History of Intel CPUs: How have they evolved until today?

History of Intel CPUs

Since it was founded by Gordon Moore and Robert Noyce on July 18, 1968, more than five decades have passed in which it has become the leading company in the manufacture and design of CPUs for personal computers. That is why we have decided to take a tour of Intel‘s history through its CPUs.

Intel is known for its processors with the ISA x86, which has had several evolutions over time and has become the universal language of all applications that run on our PCs and although other sets of registers and instructions have been appearing to Today the ISA invented by the company that Pat Gelsinger now runs in the late 1970s is still as strong as the first day if not even stronger.

Intel’s early history, before 8086

Gordon Moore

In the first section of this story we are going to discuss the CPUs that Intel made before the arrival of the IBM PC that would lead them to success and to be the largest producer of processors for personal computers in history.

The Intel 4004, the first CPU on a chip

Intel 4004

If you look at any history book they will tell you that the first CPU that Gordon Moore and Robert Noyce designed under the Intel brand was the i4004, which is considered the first complete CPU built in one piece, since before a CPU used to have all its components divided into several different chips.

However, it was not the company’s flagship product at the time, since its capital came from the sale of DRAM and SRAM memory for the minicomputers of the time. At that time there were no personal computers and random access memory was state of the art. Although it was not Noyce and Moore who came to the design table to create the 4004, it was the work of three engineers: Federico Faggin, Ted Hoff, and Stanley Mazor who designed a complete 4-bit CPU for Busicom, a company of electronic calculators from Japan that had commissioned the project from Intel.

Thus, the one that at that time was an emerging company compared to other giants of the time such as Fairchild Semiconductor or Texas Instruments was responsible for creating the first complete CPU on a chip.

The Intel 8008 and the Datapoint Terminal

The i4004 experiment served Intel to validate itself for the rest of the industry, but they were still not the giant they would become years later. The era of the PC had not yet arrived and the few computer users continued to use time-sharing terminals connected to a minicomputer, which, paradoxically, could be as big as a refrigerator.

One of these companies was Datapoint Corporation and its 2200 terminal model had a particularity and that is that its circuitry not only served as a remote terminal, but also had a complete computer, since it could execute programs thanks to the fact of having the integrated circuitry of a CPU inside. Again the three architects in charge of the i4004 were tasked with reducing all the complex 8008 circuitry that you see above these images on the i8008.

Unfortunately the project with Datapoint did not come to fruition and in the end the relations between both companies ceased completely. This did not stop them from finishing their first 8-bit CPU, but Intel was not very happy with the result.

The Intel 8080 and the S-100 boom

Intel 8080 Historia Intel

Intel’s third processor was born out of fact: the 8008 was too limited to use and run complex programs. That is why Intel considered another 8-bit processor that could have the ability to run any computer program and from there the i8080 was born.

Again Federico Faggin and Stan Mazor got back to work, this time helped by Masatoshi Shima who had already collaborated with them in the design of the 4004. As a curiosity, Faggin would leave Intel in 1974 to found Zilog and create an improved version in the form of of the famous Z80 that was the central CPU of systems like the Spectrum, MSX, Amstrad CPC and many others.

The 8080 was not a CPU built from scratch, it was built from the design of the 8008. They kept the interrupt system, but improved the addressing from 16 KB of maximum RAM to 64 KB, added more peripheral ports and new instructions. As for the clock speed, it rose from 0.5 MHz to 2 MHz, making it four times faster than its predecessor.

Altair 8800 8080

This CPU stood out for being the main processor of the Altair 8800 created by the Massachusetts Technical Institute that brought with it the first generation of personal computers called S-100 due to the fact that the different components of the computer were connected to a common board that used an interface of the same name. The operating system for the MITS machine was the CP / M created by Gary Kildall, the forerunner of MS-DOS.

Thanks to the use of the S-100 bus as standard, clones of the Altair 8800 soon appeared everywhere, unfortunately the existence of the improved Z80 left Intel out of the leadership of these forerunners of personal computers.

The road to the Intel 8086

Microscopio 8086

There is a belief that the Intel 8086 was a processor made by Intel at the request of IBM for its first PC, which is false, since it already existed for two years before and its history has nothing to do with the first personal computer of the back then blue giant. What’s more, it was initially a minor project within Gordon Moore’s company, as the 8800, released as iAPX 432, was much better on paper with a 32-bit ALU, an integrated MMU, and everything you need to do it. multitask.

