difference between 32 bit and 64 bit computers

What is the difference between 32 bit and 64 bit computers. Is 64 faster? Can it make better resolution pictures?
Can anyone explain it like, in an understandable way? I mean for a low-tech person like me.

It has more to do with storage and security. I’m a little lite on the details but 64 bit allows for more combinations.

32 bit is just fine for graphics and sound. 32 bit graphics is essentially 24 bit true color with an 8 bit alpha channel. The 24 bits are for R, G and B channels.

32 bit float sound is essentially 24 bits of audio processing and 8 bits of post. The 24 bits is three 8 bit layers.

With all computers needing unique mac address 64 bit is probably going to be a must. It’s just a way to enable more combinations at this point.

Well, it seems that a 32-bit OS will only allow a maximum of 4GB(-ish) per application, whereas a 64-bit OS will allow the application to go beyond that limit.

This is obviously useful for very large scenes in a 3D package.

A “32-bit” machine naturally works with integers that are 32-bits wide, and it uses a memory-address that is 32-bits wide. This limits the machine to 4 gigabytes of RAM.

A “64-bit” machine naturally works with integers that are 64-bits wide, and, once again, uses (up to …) 64-bit memory addresses. This enables the microprocessor to address much more memory.

A 64-bit microprocessor can put itself into a “32-bit mode,” and some even into a “16-bit mode,” to allow for backwards compatibility, doing so on-the-fly to suit the needs of any particular application.

Well to understand the difference, you need to know what a bit is (if you don’t know), and how memory works.

A bit is only a 0 or a 1. And you can count with only that: 0 is 0, 1 is 1, 10 is 2, 11 is 3, 100 is 4, etc.
32/64 bits are the number of bits allowed per number, for 32 the maximum is 11111111111111111111111111111111, or 4 294 967 295.
This represents the highest number you can store in a 32-bit number, if it’s not signed, and is an integer (= has no decimal part).
For 64-bit, the highest is 4 611 686 018 427 387 903.

It means that for example you can store much more information on the Z-depth map than you can with a 32-bit computer.

One of the most known example of the limitations are the Minecraft’s Far Lands back on Beta 1.8 where terrain generation beyond 2 147 483 647 on any coordinates (because the number is signed, you got half on the positive and half on the negative) there were Stack Overflow (that is, the error when you try to input a value higher then the limit described above) errors and terrain was generated weirdly.

Does this mean greater precision for floats?

Yes. limits for floats are the same as for integers, with the only change that they hold decimal information.

Did the representation / precision for floating-point​ change with (Intel) 64-bit? Didn’t think so …

wow, I still dont’ understand it. But thanks.

To keep it simple: The main difference is the amount of RAM the machine and the programs running on it can use.

Windows 32bit e. g. is limited to about 3.25 GB of RAM - even if you have much more RAM physically in your machine, Windows 32bit just can’t use it. And even worse: Any piece of software running on Windows 32bit is limited to 2 GB RAM. So, if we speak Blender: If you’re rendering complex scenes or simulate water or smoke in high resolutions, Blender will crash when hitting that 2 GB memory limit.

On 64bit machines on the other hand there practically is no memory limit for the system or a specific software process (if that software is built for 64bit). The operating system can use all physically available RAM in your system (well, there are exceptions with some versions of Windows, which are arbitrarily limited by Microsoft, but that’s not our topic here). So, yes, a 64bit OS can e. g. render complexer scenes and handle much higher resolutions without crashing.

64 bit means you can use more memory. When your RAM is greater than 4GB, consider installing a 64 bit OS so you can use all that 8GB worth of RAM.

We were looking at this in my computer architecture course last week. The only difference is to do with RAM.
You see, for the CPU to get data from the ram, it sends a memory address along the … address bus. The ram then replies with the data (1 byte, 8bits) along the data bus.
Every point in memory has to have a unique address, right? So when you use a 32 bit integer as the address, you have 2^32 memory bytes you can use, which comes to about 4gb. No matter how hard you try, 32 bits won’t allow you to access more than that.
So you might think that doubling the memory address integer would give you a max of 8gb, but it gives you 2^64 = 1.9x10^19, or according to wikipedia, 16 exbibytes of ram (that’s exbibytes, not exabytes).

It makes no difference to the float precision, speed or anything else. Only the usable ram.