Quote:

What if in 10 years we have standards of 256 bits?

It isn't hard to imagine a 256-bit Vector type appearing in that timeframe.

Altivec and SSE2 are 128-bit at the moment. For example, Altivec currently supports:

# 16-way parallelism for 8-bit signed and unsigned integers and characters
# 8-way parallelism for 16-bit signed and unsigned integers
# 4-way parallelism for 32-bit signed and unsigned integers and IEEE floating-point numbers

Which could mean some interesting type names already. A 256-bit variant is not out of the question either in the future, should it become useful. I think the standard is to have:

vector int8 vi;
vector float32 vf;

to describe the types though in this way. Operations are done via vector operations, e.g., vec_madd(multiplier, first, second) to do first*multiplier + second (a very common DSP function) in a vectorised manner.

I don't see general purpose 128-bit integer units being common for a long time though.