he hired a group of what he considered to be the best young electronics professionals fresh out of American universities. In September 1957, due to a conflict with Shockley, who decided to stop researching silicon semiconductors, eight key employees of Shokley Transistor decided to leave their jobs and start their own business. The eight people are now forever known as the Treacherous Eight. This epithet was given to them by Shockley when they left work. The G8 included Robert Noyce, Gordon Moore, Jay Last, Gene Hourney, Victor Grinich, Eugene Kleiner, Sheldon Roberts and Julius Blank.
After leaving, they decided to create their own company, but there was nowhere to take investments from. As a result of calling 30 firms, they came across Fairchild, the owner of Fairchild Camera and Instrument. He happily invested a million and a half dollars in a new company, which was almost twice as much as its eight founders initially considered necessary. A so-called premium deal was struck: if the company was successful, he could buy it out in full for three million. Fairchild Camera and Instrument exercised this right as early as 1958. The subsidiary was named Fairchild Semiconductor.
In January 1959, one of the eight founders of Fairchild, Robert Noyce, invented the silicon integrated circuit. At the same time, Jack Kilby at Texas Instruments invented the germanium integrated circuit six months earlier - in the summer of 1958, but Noyce's model turned out to be more suitable for mass production, and it is she who is used in modern chips. In 1959, Kilby and Noyce independently filed patents for the integrated circuit, and both were successfully granted, with Noyce receiving his patent first.
In the 1960s, Fairchild became one of the leading manufacturers of operational amplifiers and other analog integrated circuits. However, at the same time, the new management of Fairchild Camera and Instrument began to restrict the freedom of action of Fairchild Semiconductor, which led to conflicts. One by one, G8 members and other experienced employees began to quit and start their own companies in Silicon Valley.
The first name Noyce and Moore chose was NM Electronics, N and M being the first letters of their last names. But it wasn't very impressive. After a large number of not very successful proposals, for example, Electronic Solid State Computer Technology Corporation, they came to the final decision: the company will be called Integrated Electronics Corporation. In itself, it was also not very impressive, but it had one merit. The company could be called Intel for short. It sounded good. The title was energetic and eloquent.
Scientists set themselves a very specific goal: to create a practical and affordable semiconductor memory. Nothing like this had been created before, given the fact that silicon-based memory cost at least a hundred times more than conventional magnetic-core memory for that time. Solid-state memory cost as much as one dollar per bit, while magnetic-core memory cost only about a cent per bit. Robert Noyce said: “We only had to do one thing - reduce the cost by a hundred times and thereby conquer the market. That's what we basically did."
In 1970, Intel released a 1 Kb memory chip, far exceeding the capacity of the chips existing at that time (1 Kb is equal to 1024 bits, one byte consists of 8 bits, that is, the chip could store only 128 bytes of information, which is negligible by modern standards. ) The resulting chip, known as Dynamic Random Access Memory (DRAM) 1103, became the world's best-selling semiconductor device by the end of next year. By this time, Intel had grown from a handful of enthusiasts to a company of more than a hundred employees.
At this time, the Japanese company Busicom approached Intel with a request to develop a chipset for a family of high performance programmable calculators. The original design of the calculator provided for a minimum of 12 chips of various types. Intel engineer Ted Hoff rejected this concept and instead designed a single-chip logic device that receives application commands from semiconductor memory. This central processing unit ran a program that made it possible to adapt the functions of the microcircuit to perform incoming tasks. The microcircuit was universal in nature, that is, its use was not limited to a calculator. Logical modules had only one purpose and a strictly defined set of commands, which were used to control its functions.
There was one problem with this chip: all rights to it belonged exclusively to Busicom. Ted Hoff and other developers realized that this design had virtually unlimited uses. They insisted that Intel buy the rights to the chip they created. Intel offered Busicom to return the $60,000 it paid for the license in exchange for the right to dispose of the developed chip. As a result, Busicom, being in a difficult financial situation, agreed.
