Monday, September 6, 2010

In which I attempt to design a CPU, part 2

At this point, we'd lay out the electronics, grouping similar parts together. All the math circuits are near the logic circuits, combining to form an ALU, Arithmetic and Logic Unit. The Control circuits are nearby, and with our leftover space, we have little chunks of super-fast, super-expensive, memory, called Registers.
This is typically so complex that developers turn to software tools, like Verilog, which takes a list of the requirements, laid out like a C program, and passes them onto a program that lays out a circuit design that accomplishes what was specified. Then that can be given to your fabricator, who cranks out the chips.
But before you crank them out, first you get one, and you test it. No sense in paying for defective chips, right? In fact, that's part of the reason for all the computer support. They can simulate your chips before you even make them, so you know your design is good before you've made even one. Then you make one and test that to prove that the simulation was accurate.

3 comments:

Anonymous said...

Perhaps you should attempt to describe how to make a quantum computer? Never understand how you can make a computer work by moving atoms up and down in place of transitors with 0 & 1

themadengineer said...

Alright, I'll go write up an article, and post it when it's done.

Anonymous said...

coolio

Related Posts Plugin for WordPress, Blogger...