Download Processors - CVCC Online

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Microprocessor wikipedia , lookup

Multi-core processor wikipedia , lookup

Transcript
Processors
A Computing
Research Project
Louis Sullivan
Processors: the construction process
Processors are mainly made using silicon, which is found in beach sand. Silicon is a natural
semiconductor, and the most abundant element on Earth, excluding oxygen. This means that
processors can continue to be made, as silicone is so abundant, and it’s a natural conductor, which
makes it perfect for computer parts.
To begin with, silicon is gathered by extraction companies and shipped to factories. Here, the silicon
is purified, melted, and cooled to form an ingot, which is then sliced into discs called wafers. Chips
are built simultaneously in a grid formation on the wafer surface in a fabrication facility or “fab.” A
chip is a complex device that forms the brains of every computing device. While chips look flat, they
are three-dimensional structures and may include as many as 30 layers of complex circuitry.
Processors are fabricated in batches of wafers in clean rooms that are thousands of times cleaner
than hospital operating rooms. Fab technicians wear special suits, nicknamed bunny suits, designed
to keep contaminants such as lint and hair off the wafers during chip manufacturing.
The "recipe" for making a chip varies depending on the chip’s use. Making chips is a complex process
requiring hundreds of precisely controlled steps that result in patterned layers of various materials
built one on top of another. A photolithographic "printing" process is used to form a chip’s multilayered transistors and interconnects (electrical circuits) on a wafer. Hundreds of identical
processors are created in batches on a single silicon wafer. Once all the layers are completed, a
computer performs a process called wafer sort test. The testing ensures that the chips perform to
design specifications.
After fabrication, it's time for packaging. The wafer is cut into individual pieces called die. The die is
packaged between a substrate and a heat spreader to form a completed processor. The package
protects the die and delivers critical power and electrical connections when placed directly into a
computer circuit board or mobile device, such as a smartphone or tablet.
These packages have to undergo final testing for functionality, performance, and power. The chips
are electrically coded, visually inspected, and packaged in protective shipping material, ready to be
sold.
The Architecture of a Processor
The way a chip works is the result of how a chip’s transistors and gates are designed and the
ultimate use of the chip. Design specifications that include chip size, number of transistors, testing,
and production factors are used to create schematics—symbolic representations of the transistors
and interconnections that control the flow of electricity though a chip.
Designers then make stencil-like patterns, called masks, of each layer. Designers use computer-aided
design (CAD) workstations to perform comprehensive simulations and tests of the chip functions. To
design, test, and fine-tune a chip and make it ready for fabrication takes hundreds of people.
The architecture defines what the chip will do, and affects how effective the chip is at different
tasks. Depending on how the chip is mapped out, it will be able to do different things better than
other processors, whose architecture is different. This doesn’t necessarily mean that it’s worse, it’s
just specialised in a different task.
Size and Speed
The speed of a processor depends on many different factors. Here are the main ones.
Clock Speed
Clock speed, basically, is how many fetch, decode, execute cycles it can do in one second. It is
measured in Hertz, although it is more common to see GHz due to the speed of modern processors.
1.8 GHz is equal to 1,800,000,000 clock cycles per second.
Number of Cores
Processors have evolved drastically over the years, and the introduction of multi-core processors has
greatly increased speeds of processors. Each core cannot do more than 1 clock cycle at the exact
same time, so a dual-core processor theoretically doubles the clock speed, as each core does the
same amount, at the same speed, but at the same time. Therefore quad-cores are four times as fast.
However, due to bottle-necking and limitations on software, sometimes only one core can be used,
as the program needs to be compatible with multi-core processors. This means that number of cores
is not as important as clock-speed, but definitely helps.
Cache
Cache is a super-fast volatile memory, located in the CPU, that contains the data for any programs
running. Whilst RAM is fast, compared to Cache it is very slow, and so if the processor had to fetch
from RAM all the time the computer would slow down greatly. However, Cache memory is very
expensive to make and buy, and so there is only a very small amount in the CPU. More demanding
programs will have to fetch from the RAM, but frequently used instructions and programs will be
stores, and used, in the cache. Think of it like a bus stop, a half-way point between RAM and the
processor. Ram delivers the most used lines of code to the cache, which is then fetched by the
processor, rather than the processor having to fetch all the way from RAM.
Binary Code
Each and every line of code executed by the processor has to be in machine (binary) code. This
essentially is a load of 0s and 1s; 0s for false, 1s for true. These are then arranged into bytes (8 ‘bits’),
kilobytes (1024 bytes), megabytes (1024 Kb), gigabytes (1024 Mb) et cetera. However, when people
code it is extremely long-winded, difficult and time consuming to code in binary. This is avoided by
writing in a higher-level programming language, such as Python, which is then translated in to binary
by utilities such as compilers, assemblers and linkers.
Even data like pictures and text is stored as Binary. For text we use character-encoding language, like
ASCII, to convert text into binary. For example, ‘A’ in ASCII is ‘100 0001’