On November 15, 1971, the first 4-bit microcomputer set 4004 appeared (the term microprocessor appeared much later). The microcircuit contained 2,300 transistors, cost $200, and was comparable in its parameters to the first ENIAC computer, created in 1946, which used 18,000 vacuum tubes and occupied 85 cubic meters.
The microprocessor performed 60 thousand operations per second, operated at a frequency of 108 kHz and was manufactured using 10-micron technology (10,000 nanometers). Data was transmitted in blocks of 4 bits per clock, and the maximum addressable memory size was 640 bytes. The 4004 was used to control traffic lights, in blood tests, and even in the Pioneer 10 research rocket launched by NASA.
In April 1972, Intel released the 8008 processor, which ran at 200 kHz.
The next processor model, the 8080, was announced in April 1974.
This processor already contained 6000 transistors and could address 64 KB of memory. The first personal computer (not PC) Altair 8800 was assembled on it. The CP / M operating system was used in this computer, and Microsoft developed an interpreter for the BASIC programming language for it. It was the first mass-produced computer for which thousands of programs were written.
Over time, the 8080 became so famous that it began to be copied.
In late 1975, several former Intel engineers involved in the development of the 8080 processor formed Zilog. In July 1976, this company released the Z-80 processor, which was a much improved version of the 8080.
This processor was not pin-compatible with the 8080, but combined many different features, such as a memory interface and a RAM upgrade circuit, which made it possible to develop cheaper and simpler computers. The Z-80 also included an extended 8080 instruction set to allow the use of its software. This processor included new instructions and internal registers, so the software developed for the Z-80 could be used with almost all versions of the 8080.
Initially, the Z-80 processor ran at 2.5 MHz (later versions ran at 10 MHz), contained 8500 transistors, and could address 64 KB of memory.
Radio Shack chose the Z-80 processor for its TRS-80 Model 1 personal computer. The Z-80 soon became the standard processor for systems running the CP/M operating system and the most common software of the day.
Intel did not stop there, and in March 1976 released the 8085 processor, which contained 6500 transistors, operated at a frequency of 5 MHz and was manufactured using 3-micron technology (3000 nanometers).
Despite being released a few months before the Z-80, it never quite caught on with the latter's popularity. It was used mainly as a control chip for various computerized devices.
In the same year, MOS Technologies released the 6502 processor, which was completely different from Intel processors.
It was developed by a group of engineers from Motorola. The same group worked on the 6800 processor, which would eventually evolve into the 68000 family of processors. The first version of the 8080 processor cost $300, while the 8-bit 6502 only cost about twenty-five dollars. Such a price was quite acceptable for Steve Wozniak, and he built the 6502 processor into the new Apple I and Apple II models. The 6502 processor was also used in systems built by Commodore and other manufacturers.
This processor and its successors worked successfully in gaming computer systems, which included the Nintendo Entertainment System. Motorola continued to work on the 68000 series of processors, which were subsequently used in Apple Macintosh computers. The second generation of Macs used the PowerPC processor, which is the successor to the 68000. Today, Macs have reverted to the PC architecture and share processors, chipsets, and other components with them.
In June 1978, Intel introduced the 8086 processor, which contained an instruction set codenamed x86.
The same instruction set is still supported in all modern microprocessors: AMD Ryzen Threadripper 1950X and Intel Core i9-7920X. The 8086 processor was completely 16-bit - internal registers and a data bus. It contained 29,000 transistors and ran at 5 MHz. Thanks to the 20-bit address bus, it could address 1 MB of memory. When creating the 8086, backward compatibility with the 8080 was not provided. But at the same time, the significant similarity of their commands and language made it possible to use earlier versions of the software. This feature subsequently played an important role in the rapid transfer of CP/M (8080) system programs to PC rails.
Despite the high efficiency of the 8086 processor, its price was still too high by the standards of that time and, more importantly, it required an expensive chip to support a 16-bit data bus. To reduce the cost of the processor, in 1979 Intel released the 8088 processor, a simplified version of the 8086.
