Code_ The Hidden Language of Computer Hardware and Software - Charles Petzold [98]
Of course, these three numbers indicate only roughly how fast the microprocessor operates. These numbers tell you nothing about the internal architecture of the microprocessor or about the efficiency and capability of the machine-code instructions. As processors have become more sophisticated, many common tasks previously done in software have been built into the processor. We'll see examples of this trend in the chapters ahead.
Even though all digital computers have the same capabilities, even though they can do nothing beyond the primitive computing machine devised by Alan Turing, the speed of a processor of course ultimately affects the over-all usefulness of a computer system. Any computer that's slower than the human brain in performing a set of calculations is useless, for example. And we can hardly expect to watch a movie on our modern computer screens if the processor needs a minute to draw a single frame.
But back to the mid-1970s. Despite the limitations of the 4004, it was a start. By April 1972, Intel had released the 8008—an 8-bit microprocessor running at 200 kHz that could address 16 KB of memory. (See how easy it is to sum up a processor with just three numbers?) And then, in a five-month period in 1974, both Intel and Motorola came out with microprocessors that were intended to improve on the 8008. These two chips changed the world.
Chapter 19. Two Classic Microprocessors
The microprocessor—a consolidation of all the components of a central processing unit (CPU) of a computer on a single chip of silicon—was born in 1971. It was a modest beginning: The first microprocessor, the Intel 4004, contained about 2300 transistors. Today, nearly three decades later, microprocessors made for home computers are approaching the 10,000,000 transistor mark.
Yet what the microprocessor actually does on a fundamental level has remained unchanged. While those millions of additional transistors in today's chips might be doing interesting things, in an initial exploration of the microprocessor they offer more distraction than enlightenment. To obtain the clearest view of what a microprocessor does, let's look at the first ready-for-prime-time microprocessors.
These microprocessors appeared in 1974, the year in which Intel introduced the 8080 (pronounced eighty eighty) in April and Motorola—a company that had been making semiconductors and transistor-based products since the 1950s—introduced the 6800 (sixty-eight hundred) in August. These weren't the only microprocessors available that year. Also in 1974, Texas Instruments introduced the 4-bit TMS 1000, which was used in many calculators, toys, and appliances; and National Semiconductor introduced the PACE, which was the first 16-bit microprocessor. In retrospect, however, the 8080 and the 6800 were certainly the two most historically significant chips.
Intel set the initial price of the 8080 at $360, a sly dig at IBM's System/360, a large mainframe system used by many large corporations that cost millions. (Today you can buy an 8080 chip for $1.95.) It's not as if the 8080 is comparable to System/360 in any way, but within a few years IBM itself would certainly be taking notice of these very small computers.
The 8080 is an 8-bit microprocessor that contains about 6000 transistors, runs at a 2 MHz clock speed, and addresses 64 kilobytes of memory. The 6800 (also selling these days for $1.95) has about 4000 transistors and also addresses 64 KB of memory. The first 6800 ran at 1 MHz, but by 1977 Motorola introduced later versions running at 1.5 and 2 MHz.
These chips are referred to as single-chip microprocessors and less accurately as computers on a chip. The processor is only one part of the whole computer. In addition to the processor, a computer