Anusha said:Without registers, CPUs won't be there. Not so with cache. Cache it just to speedup the fetching of data and instructions to be processed by the CPU using the locality of reference factor. CPU has no clue as to the existence or the size of the cache.
Anyways...registers are not important in this case, because every CPU has registers that is enough to do what it is supposed to do (except certain RISC architecture CPUs which can create register stacks instead of memory stacks for instruction stacking)
I would like to compare 7-zip performance with a Conroe CPU running at my specs. Mine is a Conroe-2M.
I'm quite mixing up my basics.

I was just trying to make an example for that cache is important argument.
I agree that there will be no CPU without it's set of registers. Different processor architectures have different instruction sets. This might be useful in designing compression algorithms. I've heard that the architecture is closely observed in creating encryption algorithms and for creating encryption crackers (decryption).
I'll very much like to do some compression tests on architectures like Sun SPARC, I also don't have much experience with AMD processors.
The laptop I did the tests was a Core 2 Duo T5600 processor. (1.83GHz 667MHz FSB 2MB L2 cache)
Last edited:
) and 2MB L2
)