The 8088 used the same internal core and 16-bit registers as the 8086, could address 1 MB of memory, but unlike the previous version, used an external 8-bit data bus. This made it possible to ensure backward compatibility with the previously developed 8-bit 8085 processor and thereby significantly reduce the cost of creating motherboards and computers. That's why IBM chose the stripped-down 8088 rather than the 8086 for its first PC. This decision had far-reaching consequences for the entire computer industry.
The 8088 processor was fully software compatible with the 8086, allowing for 16-bit software. The 8085 and 8080 processors used a very similar instruction set, so programs written for older processors could be easily converted to the 8088 processor. This, in turn, made it possible to develop a variety of programs for the IBM PC, which was the key to its future success. Not wanting to stop halfway, Intel was forced to provide 8086/8088 backward compatibility support with most of the processors released at the time.
Intel immediately began to develop a new microprocessor after the release of 8086/8088. The 8086 and 8088 processors required a large number of support chips, and the company decides to develop a microprocessor that already contains all the necessary modules on a chip. The new processor included many components that were previously available as separate chips, which would dramatically reduce the number of chips in a computer, and, consequently, reduce its cost. In addition, the system of internal commands has been expanded.
In the second half of 1982, Intel released the 80186 embedded processor, which, in addition to the improved 8086 core, also contained additional modules that replaced some of the support chips.
Also in 1982, the 80188 was released, which is a variant of the 80186 microprocessor with an 8-bit external data bus.
Released on February 1, 1982, the 16-bit x86-compatible 80286 microprocessor was an improvement on the 8086 processor with 3 to 6 times the performance.
This qualitatively new microprocessor was then used in the landmark IBM PC-AT computer.
The 286th processor was developed in parallel with the 80186/80188 processors, but it lacked some modules that were available in the Intel 80186 processor. The Intel 80286 processor was produced in exactly the same package as the Intel 80186 - LCC, as well as in PGA-type packages with sixty-eight conclusions.
In those years, backward compatibility of processors was still supported, which did not interfere with the introduction of various innovations and additional features. One of the major changes was the transition from the 16-bit internal architecture of the 286 and earlier processors to the 32-bit internal architecture of the 386 and later IA-32 processors. This architecture was introduced in 1985, but it took another 10 years for operating systems like Windows 95 (partially 32-bit) and Windows NT (requiring 32-bit drivers only) to hit the market. And only 10 years later, the Windows XP operating system appeared, which was 32-bit both at the driver level and at the level of all components. So, it took 16 years to adapt 32-bit computing. For the computer industry, this is quite a long time.
80386th appeared in 1985. It contained 275,000 transistors and performed over 5 million operations per second.
Compaq's DESKPRO 386 computer was the first PC based on the new microprocessor.
The next in the x86 family of processors was the 486th, which appeared in 1989.
Meanwhile, the US Department of Defense was not happy with the prospect of being left with a single chip supplier. As the latter became less and less (remember what a zoo was observed back in the early nineties), the importance of AMD as an alternative manufacturer grew. Under an agreement from 1982, AMD had all the licenses for the production of 8086, 80186 and 80286 processors, however, the newly developed Intel 80386 processor categorically refused to transfer to AMD. And broke the deal. What followed was a long and high-profile lawsuit - the first in the history of companies. It ended only in 1991 with the victory of AMD. For its position, Intel paid the plaintiff a billion dollars.
But still, the relationship was spoiled, and there was no talk of former trust. Moreover, AMD took the path of reverse engineering. The company continued to release Am386 processors, which differed in hardware, but completely coincided in microcode, and then Am486. Intel has already gone to court. Again, the process dragged on for a long time, and success turned out to be on one side, then on the other. But on December 30, 1994, a court decision was made, according to which the Intel microcode is still the property of Intel, and it is somehow not good for other companies to use it if the owner does not like it. So things have changed since 1995. On Intel Pentium and AMD K5 processors, any applications for the x86 platform were launched, but from the point of view of architecture, they were fundamentally different. And it turns out that the real competition between Intel and AMD began only a quarter of a century after the creation of the companies.
