» System Architecture - A look at the CPU from its inception «

The central processing unit (CPU) is the "brain" of the computer. Its function is to execute programs stored in the main memory by fetching their instructions, examining them, and then executing them one after another.

The development of the microprocessor is directly related to the development of the transistor. William Shockley, of Transistor fame, started a company known as Shockley Research Laboratories in New Jersey. He later moved to California to start Fairchild Electronics. Later, in 1968, two top Fairchild Electronics executives formed their own company (with some investment help from IBM) known as Integrated Electronics; or Intel for short!

Register

Used as the implicit source and destination of an operation (the register doesn't have to be specified separately). RISC processors use a load/store architecture instead - to add memory to a register, it must be loaded into an intermediate register first.

Cache

A small amount of fast memory which holds recently accessed data or instructions so that if they are used by the programs again, the cache can supply them transparently faster than main memory. Cache memory is typically organised into lines (several bytes are loaded at once, on the assumption that nearby memory will be used next). The lines are organised into sets, each set is mapped to a separate group of memory addresses, and there are usually between two and sixty-four lines per set (fewer lines per set are simpler, but access to more addresses than cache lines in the same set can cause data in the cache to be discarded before it can be used).

Smaller caches are faster, so often a small level 1 cache is used, with a larger but slower level 2 cache supporting it. Level 3 caches can even be used in some cases. Some cache controllers monitor the memory bus to detect when a cached memory value has been modified by another CPU, or a peripheral.

ALU (Arithmetic and Logic Unit)

The smart part of a processor chip that performs commands like adding, subtracting, multiplying and dividing. It also knows how to read logic commands like OR, AND, or NOT. Messages from the Control Unit instruct the ALU what to do and then it takes data from its close companion, the Registers, to perform the task. The source of the operands and the destination of the result.

The width in bits of the words which the ALU handles is usually the same as that quoted for the processor as a whole whereas its external busses may be narrower. Some processors use the ALU for address calculations (e.g. incrementing the program counter), others have separate logic for this.

The Registers are a mini-storage area for data used by the Arithmetic Logic Unit (ALU) to complete the tasks the Control Unit has requested. The data can come from the data cache, main memory or the control unit and are all stored at special locations within the Registers. This makes retrieval for the ALU quick and efficient. The limited size and high speed of the register set makes it one of the critical resources in most computer architectures. Register allocation, typically one phase of the back-end, controls the use of registers by a compiled program.

FPU (Floating Point Unit)

A formal term for the math co-processors (also called numeric data processors, or NDPS) normally integrated with the CPU rather than packaged and sold separately. FPUs perform certain calculations faster than CPUs because they specialise in floating-point math, such as addition, multiplication, logarithms, exponentials, trigonometric functions and various kinds of rounding and error detection, whereas CPUs are geared for integer math.

Control unit

The Control Unit is one of the most important parts of the microprocessor because it is in charge of the entire process. The control unit it fetches instructions from memory and decodes them to produce signals which control the other part of the computer. Based on instructions from the Decode Unit, it creates control signals that tell the Arithmetic Logic Unit (ALU) and the Registers how to operate, what to operate on, and what to do with the result. The Control Unit makes sure everything happens in the right place at the right time.

CPU : 4004 to 8088

The 4004 and 4040 were 4-bit processors and Intel's first venture into microprocessors. This breakthrough invention was used in the Busicom calculator, then in other low-end devices. It paved the way for embedding intelligence in inanimate objects as well as the personal computer.

The 8008 was twice as powerful as the 4004. It was designed for use as a Datapoint CRT controller It was mainly used to give some intelligence to hardware. One application was a serial terminal where the 8008 co-ordinated the terminal's activities.

According to the magazine Radio Electronics, Don Lancaster, a dedicated computer hobbyist, used the 8008 to create a predecessor to the first personal computer, a device Radio Electronics dubbed a "TV typewriter."

The 8080 was Intel's first useful 8-bit processor. It was quite capable and was ultimately used in MANY early personal computers.

