DADDY BOB'S COMPUTER Q & A

32 VS. 64 Bit

CPU or processor is the Central Processing Unit, the Pentium or AMD chip that is the brains of any computer.

Bit is short for "binary digit." It is how a computer stores and makes references to data, memory, etc. There are 8 bits in a byte.

A bit can have a value of 1 or 0, sometimes said to be on/set or off/cleared. (Some explanations compare the bit to a little switch that can be turned on indicating a 1 or off indicating a 0.) This is called "binary" code, and all data is streamed throughout the CPU, RAM memory, and hard drive as 1's and 0's. These bits are what your processor uses to do its calculations and data processing. These 1's and 0's are the ONLY thing the CPU understands. 

Here's an example that should help:

Note: In this example I'll compare 8 bits to 16 bits to keep the numbers smaller. The same principal will hold true for 32 and 64 bits.

The decimal number 100 is 1100100 in binary. It uses 7 bits, (count them) and would fit in the 8 bit CPU's registers because they can hold 8 bits. The decimal number 1000 is 1111101000 in binary and it requires 10 bits, so it would NOT fit in a 8 bit CPU's registers. The 8 bit CPU would have to handle it in two steps, taking nearly twice as long to do it. A larger 16 bit CPU could handle it in one pass. This then would make the 16 bit CPU faster than a 8 bit one in this example.

8 bits can hold numbers from 0 to 255 (0-11111111 in binary). 16 bits can hold numbers from 0 to 65,535 (0-1111111111111111 in binary). So you see as the amount of the bits doubles, the amount of numbers that it can hold, represent, or address is squared. It therefore follows that a 32 bit processor can hold or represent numbers from 0 to 4,294,967,295, equivalent in binary to a stream of 32 consecutive 1's. A 64-bit processor can hold or represent numbers from 0 to 18,446,744,073,709,551,615, equivalent in binary to a stream of 64 consecutive 1's.

The reason for the increase in steps like 8-16-32-64 is that the size is increased to the next power of 2 to keep it in even full bytes. Remember since binary only has 2 numbers, 1 and 0, everything is done in 2's or powers of 2. 

   2³ =  8 bits

   2⁴ = 16 bits

   2⁵ = 32 bits

   2⁶ = 64 bits

The more bits a CPU can handle at one time, the larger each packet of data sent to it can be. The more instructions it can do in the same time period, the more efficient it will be doing all kinds of calculations. It can also address more memory allowing for more RAM and a larger hard drive.

The first real personal computers, the Apple, and Commodore, had an 8 bit CPU with a 16 bit address bus. In these early 8 bit computers, the amount of memory a CPU could address was so limited, something called "Bank Switching" was employed, that nearly doubled the memory they could address. "Bank Switching" is rarely used now, so I won't get into how it worked here.

The first IBM PC had a 16 bit CPU with a 32 bit address bus. The current Pentium CPU is a 32 bit CPU with a 64 bit address bus. As of this writing, first quarter of 2010, the latest CPUs are 128 bit with a 256 bit address bus, and it won't stop here.

The more calculations, the larger the numbers, and the more memory the processor can handle and address, the faster it will perform its instructions. The faster it performs its instructions, the more of them it can perform per clock cycle. This will result in an increase in its overall performance.

This performance increase will only become evident if the software it is running was written to take advantage of it.

Click HERE to view or download this article in MS Word format