Here's a bit of a guide on 64 vs. 32:
http://informer.rsbandb.com/article/2009/06/2147/64vs32.htmI'd be willing to wager that more address space won't be one of the features that sells 128 bit. My guess would be that its primary focus would be to enhance processing speed of large integers. The same happened with 64 bit, larger integers were able to be processed faster due to the larger registers.
The upper limit of 32 bits is 2
32 - 1: 4,294,967,295
The upper limit of 64 bits is 2
64 - 1: 18,446,744,073,709,551,616
The upper limit of 128 bits is 2
128 - 1: 3.4028236692093846346337460743177e+38
While no normal human being will ever need the capacity of 128 bits (I hope not at least) lots of intense scientific applications do. Now to the reason why this is important.
In older architectures large numbers like this had to be stored across multiple registers (a register is a hardware component on the CPU that is used for storing data and other things). Basically this means you take a 64 bit integer and chop the bits in half. Put the upper 32 bits in one register and the lower 32 bits in another register. This requires you to stitch them together when you want them to be accessed. The same process can be used for a larger 128 bit number, chopping it into 4 32 bit, bit blocks.
Finally the slowest option for storing large numbers is to either write them to memory in the same fashion or even worse write them to disk.
So this is why high end computational tasks will value 128 bits from a hardare angle. There are also likely to be other security features on the 128 bit CPUs that the OS will implement.