The 8085 was simply an improved 8080 with the features which I've already mentioned. Note that it was still an 8-bit processor.

The 8086 was Intel's first venture in 16-bit computing. It had a totally different instruction set, different hardware architecture, and much greater capability. It was in no direct way related to ANY of the previous devices (including the 8085). It just happened to be the next number in sequence.

The 8088 was an 8086 internally with an 8-bit data bus. It was a 16-bit processor but had an 8-bit external bus to make connection to older hardware less difficult. It also had no direct relation to previous processors. The 8088 was used in the IBM PC and XT and most XT-class clones. A pivotal sale to IBM's new personal computer division made the 8088 the brains of IBM's new hit product--the IBM PC. The 8088's success propelled Intel into the ranks of the Fortune 500, and Fortune magazine named the company one of the "Business Triumphs of the Seventies."

CPU : 80186 (and 80188)

The 80186 was an enhanced 8086 with some added instructions. One major function was the reduced number of chips required. Its features included 2 programmable DMA controllers, 3 timers, a PIC (Programmable Interrupt Controller), and 32 bit watchdog timer. Enhanced military versions were able to operate at extremes of temperature (-25 to 125 Celsius)


"To satisfy this market, we defined a processor with a significant performance increase over the 8086 that also included such common peripheral functions as software-controlled wait state and chip select logic, three timers, priority interrupt controller, and two channels of DMA (direct memory access). This processor, the 80186, could replace up to 22 separate VLSI (very large scale integration) and TTL (transistor-transistor logic) packages and sell for less than the cost of the parts it replaced."
-- Paul Wells of Intel Corporation writing in Byte Magazine


This is the chip that never saw life in PC's - or did it ?



Actually, yes, contrary to popular belief, it did. With apologies to the reference books that say otherwise, a handful of companies actually used it. The long gone Mindset was one and Tandy was another with their TRS-80 Model 2000.



I seem to recall there was something queer about how it interacted with video and as a result it was side-lined into use as an embedded controller chip. As an embedded processor however it excelled and is still widely used today in a vast number of devices. These include(d) intelligent peripherals like disk controllers, high end network switches and controllers, in-circuit emulators and intelligent multi-port serial adapters.

For example:

• Cybron Technologies, Malaysia use it as a power for monitoring device for their "Nucleus RTX" ISDN board.
• Plessey Naval Systems Ltd used it in the design and development of a new sonar for the Royal Navy. The main task areas were: the development of high integrity embedded software, targeted to a 80186 based processor card used to control a large mechanical winch. This safety critical software required continuous fault monitoring while actioning operator requests passed via RS-422 serial links.
• The Licom Corporation used it for a synchronous M1-3 multiplexer controller card and telecommunications switching devices.
• VAutomation's V186-ES1 Evaluation Silicon allows designers of application-specific integrated circuits (ASICs) to easily evaluate and verify its Intel 80186-compatible core, known as V186, and to quickly prototype their designs containing the core. They offer a suite of synthesisable cores for microprocessors and communications controllers. Their V186 allows an Intel 81086-compatible microprocessor to be implemented within an ASIC for true system-on-chip integration.

And, as a complete revival, of sorts…
XT-CE A 80186 PC Emulator for the Windows CE Operating system.

The XT-CE PC/XT emulator for Windows CE emulates a PC with an 80186 processor, 1Mb of RAM, a CGA-compatible display and the standard support chips. This allows most applications designed for PC-compatible computers running MS-DOS to run on the H/PC device. As XT-CE does not emulate DOS (or any other operating system), one needs to be provided (much like a real PC). In conjunction with running a 'real' DOS, XT-CE is able to provide the same level of application compatibility as the original IBM PC/XT, while still allowing the H/PC user access to all of the features and applications of the Windows CE operating system.

CPU emulation: Intel 80188/80186

CPU : 80286 to Pentium III

The 80286 was the next major jump in the 80x86 line. Its instruction set was expanded to include some additional sophisticated 16-bit instructions. The 286 was the first Intel processor that could run all the software written for its predecessor, the 8086. This software compatibility remains a hallmark of Intel's family of microprocessors. Within 6 years of it release, there were an estimated 15 million 286-based personal computers installed around the world.

