X-Bit.

Hi, ya'll.

When a company boasts that they're console is 16-bit, 32-bit, or even 128-bit. What does this mean? I'm guessing that this has something to do with the video memory of the console, maybe how many colours it can display at once?

I know it's a basic question, but I've always wondered.

It's supposed to be in relation to the processor of the system I assumed, but I really doubt the current gen systems use a 128-bit CPU in them when PCs only have 64-bit accessible at a consumer level.

Yeah, I justed assumed what with the PSP was boasting that it's 32-bit in your hands.

The PSP in my hands right now boasts alot of things............stupid web browser. =/

Sort of related to a) number of opcodes , how much memory can be read in a single go, that sort of thing

an 8 bit processor (6502 for instance) could read 8 bits of memory in one cycle. (or something- im no real expert) and had 8 bit opcodes.

8 bits=256 (0-255) opcodes max.

16bits=65525 opcodes max. Lots more to do here-
32bits=65525^2 opcodes max.

As you can see, theyre not going to fill all those opcodes anytime soon, so i think now its a simple case of how much the processor can shift each clock cycle.

I doubt they ran out of opcodes in their 32 bit processors- and had to move up to 64 bit ;)

usually most console cpu's are still 32 Bit or even 16 bit. CPUs don't have to be 64 or 128 Bit. But there's a number of special purpose processors such as RISC processors, 3D Processors and especially generic DSPs(digital signal processor) that are using more than 32 Bit. There is a simple reason why they do it. It's not only speed, tho they all benefit of high speed. It's rather floating point accuracy at high speed. Imagine a 24 Bit RGB Color still can hold only 256 shades of grey. you can easily hit that limit and get visual artefacts, like a gradient that isn't smooth anymore after some channel amplification, due to rounding errors etc.

For the same reason they wouldn't edit a cinema movie in 24 Bit RGB, but preferably in something like 48 Bit.

In consoles the 128 are only Chip-internal, this means values are sent to the DSPs in 24 Bit, then they are shifted to higher resolutions, then computed in eg. 128 Bit range and finally converted back to 32 or 24 Bit.

The "X" usually refers to the width of the CPU databus, and in most cases also the size of the on-die registers.

My understanding of x-bit where CPU's are concerned is simply how many bits the CPU can operate on in a single instruction. i.e. the bit-size of the registers. I know hardware manufacturers have taken liberties with this meaning, though. e.g. The Atari Jaguar was called 64-bit but in reality that was simply the sum of all the processors it used, I think.

>When a company boasts that they're console is 16-bit, 32-
>bit, or even 128-bit. What does this mean?

It means that the databus of the CPU is XX bits wide.

EDIT: Doh! FD nailed it...


Andy

FD and Big10p have it. For example a 32-bit processor has 32-bit registers and can access 4GB of RAM. When it reads and writes memory it does so in 32bit (4 byte) chunks this is much better than 2x16bit chunks or of course 4x8bit chunks.
It can also add up, subract, multiply and divide 32-bit numbers with a single instruction which is handy.