However, to ensure compatibility, cross-pollination by technologies has not gone anywhere. Modern Intel processors have a lot of AMD's patents, and conversely, AMD neatly adds Intel-designed instruction sets.
In 1993, Intel introduced the first Pentium processor, which was five times faster than the 486 family. This processor contained 3.1 million transistors and performed up to 90 million operations per second, which is about 1500 times faster than the 4004.
When the next generation of processors appeared, those who had counted on the Sexium name were disappointed.
The P6 family processor, called the Pentium Pro, was born in 1995.
Revisiting the P6 architecture, Intel introduced the Pentium II processor in May 1997.
It contained 7.5 million transistors, packed in a cartridge, unlike a traditional processor, which made it possible to place the L2 cache memory directly in the processor module. This helped to significantly increase its performance. In April 1998, the Pentium II family was expanded with the low cost Celeron processor used in home PCs and the professional Pentium II Xeon processor for servers and workstations. Also in 1998, Intel for the first time integrated L2 cache (which ran at the full processor core frequency) directly into the die, which made it possible to significantly increase its performance.
While the Pentium processor was rapidly gaining market dominance, AMD acquired NexGen, which was working on the Nx686 processor. The merger resulted in the AMD K6 processor.
This processor was both hardware and software compatible with the Pentium processor, that is, it was installed in a Socket 7 socket and executed the same programs. AMD continued to develop faster versions of the K6 processor and captured a significant portion of the mid-range PC market.
The first high-end desktop processor to include an on-chip L2 cache and run at full core frequency was the Pentium III processor, based on the Coppermine core, introduced in late 1999, which was essentially a Pentium II, containing SSE instructions.
In 1998, AMD introduced the Athlon processor, which allowed it to compete with Intel in the high-speed desktop PC market almost on a par. 
This processor turned out to be very successful, and Intel received it in the face of a worthy rival in the field of high-performance systems. Today, the success of the Athlon processor is beyond doubt, but at the time of its entry into the market, there were concerns about this. The fact is that, unlike its predecessor K6, which was compatible both at the software and hardware levels with the Intel processor, Athlon was compatible only at the software level - it required a specific system logic chipset and a special socket.
The new AMD processors were produced using 250nm technology with 22 million transistors. They had a new integer calculation unit (ALU). The EV6 system bus provided data transfer on both edges of the clock signal, which made it possible to obtain an effective frequency of 200 MHz at a physical frequency of 100 megahertz. The first level cache was 128 KB (64 KB instructions and 64 KB data). The second level cache reached 512 KB.
The year 2000 was marked by the appearance on the market of new developments of both companies. On March 6, 2000, AMD released the world's first 1GHz processor. It was a representative of the increasingly popular Athlon family based on the Orion core. AMD also introduced the Athlon Thunderbird and Duron processors for the first time. The Duron processor was essentially identical to the Athlon processor and differed from it only in a smaller L2 cache. Thunderbird, in turn, used an integrated cache memory, which made it possible to increase its performance. Duron was a cheaper version of the Athlon processor, which was designed primarily to compete with the inexpensive Celeron processors. And Intel at the end of the year introduced a new Pentium 4 processor.
In 2001, Intel released a new version of the 2 GHz Pentium 4 processor, which was the first processor to achieve this frequency. In addition, AMD introduced the Athlon XP processor, based on the Palomino core, as well as the Athlon MP, designed specifically for multiprocessor server systems. During 2001, AMD and Intel continued to work on improving the performance of microchips under development and improving the parameters of existing processors.
In 2002, Intel introduced the Pentium 4 processor, which for the first time reached an operating frequency of 3.06 GHz. Subsequent processors will also support Hyper-Threading technology. Simultaneous execution of two threads gives processors with Hyper-Threading technology a performance boost of 25-40% compared to conventional Pentium 4 processors. This inspired programmers to develop multi-threaded programs, and set the stage for the emergence of multi-core processors in the near future.
In 2003, AMD released the first 64-bit Athlon 64 processor (codenamed ClawHammer, or K8).