While still a 16-bit unit, the 80286 could address 16 Megabytes of memory with its 24-bit address bus. It also offered a new mode of memory addressing, called protected mode, which expanded the addressing to allow more memory to be accessed. It would boot in the 8088/86 mode (called real mode), but could switch to protected mode if instructed to access the additional memory.

It also had multitasking protection built in, however, its inability to rapidly switch from protected mode to real mode inhibited its ability to fully support multitasking Operating Systems such as Windows or OS/2.


The 80386 was the Intel's first popular 32-bit processor. It included full 8086 and 80286 compatibility but added some extra modes including a software switch-able protected mode and a virtual machine mode. It had larger addressing and also 32-bit instructions.

This also took advantage of a scalar architecture, meaning that it offered multiple instruction pipelines and could start working on a new instruction before the last instruction was completed. The 80386, like the 80286, boots up in real mode but offered efficient switching between real mode and protected mode. It could even support multiple real mode (virtual) sessions.


The 80486 was an enhanced 386 and contained an onboard static RAM cache that reduced memory access times. It also include an onboard FPU or Floating Point Processor (also known as a "math coprocessor"), which speeds up computing because it offers complex math functions from the central processor.


The Pentium retained the 32 bit address bus from the 80486, but doubled the data bus to 64 bits. Again, it was fully backwards compatibility, but featured many new features including some RISC-like features and some dual pipelined or superscalar architecture for concurrent processing of instructions. They included caches, pipelining, anticipatory fetching, predictive branching, and many other things.

The Pentium processor allowed computers to more easily incorporate "real world" data such as speech, sound, handwriting and photographic images. The name Pentium®, mentioned in the comics, on television talk shows and a truly massive promotional campaign became a household word soon after introduction.

Later versions included MMX technology, which is designed specifically to process video, audio and graphics data efficiently. Despite the hype in offer little extra benefits unless specifically programmed for the extra instructions. (Initially reported as MultiMedia eXtension, but later said by Intel to mean Matrix Math eXtension).

The Pentium Pro was the new P6 processor and was packaged together with a second speed-enhancing cache memory chip and many more advanced RISC-type features which supersede those found in the much cheaper Pentium.

The Pentium® II processor is packaged along with a high-speed cache memory chip in an innovative Single Edge Contact (S.E.C.) cartridge that connects to a motherboard via a single edge connector, as opposed to multiple pins. With the aid of this powerful chip, PC users could capture, edit and share digital photos with friends and family via the Internet; edit and add text, music or between-scene transitions to home movies; and, with a video phone, send video over standard phone lines and the Internet.

The horrendously expensive Pentium II Xeon is aimed at the server market and incorporates extra on chip cache

Fall of 1999 saw the introduction of Coppermine, an enhanced new version of the Pentium III. Built on a 0.18-micron process, the chip includes 256Kb of on-chip L2 cache that runs at the speed of the processor. Power users await the release of Cascades, a server version of Coppermine, (due in late 1999) which offeres more L2 cache and different packaging.

Beyond the Pentium III

The immediate future offers Merced - now Itanium - Intel's first full 64 bit processor. Scheduled to ship Mid 2000, and built using a 0.18-micron process it will debut at about 800 MHz (perhaps as high as 1200Mhz!) and will execute 6 to 8 instructions per cycle.

Late 2000 will see Willamette, Intel's seventh-generation processor, also built using a 0.18-micron process and promising to deliver both higher clock speeds and higher performance per clock cycle than the Pentium III. Foster, the server version, will follow.

MID-2001 promises the McKinley, Intel's second IA-64 processor, which should be twice as fast as Merced, using the same 0.18-micron process, and is likely to be the chip that gives IA-64 its real start. But IA-64 CPUs won't have any role in PCs until a low-cost derivative of McKinley, code-named Deerfield, is introduced in 2002.