What is actually the difference between a 64 bit cpu and a 32 bit cpu? Is it the instruction set which is 64 bit? or can it add 2 64bit numbers to each other in one cycle or some thing completely differnt?
There are different forms of 64-bit, which provide different things. IA64, which is currently out, is exclusively 64-bit (for now). It runs 32-bit programs very slowly. x86-64, still in development, promises to run 32-bit well, and give good 64-bit performance. However, there are no actual x86-64 processors around yet. It's still an unkown.
BTW, a little known fact is that the assembly language is completely different. A friend of mine (who's a game developer) bitches about IA64, because he doesn't want to spend a year learning a new assembly language.
Sorry I didn't really answer your question, but hopefully I got the ball rolling.
<font color=green>I post so you don't have to!
9/11 - RIP</font color=green>
As I understand it 64 bit computing unlocks some potentially very powerful possibilties. Firstly a 64 bit platfrom can access more than 4 Gb of memory without bank switching, can process much larger numbers (18 x10^18) compared to 4x10^9. Other such advantages include a wider processing pipeline (benchmarks show Itanium to be the mutts nuts at FPU perfomance better than P4, Athlon or UltraSparc.
6 replies and none were relative. Just sad.
The main difference is in the register and alu/fpu bit width. I am no CPU expert but I know enough. The 386 was Intel's first 32 bit x86 cpu. Then as the generations progressed we have seen the 486, pentium, pentium pro/2/3/4. All of which are just extensions to the aging 80x86 core developed around 1979 to extend the 8085 family. The 64 bit cpu's consist of high end expensive families such as the MIPS processors which were used in Intends N64 and Silicon Graphics Systems (SGI). Then there is the well known ALPHA processor which was started by digital and then sold to compaq who sold it to intel to work on the Itanium.
Now the difference works like this if you know a little Hex then the largest number that you can fit into a 32 bit register is FFFFFFFF but a 64 bit register can hold FFFFFFFFFFFFFFFF. Higher numbers means more more work can be done per clock cycle (almost double if you like to think of it that way). Intel's Itanium is a new generation of a powerful 64 bit CPU. It has 128 64 bit bit general purpose registers compared to the pentiums 4 32-bit (I could be wrong). This means that allot of instructions could be carried out with more efficiency because the CPU can hold allot of data before it starts hitting the cache or ram. I hope this explains it in a better light.
Digital invested so much into Alpha that they bought out by compaq. For the sake of interset, the 8086 (I assume that is what you meant by 8085, there was no 8085) was a stop gap to the 8088 which was meant to thrash the Motorla 6800. However the 8088 was proving awkward and the 8086 was pushed instead during Intels Operation Crush. When the 8088 was finally released it was no longer competitive.
What you say is in reference to IA-64 is broadly true.