Unlike the Itanium and Itanium 2 64-bit server processors, which are optimized for the new 64-bit software system architecture and are rather slow running traditional 32-bit programs, the Athlon 64 is a 64-bit extension of the x86 family. Some time later, Intel introduced its own set of 64-bit extensions, which it called EM64T or IA-32e. The Intel extensions were almost identical to the AMD extensions, which meant they were compatible at the software level. Until now, some operating systems call them AMD64, although competitors prefer their own brands in marketing documents.
In the same year, Intel released the first processor with L3 cache, the Pentium 4 Extreme Edition. A 2 MB cache was built into it, the number of transistors was significantly increased and, as a result, performance was increased. The Pentium M chip for portable computers also appeared. It was conceived as an integral part of the new Centrino architecture, which was supposed, firstly, to reduce power consumption, thereby increasing battery life, and secondly, to provide the possibility of producing more compact and lightweight cases.
To make 64-bit computing a reality, 64-bit operating systems and drivers are required. In April 2005, Microsoft began distributing a trial version of Windows XP Professional x64 Edition that supports additional AMD64 and EM64T instructions.
Without slowing down, AMD in 2004 releases the world's first dual-core x86 processors Athlon 64 X2.
At that time, very few applications could use two cores at the same time, but in specialized software, the performance gain was quite impressive.
In November 2004, Intel was forced to cancel the 4 GHz Pentium 4 model due to heat dissipation problems.
On May 25, 2005, the Intel Pentium D processors were demonstrated for the first time. There is not much to say about them, except perhaps only for a heat dissipation of 130 watts.
In 2006, AMD introduces the world's first 4-core server processor, where all 4 cores are grown on a single chip, and not "glued" from two, as in business colleagues. The most complex engineering problems have been solved - both at the development stage and in production.
In the same year, Intel changed the name of the Pentium brand to Core and released the dual-core Core 2 Duo chip.
Unlike the NetBurst architecture processors (Pentium 4 and Pentium D), the Core 2 architecture did not focus on increasing the clock speed, but on improving other processor parameters, such as cache, efficiency, and the number of cores. The power dissipation of these processors was significantly lower than that of the Pentium desktop line. With a TDP of 65W, the Core 2 had the lowest power dissipation of any desktop microprocessor then commercially available, including Prescott (Intel) cores with a TDP of 130W and San Diego (AMD) cores with a TDP of 89 W.
The first desktop quad-core processor was the Intel Core 2 Extreme QX6700 clocked at 2.67 GHz with 8 MB L2 cache.
In 2007, the 45nm Penryn microarchitecture was released using lead-free Hi-k metal gates. The technology was used in the Intel Core 2 Duo processor family. Support for SSE4 instructions has been added to the architecture, and the maximum amount of L2 cache for dual-core processors has increased from 4 MB to 6 MB.
In 2008, the next generation architecture, Nehalem, was released. The processors have an integrated memory controller that supports 2 or 3 DDR3 SDRAM channels or 4 FB-DIMM channels. The FSB bus was replaced by a new QPI bus. The L2 cache has been reduced to 256 KB per core.
Soon, Intel moved the Nehalem architecture to a new 32nm process technology. This line of processors was named Westmere.
The first model of the new microarchitecture was Clarkdale, which has two cores and an integrated graphics core, manufactured using a 45-nm process technology.
AMD has tried to keep up with Intel. In 2007, she released a new generation of x86 microprocessor architecture - Phenom (K10).
Four processor cores were combined on one chip. In addition to L1 and L2 cache, the K10 models finally received 2MB L3. The size of the data and instruction cache of the 1st level was 64 KB each, and the cache of the 2nd level was 512 KB. There is also promising support for a DDR3 memory controller. The K10 used two 64-bit controllers. Each processor core had a 128-bit floating point module. On top of that, the new processors worked through the HyperTransport 3.0 interface.
