Explaining 32-bit vs 64-bit: Understanding Computer Applications and Processors

[sen_almighty]

Can someoe explain what 32bit and 64bit mean when you're talkin bout comp applications or processers or what not.

 

[LogicalAtheist]

Bit - A fundamental unit of information having just two possible values, as either of the binary digits 0 or 1.

7 bits make up a byte.

Think of a byte as the atom of information processing. Things are measured by how many bytes they are. Larger units are megabyte, and gegabyte.

Think of the bit as maybe a quark, or a string.

It's a measure of information, of which the computer computes. Other parts of the computer have measurements of how they can compute it as well.

32bit and 64bit concern usually the video card. Meaning it is able to accept software information written in one of those amounts.

If your card can accept 32bit, then it will take in the software information (from a video game perhaps) that's fit for that much information.

The higher the value, the more information. Thus if you can use 64bit, and the same software, the software will let you use the 64 bit information - which would mean you could see the game in it's higher video quality.

Any specific questions? Lemme know!

 

[sen_almighty]

yes, i know that, but 32bit OS's and software, and 64bit, like the Operton is 64bit and the P4 is 32bit and stuff, wat does that mean?

 

[dav2008]

Quick thing, Logical...I thought it was 8 bits to a byte..

 

[LogicalAtheist]

The software in this case is made to send out X bits in an instant.

Meaning one packet of information is larger. Each 64 bit packet is sent out instantaneously.

So, if the software is made for higher bits, it'll hand more instantaneous information to the processor.

So if your processor cannot process X bits at the same speed it receives them, you'll get lag

 

[LogicalAtheist]

Dav - I may be wrong. I also thought it was 8, but then I thought something was 7.

I think there might be some bytes that have only 7? Someone with a sure knowledge let us both know.

 

[LogicalAtheist]

Originally posted by sen_almighty
yes, i know that, but 32bit OS's and software, and 64bit, like the Operton is 64bit and the P4 is 32bit and stuff, wat does that mean?

Sen - I've never heard of these OS's. What are they all about?

As you can see in the new "longhorn" (code name for new windows) I'm dying for a revolutionary OS. But I feel uncomfortable without windows, heh

I hate desktop icons, start menu, and the stupid bar at the bottom. I want something much better.

I have drawn out advanced GUI ideas in the past, for a programmer. But alas making an OS from scratch is a sickening task.

 

[BoulderHead]

Hello sen_almighty,
Generally these terms are applied to the processors, and they are speaking of how many individual bits the CPU can perform functions on in one clock cycle. The 64-bit processor can operate on twice as many bits as its 32-bit little brother. There are other considerations such as software having been written to transfer a certain number of bits each clock cycle, and this means the jump to 64-bits will necessitate some software revision.

 

[russ_watters]

Some clarifications:

A byte is 8 bits. Sometimes a 9th bit is added for error correction (a checksum) but it isn't part of the byte.

The Opteron is AMD's 64 bit processor (the Itanium is Intel's 64 bit processor).

Nothing in a computer is instantaneous. If a data path is 64 bits, all 64 bits get to their destination at the same time, but it still takes time for them to get there.

In order to use a 64 bit processor, you need a 64 bit operating system (the Opteron actually does BOTH 32 bit and 64 bit). Microsoft is currently working on several flavors of 64 bit OSs.

The benefit of 64 bit is basically just allowing you to use larger numbers. Sometimes it makes applications faster, sometimes it doesn't.

For video cards, 64 bit vs 32 bit is talking about memory bandwidth. Its the same basic idea - information is transferred in groups of 32 or 64 (or 16 or 128) bits but the graphics processors are all still 32 bits. Memory bandwidth is EXTREMELY important for 3d rendering which is why you hear it advertised.

 

[LogicalAtheist]

Originally posted by BoulderHead
Hello sen_almighty,
Generally these terms are applied to the processors, and they are speaking of how many individual bits the CPU can perform functions on in one clock cycle. The 64-bit processor can operate on twice as many bits as its 32-bit little brother. There are other considerations such as software having been written to transfer a certain number of bits each clock cycle, and this means the jump to 64-bits will necessitate some software revision.

Here we have boulder saying in real computer terms what I was trying to say without knowing the terms, thanks boulder
So that "instant" i said is one clock cycle.

So processor can do a certain amount of bits per clock cycle,. and remember a faster prcoessor has fast clock cycles...

and the software says "i want to send you 64 bits a clock cycle" and processor says "no no, i can only do 32bit".

Boulder said it right.

 

[LogicalAtheist]

Originally posted by russ_watters
Some clarifications:

A byte is 8 bits. Sometimes a 9th bit is added for error correction (a checksum) but it isn't part of the byte.