To make matters worse, the three musketeers who had created the first three Intel CPUs left, it was a lone Stephen Morse who was chosen to design the 8086 for two reasons: He was an electrical and software engineer. So he knew both worlds perfectly. However, the 8086 had a problem and that is that it started from a totally new ISA and the binary was not compatible. If the programs for CP / M suited well then it was much faster than the Z80, but no one was willing to do it.

Morse left the company shortly after, the 8086 had become Intel’s biggest failure to date, or not.

The history of Intel in the 80s: the beginnings of the PC

In 1981 there were already personal computers in stores, so IBM did not invent the personal computer, but little by little they were a threat to the blue giant and that is why they decided to create a low-budget project called Project Chess, based on use existing hardware and for this they would take as an example the S-100 computers derived from the Altair 8800.

At that time the most advanced CPU for these computers was the 68000, but this CPU was not ready for mass production. The solution? It came from Intel, which had an Intel 8086 and its 8-bit version, the 8088. In order to get the contract for the CPU of the first IBM PC, they had to accept that there would be a second supplier of the processor, a certain company called AMD . Overnight, a project that had been the biggest failure in its history became its biggest success.

Intel 8088

Both the 8088 and 8086 were the first CPUs with the x86 ISA, they were also the first processors in a home system to have a 16-bit ALU. The difference between the two models? The 8088 had an 8-bit and not a 16-bit data bus. Although what stood out the most was its 20-bit memory addressing, which allowed a system to have up to 1 MB of memory, 16 times more than what the 8080 supported.

The 80286, the biggest IPC jump from an x86 CPU

Intel 80286

The success of the IBM PC was overwhelming, already in 1984 it had become the platform par excellence as a personal computer for businesses all over the world, but it was necessary to create a faster version that came with the creation of the AT PC and this time yes that IBM in total trust with Intel asked him to make a new CPU fully backward compatible at the binary level, but much faster.

The 80286 was the biggest jump in IPC that has been made in the history of Intel CPUs, since its performance per clock cycle was more than double that of the 8086. To achieve this, its internal buses were improved that stopped these multiplexed, the process of capturing instructions was optimized and the functionality of the jump instructions were improved. In addition, Intel for the first time included an MMU, which worked by segments.

As a curiosity, it is the longest-lived CPU for PC in history, since even in 1990 you could find computers for sale with this CPU.

Finally the 32 bits, the 80386

Intel 80386

The 80386 was not the first 32-bit CPU in Intel’s history, but the iAPX 432 has the honor of being, however, when it came time to create a 32-bit version of the 8086 Intel took many concepts from that project, although the biggest challenge was moving the code. The solution? Use the same registers, but extended in size up to 32 bits. Additionally, the 80386 was the first CPU in Intel’s history with a segmented pipeline, which was three-stage.

Although the great challenge had to do with multitasking, the idea of having a processor that can run an operating system capable of running several programs at the same time on a personal computer or workstation. For this they designed a new MMU, this time supporting virtual memory per paging, a method that is used today.

Thanks to this CPU, the PC took a leap in quality as a platform, since it brought the news that allowed Windows from its third version to stop being a joke, as well as Linux and above all it marked the dominance of the PC over others platforms.

The 80486, a minor but important update

80486

The fourth generation of x86 processors in the history of Intel brought us two important novelties, the first of them the first level caches of data and instructions. The second was the integration of the floating point unit, which was no longer a separate coprocessor to become an integral part of the CPU. Its pipeline or elongated segmentation at 5 cycles allowed it to reach 100 MHz with the DX4 model.

It was also the nail in the coffin of rival Motorola, who ended up teaming up with IBM and Apple to create the PowerPCs. The reason? Intel was not satisfied with a slightly improved version of the 386 and repeated the move of the 286 with an incredible increase in the IPC, which certified the greater power of the PC compared to its competitors, but especially gave Intel the final victory over Motorola.

As a curiosity, its chief architect was Pat Gelsinger, the current CEO of Intel.

The Intel Pentium, the first multimedia CPU in history

Intel Pentium

The Intel Pentium was the first superscalar CPU under ISA x86, which brought with it the ability to execute 2 instructions in parallel and at the same time. Thanks to this, the CPI compared to 80486 improved by 40% in the process.

Years later, Intel launched the Pentium MMX, which made a change in the floating point unit, since it implemented the SIMD over register in it to accelerate the incipient multimedia programs of the time. But the commercial life of the Pentium MMX was short, since in 1997 when they were launched they coincided in the market with the Pentium II, just a year after the improved version of Intel’s fifth generation x86 CPU.