In 2009, a long-term conflict between Intel and AMD corporations related to patent law and antitrust law was completed. So, for almost a decade, Intel used a number of dishonest decisions and techniques that interfered with the fair development of competition in the semiconductor market. Intel put pressure on its partners, forcing them to refuse to purchase AMD processors. Bribery of customers, the provision of large discounts and the conclusion of agreements were used. As a result, Intel paid AMD $1.25 billion and committed to following a certain set of business rules for the next 5 years.
By 2011, the era of Athlons and the competitive struggle in the processor market had already turned into a lull, but it did not last long - already in January, Intel introduced its new Sandy Bridge architecture, which became the ideological development of the first generation Core - a milestone that allowed blue giant to take the lead in the market. AMD fans have been waiting for a response from the Reds for quite a long time - only in October the long-awaited Bulldozer appeared on the market - the return to the market of the AMD FX brand associated with breakthrough processors for the company at the beginning of the century. 
The new AMD architecture took on a lot - confrontation with the best Intel solutions (which later became legendary) cost the chipmaker from Sunnyvale dearly. Already traditional for the Reds, inflated marketing, associated with loud statements and incredible promises, crossed all boundaries - the Bulldozer was called a real revolution, and the architecture was predicted a worthy battle against new products from a competitor. What has FX prepared to win the market?
A bet on multi-threading and uncompromising multi-core - in 2011, AMD FX was proudly called "the most multi-core desktop processor on the market", and this was not an exaggeration - the architecture was based on as many as eight cores (albeit logical ones), each of which accounted for one thread. At the time of the announcement of the architecture, the new FX was an innovative and bold solution against the background of the competitor's four cores, looking far ahead. But alas, AMD has always relied on only one direction, and in the case of Bulldozer, this was by no means the area that the mass consumer was counting on.
The productivity of the new AMD chips was very high, and FX easily showed impressive results in synthetics - unfortunately, the same could not be said about gaming loads: the fashion for 1-2 cores and the lack of support for normal parallelization of cores led to the fact that the "Bulldozer" coped with loads with great creaking where Sandy Bridge did not even feel difficulties. Add to this the two Achilles heels of the series - dependence on fast memory and a rudimentary northbridge, as well as the presence of only one FPU unit for every two cores - and the result is very deplorable. AMD FX was called a hot and sluggish alternative to fast and powerful blue processors, which took only relative cheapness and compatibility with older motherboards. At first glance, it was a complete failure, but AMD has never been squeamish about working on bugs - and Vishera became such a job - a kind of reboot of the Bulldozer architecture, which entered the market at the end of 2012.
The updated Bulldozer was named Piledriver, and the architecture itself added instructions, increased muscle in single-threaded loads, and optimized the operation of a large number of cores, which increased multi-threaded performance. However, in those days, the notorious Ivy Bridge, which only increased the number of Intel admirers, acted as a competitor for the updated and refreshed series of reds. AMD decided to follow the already proven strategy of attracting budget users, overall savings on components and the opportunity to get more for less money (without encroaching on the segment above).
But the funniest thing in the history of the appearance of the most unsuccessful (according to most) architecture in AMD's arsenal is that sales of AMD FX can hardly be called not only a failure, but even mediocre - so, according to the Newegg store for 2016, AMD FX became the second most popular processor -6300 (behind only the i7 6700k), and the notorious leader of the budget red segment FX-8350 entered the top five best-selling processors, slightly behind the i7 4790k. At the same time, even the relatively cheap i5, which was cited as an example of marketing success and “people's” status, fell far behind the time-tested oldies based on Piledriver.
Finally, it is worth noting a rather funny fact, which a few years ago was considered an excuse from AMD fans - we are talking about the confrontation between the FX-8350 and the i5 2500k, which originated at the time of the release of Bulldozer. For a long time, it was believed that the red processor was significantly behind the 2500k chosen by many enthusiasts, but in the latest tests of 2017, paired with the most powerful FX-8350 GPU, it turns out to be faster in almost all gaming tests. It would be appropriate to say "Hurrah, wait!".
Meanwhile, Intel continues to conquer the market.