The Opteron is AMD's 64 bit processor (the Itanium is Intel's 64 bit processor).

Nothing in a computer is instantaneous. If a data path is 64 bits, all 64 bits get to their destination at the same time, but it still takes time for them to get there.

In order to use a 64 bit processor, you need a 64 bit operating system (the Opteron actually does BOTH 32 bit and 64 bit). Microsoft is currently working on several flavors of 64 bit OSs.

The benefit of 64 bit is basically just allowing you to use larger numbers. Sometimes it makes applications faster, sometimes it doesn't.

For video cards, 64 bit vs 32 bit is talking about memory bandwidth. Its the same basic idea - information is transferred in groups of 32 or 64 (or 16 or 128) bits but the graphics processors are all still 32 bits. Memory bandwidth is EXTREMELY important for 3d rendering which is why you hear it advertised.

yeah he said it to.

I didn't mean an isntant was an instnat heh. It was me forgetting the damn terms and meaning one unit of time.

Thus we have oru answer.

I want to know more about these 64 bit OS's

Anyone got some links? with pics? thanks!

 

[BoulderHead]

Originally posted by LogicalAtheist
I want to know more about these 64 bit OS's

Anyone got some links? with pics? thanks!

Tey here;

http://www6.tomshardware.com/business/20030506/WinHEC-02.html

http://www.winsupersite.com/reviews/

 

[Adam]

A byte is 8 bits, no more and no less. Sometimes the left-hand bit (MSB) is used for odd or even parity checking (left or right side representation is just convention really). Ninth bits are not added. ASCII codes require 7 bits, and the eighth is often used for a parity check.

PARITY works by simply adding up the number of 1s in any given byte. If you are doing even-parity, you want the end result to be an even number of 1s. If an odd-parity system is being used, you want the total number of 1s to be odd. If your system is using even-parity, the byte 00011001 will be considered erroneous, as it has an odd number of 1s. So we change the MSB to 1 like so: 10011001. Now it has an even number of 1s, so it's ok. Odd is just the reverse, making sure there is an odd number of 1s.

Regarding architectures and OSs and such, 32-bit or n-bit refers to how many bits are transported and processed at any given moment. See below:

8-bit:
----------
----------
----------
----------
----------
----------
----------
----------

16-bit:
----------
----------
----------
----------
----------
----------
----------
----------
----------
----------
----------
----------
----------
----------
----------
----------

Imagine those as wires running through the system. 8-bit is eight wires wide, trasnporting eight bits (1s or 0s in some combination) at a time. 16-bit can move 16 at a time.

 

[damgo]

Yes, it's the size of the data path -- mainly this is important for memory addressing, but arithmetic too. A 32-bit processor/program can only easily access 4 gigabytes of memory, and a 16-bit one only 64 megabytes; anything more and it slows way down. A 64-bit one can access ~16,000,000,000 gigabytes. These also give the largest numbers the program/processor can operate on at once.

 

[Artman]

Originally posted by damgo
Yes, it's the size of the data path -- mainly this is important for memory addressing, but arithmetic too. A 32-bit processor/program can only easily access 4 gigabytes of memory, and a 16-bit one only 64 megabytes; anything more and it slows way down. A 64-bit one can access ~16,000,000,000 gigabytes. These also give the largest numbers the program/processor can operate on at once.

The calculation is 2^64 for the memory that can be addressed. You get approx. what damgo said.

The CPU can handle longer files names, as in the change from Win 3.1 to Win95 allowed longer file names.

The quantity of calculations that can be performed on the data per clock cylce is also effected by pipelining in the CPU (basically the flow path of data in a CPU).

 

[sen_almighty]

wow guys, thanks for the info, i get it now

 

[grady]

The bit number is the "word size" or size of a register. Any processor has registers which are pieces of ram on the CPU itself. The register size is the largest piece of data the CPU can look at at a time, but the CPU has several registers so you can do things like add two number together.

movl $1,%EAX
movl $2,%EBX
addl %EAX,EBX

would put the number 1 in the A register 2 in the B register and then add them together with the result going into the C register (the processor always know to put the result in the C register). A 16 bit assembler would barf at this because it doesn't understand movl or addl, only mov and add.

So for "64 bit OS's" there isn't much magic involved, especially not at microsoft. Most of what this means is that they both compiled the OS on a compiler that produces the correct executable code for the 64 bit processor and changed any assembly language that was part of the OS that needed to be rewritten with the new instruction set that that processor uses. For example the mov and movl instructions.

 

[J-Man]

Originally posted by Artman
The CPU can handle longer files names, as in the change from Win 3.1 to Win95 allowed longer file names.

This has nothing to do with the bus/register size. It is a limit set in the specification for the filesystem. It is more or less an arbitrary limit.