The closer you are to the processor, the faster the memories, the faster your pc runs....
L0 cache is bascially the 8 crummy processor intel registers.
For example, PPC processors, like the G5, have 32 processor registers....
Of those 8 registers, only 5 can be used because the others are used for instruction pointing and stack management.
So as you can see, intel processor are actually, really crummy, but intel has market share.....
L1 cache is close to the register file, and is like an extension of it.
L2 may be ondie, or off die.....it is further from the register file
L3 is usually off die, if a processor contains it.....if no L3 on chip is present, then RAM is considered L3, L4 would be harddrive, L5 would removable drive, L6 would be network.....and so on.....
The further you get from the register file, the slower data transfer gets.......
2x L2 cache means that each core has its own L2 cache that it can address.
The processor also has 2x L0 cache at about 32 bytes...only 20 of which that is addressable.....
and a L1 cache........
The more memory you have closer to the processor, the faster the processor will run....thats what it comes down to....
If your really interested in this and not majoring in computer science...then pick up the CSAPP book....but it is still very hard to read....