At the architectural level the changes were minor compared to the 80486, but the design returned again to the hands of those who had created the original 80386 who finished the design in 1992. Outside of its superscalar design it was very conservative, after all. , the company had ceased to have competition, however, this did not mean that they had stood idly by.

The RISC assassination, the P6 era

It is said that in each decade of history Intel makes a great change in its CPUs, in the 80s it was the 80386 that updated the ISA to 32 bits and in the 90s it was the Pentium Pro, a processor that was thought as a spearhead from x86 to the workstation market, which was dominated by RISC architectures at the time.

From Intel they knew very well that the x86 instruction set had limitations, that is why Intel created both out-of-order and speculative execution for this processor. In addition, it increased the number of stages from 5 to 14 and added the second-level cache for the first time.

Intel Pentium II

Under the architecture of the Pentium Pro or P6 came various CPUs, released as generations of different commercial products.

  • The Pentium Pro was the first CPU to have the L2 cache built into the processor, until then it used to be included on the board near the processor.
  • As for the Pentium II, it was based on the Pentium Pro, but it moved the L2 cache out, although it left it in the same package, unlike its predecessor, it was launched for the home PC market, bringing the power of the work stations.
  • The Pentium III on the other hand included the SSE instructions and eventually ended up integrating the L2 cache back into the processor.

Intel’s strategy worked and in the late 90’s most RISC architectures languished completely awaiting their final death. Only ARMs and the PowerPCs used in the Macintosh survived by far, the rest had the sword of Damocles on them and would soon succumb.

Pentium 4, the end of an era

Intel Pentium 4

For the Pentium 4, Intel created a new architecture called Netburst, which followed the trend of the day to add a large number of stages to achieve a high clock speed. It was with this processor that Intel hit the speed ceiling and it was discovered that the race based on that metric had no future, due to the high consumption of the processors and the temperature they generated.

It was from the experience of the Pentium 4 that the metric “power per watt” came to have importance and began to design CPUs already based on the multicore concept. Especially the reason was because they were unfeasible to be mounted on laptops and had had to extend the life of the P6 to be able to launch CPUs for that type of computer that was in the making at that time.

The gossips say that it was one Apple, obsessed with its industrial designs that gave Intel a pulse on whether it could create a core with enough performance per watt and with more power than the PowerPC, only then would it make the leap to the x86. And so it was, but with it they also abandoned the Pentium name to adopt another.

The multi-core era of Intel history

Intel Core 2 Historia Intel

The first Intel Core were based on the P6 architecture, but with a dual-core configuration. However, they had developed a second generation that was revolutionary and is considered the best CPU of the 2000s.

For its development Intel copied several ideas to the Opteron of AMD; as is the implementation of the Northbridge within the processor and the adoption of the x86 64-bit extension. In terms of performance, they almost doubled the IPC and it is the third largest performance jump after the one achieved with the 80286 and 80486. To do this, they improved out-of-order execution, made the CPU capable of handling more instructions in parallel and they first added the Smart Cache on Intel processors.

Intel-Core-i9-12900K-alder-lake

However, the Core 2 could be considered the zero generation, since Intel began to use the Intel Core i3, i5 and i7 brands from the Nehalem architecture, considered as the first generation Intel Core. Since then until now we have had several generations with gradual improvements.

  • Sandy / Ivy Bridge: Intel again improved the jump prediction unit, apart from this it improved elements such as the micro-op cache, the integer and floating point units and the performance of certain instructions for fetching data from memory.
  • Haswell / Broadwell: Intel again expanded the number of instructions that the CPU can execute per cycle, in addition to increasing the bandwidth of the internal processor caches and improving the memory controller. They also included the In-CPU Voltage Controller (FIVR).
  • SkyLake generation: Intel improved the number of instructions the CPU can decode, but did not increase the number of instructions it can perform in parallel. The changes compared to previous generations are very small (they removed the FIVR).
  • Rocket Lake-S / Tiger Lake: It is the current Rocket Lake-S and Tiger Lake, After years with slight improvements in the IPC Intel has decided to follow the path of AMD not to be left behind.

Its latest release is the Intel Core 12 with Alder Lake-S architecture, which is the one that adds more novelties since the launch of Intel Core 2 such as execution with heterogeneous cores, the addition of the Thread Director and other novelties, we do not know if It will be the great generational leap that the company makes in each decade or there is something more under the hood, what is clear is that the history of Intel is at least exciting.