In 2011, a batch of new processors based on the Sandy Bridge architecture was announced, and then a little later, a batch of new processors based on the Sandy Bridge architecture was released for the new LGA 1155 socket released in the same year. This is the second generation of modern Intel processors, a complete update of the line, which paved the way for commercial success for the company, because there were no analogues in terms of power per core and overclocking. You may remember the i5 2500K - the legendary processor, it overclocked to almost 5 GHz, with the appropriate cooling tower, and is able even today, in 2017, to provide acceptable performance in a system with one, and possibly two video cards in modern games. On the hwbot.org resource, the processor overcame a frequency of 6014.1 megahertz from the Russian SAV overclocker. It was a 4-core processor with a 6 MB level 3 cache, the base frequency was only 3.3 GHz, nothing special, but due to solder, the processors of this generation overclocked very strongly and did not overheat. Also absolutely successful in this generation were the i7 2600K and 2700K - 4 core processors with hyperthreading, which gave them as many as 8 threads. They overclocked, however, they were a little weaker, but they had higher performance, and, accordingly, heat dissipation. They were taken as systems for fast and efficient video editing, as well as for broadcasting on the Internet. Interestingly, the 2600K, like the i5 2500K, is also used today not only by gamers, but also by streamers. We can say that this generation has become a national treasure, since everyone wanted exactly the processors from Intel, which affected their price, not in the best direction for the consumer.
In 2012, Intel releases the 3rd generation of processors, called Ivy Bridge, which looks strange, because only a year has passed, could they really invent something fundamentally new that would give a noticeable performance boost? Anyway, the new generation of processors is based on the same socket - LGA 1155, and the processors of this generation are not much ahead of the previous ones, this is, of course, due to the fact that there was no competition in the top segment. All the same AMD, not to say that it would breathe tightly in the back of the first, therefore, Intel could afford to release processors a little more powerful than their own, because they actually became monopolists in the market. But then another catch crept in, now in the form of a thermal interface under the lid, Intel did not use solder, but some kind of its own, as the people called it - chewing gum, this was done to save money, which brought even more income. This topic simply blew up the network, it was no longer possible to overclock processors to the eyeballs, because they received an average temperature of 10 degrees more than the previous ones, because the frequencies came closer to the border of 4 - 4.2 GHz. Special extremals even opened the processor cover, in order to replace the thermal paste with a more efficient one, not everyone managed to do this without chipping the crystal or damaging the processor contacts, but the method turned out to be effective. However, I can highlight some processors that have been successful.
You may have noticed that I did not mention i3 when talking about the second generation, this is due to the fact that processors of this power were not particularly popular. Everyone always wanted an i5, whoever had the money took the i7 of course.
In the 3rd generation, which we will talk about now, the situation has not changed dramatically.
Successful among this generation can be identified i5 3340 and i5 3570K, they did not differ in performance, everything depended on frequency, the cache was still the same - 6 MB, 3340 did not have the ability to overclock, because 3570K was more desirable, but what one, what the second - provided good performance in games. Out of the i7 on the 1155, this was the only K-index 3770 with 8MB cache and 3.5-3.9GHz. In boost, they usually overclocked it to 4.2 - 4.5 GHz. Interestingly, in the same 2011, a new LGA 2011 socket was released, for which two super-processors i7 4820K (4 cores, 8 threads, with L3 cache - 10 MB) and i7 4930K (6 cores, 12 threads, L3 cache was is equal to as much as 12 MB), what kind of monsters they were - it's hard to say, such a percentage cost 1000 bucks and was the dream of many schoolchildren at that time, although for games, of course, it was too powerful, more suitable for professional tasks.
Haswell comes out in 2013, yes, yes, another year, another generation, traditionally a little more powerful than the previous one, because AMD failed again. Known as the hottest generation. However, this generation of i5s were pretty successful. This is due to the fact, in my opinion, that the guys from Sendik ran to change their, as they thought, outdated processors for a new “revolution” from Intel, which then burned all the “Internets”. Processors overclocked even worse than the previous generation, which is why many still dislike this generation. The performance of this generation was slightly higher than the previous one (by 15 percent, which is not much, but the monopoly does its job), and the overclocking limit is a good option for Intel to give less "free" performance to the user.
