64 bit architecture
 

I am a long ways away from doing any 64-bit programming as I am just learning 32-bit, but I was curious about the new 64-bit processors. Such as what are the registers called and are they broken down into smaller registers like the 32-bit general purpose are ie.(EAX,AX,AH,AL)? Also what size of address space is available? 2^64? If so thats humongous.

Thanks

The 64-bit thangy is the width of the data bus. The wider the bus, the more data you can transfer with a clock edge. That makes for a performance increase. While machines may have an address bus the same width, it ain't necessarily so. The width of the address bus, how many of them are brought outside the device, and how many are wired up will determine the general memory space. Some systems will set aside part of the space available for memory-mapped I/O, also.

I thought terms like 16-bit, 32-bit, etc referred to the native word size of the CPU, not the bus widths (though on many, the latter are at least equal to the former). For example, the 80486 and later x86 chips, with the embedded FPU, could often do atomic writes larger than 32 bits (ie, a wider data bus), though they were 32-bit processors. Likewise, certain chips like the Intel 80386SX or Motorola 68K had a data bus of only 16 bits (and sometimes a more limited address bus as well), though the chips were otherwise 32-bit.

To me and a majority of my peers, it refers to the width of the internal data bus. This may be the same thing as your definition of 'native word size'. This is not the same thing as the number of pins brought out, witness the 32008, etc.

By 'internal data bus', do you mean the number of bits which can be manipulated internally in a single operation, such as copying from one register to another? If so, that's what I meant by 'native word size', though I tend to think of it as the width of the 'general registers' of the CPU in question, particularly in regards to addressing, logical, and integer arithmetic operations, and usually excluding floating-point operations.

I tend to look at it like a schematic of the appropriate devices. Nothing fuzzy about it that way. But that's just me and my background.

>Such as what are the registers called and are they broken down into smaller
>registers like the 32-bit general purpose are ie.(EAX,AX,AH,AL)?
The 64-bit general purpose registers behave exactly the same, they're just extended by another dword and R is used instead of E: RAX, RBX, RCX, RDX, RDI, RSI, RBP, and RSP.

So there are still eight general purpose registers that would use quadword data and a 64-bit data bus. I am assuming to gain much of a performance increase, a program would have to be targeted at a 64-bit OS.

Thanks to all for the info!

there are also 8 new general purpose registers, named r8 - r15

Thanks Tegelane

Here is an interesting article I found on 64 bit Windows programming
64 bit Windows programming