All i5s traditionally were without hyperthreading. They worked at a frequency of 3 to 3.9 GHz in boost, you could take any with the “K” index, as this guaranteed good performance, albeit with not very high overclocking. At first there was only one i7 here, it's 4770K - 4 cores 8 threads, 3.5 - 3.9 GHz, a workhorse, but it heats up very much without good cooling, I won't say that it was popular with scalpers, but people who scalped the lid, they say that the result is much better, it takes about 5 gigahertz on the water, if you're lucky. This has been the case for every processor since Sendik. However, this is not the end, in this generation there was such a Xeon E3-1231V3, which, in fact, was the same i7 4770, only without integrated graphics and overclocking. It is interesting in that it was inserted into an ordinary mother with a socket 1150 and was much cheaper than the seventh one. A little later, the i7 4790K comes out and it has an already improved thermal interface, but it's still not the same solder as it was before. Nevertheless, the processor overclocks more than the 4770. There were even talks about cases of overclocking to 4.7 GHz in the air, of course, in good cooling.
There are also "Monsters" of this generation (Haswell-E): i7-5960X Extreme Edition, i7-5930K and 5820K, server solutions adapted for the desktop market. These were the most stuffed processors at that time. They are based on the new 2011 v3 socket and cost a lot of money, but their performance is exceptional, which is not surprising, because the older processor in the line has as many as 16 threads and 20 MB of cache. Pick up the jaw and move on.
In 2015, Skylake comes out, on socket 1151, and everything would be fine and it seems almost the same performance, but this generation differs from all previous ones: firstly, the reduced size of the heat-distributing cover, for improved heat transfer with the cooling system on the processor, and secondly, support for DDR4 memory and software support for DirectX 12, Open GL 4.4, Open CL 2.0, which indicates better performance in modern games that will use these APUs. It also turned out that even processors without the K index can be overclocked, this was done using the memory bus, but this case was quickly covered up. Whether this method works through crutches - we do not know.
There were few processors here, Intel again improved the business model, why release 6 processors, if 3-4 out of the entire line are popular? So we will release 4 processors of the medium and 2 expensive segments. Personally, according to my observations, most often they take i5 6500 or 6600K, all the same 4 cores with 6 MB cache and turbo boost.
In 2016, Intel introduced the fifth generation of processors - Broadwell-E. The Core i7-6950X was the world's first ever 10-core desktop processor. The price of such a processor at the time of the start of sales was $ 1,723. To many, such a move by Intel seemed very strange.
On March 2, 2017, the new processors of the older AMD Ryzen 7 line went on sale, which included 3 models: 1800X, 1700X and 1700. As you already know, on February 22 this year, the official presentation of Ryzen took place, at which Lisa Su stated that engineers exceeded the forecast by 40%. In fact, Ryzen is ahead of Excavator by 52%, and given that more than six months have passed since the start of Ryzen sales, the release of new BIOS updates that increase performance and fix minor bugs in the Zen architecture, we can say that this figure has grown to 60% . Today, the older Ryzen is the fastest eight-core processor in the world. And here another assumption was confirmed. About the ten-core Intel. In fact, this was Ryzen's real and only answer. Intel stole the victory from AMD in advance, like, no matter what you release there, the fastest processor will remain with us anyway. And then at the presentation, Lisa Su could not call Ryzen the absolute champion, but only the best of the eight-core. Such is the subtle trolling from Intel.
Now AMD and Intel are introducing new flagship processors. AMD has Ryzen Threadripper, Intel has Core i9. The price of eighteen nuclear thirty-six in-line flagship Intel Core i9-7980XE is about two thousand dollars. The price of a sixteen-core thirty-two threaded Intel Core i9-7960X processor is $1,700, while a similar sixteen-core thirty-two threaded AMD Ryzen Threadripper 1950X costs about a thousand dollars. Draw reasonable conclusions yourself, gentlemen.
Video for this item.