A Quick Tour of How Cache Works
The p2 has 512kb, that is 512 thousand bytes (well, a kilobyte is really 1024 bytes,
but close enough) of l2 cache. The celeron has 128kb of l2.
What is cache? Processors basically are increasing in speed faster than the memory
is. It is also more expensive to have faster memory, so instead of having alot of really
expensive fast memory, you have a small amount of fast memory, and alot of slower
memory. There are also physical reasons why smaller memories are intrinsically faster
(it is faster, after all, to find something in a small dictionary). Why does this help?
Well, data/instructions tend to be localised, that is when you read information from
memory, the next bit you want to read is most probably near it. If you read one
instruction, the odds are the next one will be near, etc. This means your small bit of
fast memory can hold most of the information you need, most of the time, so your
processor can work out of the faster cache memory most of the time. (this is actually
the same deal as you have with your disk drive and main memory, a small, expensive
and fast memory, vs a larger, cheaper slower memory.)
What is the difference between the p2 and celeron then? Well, the cache on the p2
runs at half the processor speed, which effectively doubles the delay the p2 will see
when it reads data from the l2 cache. The celeron cache is clocked at the same
speed, so can retrieve data faster. The p2 has 4x the cache memory, so there is a
larger chance that the data that the processor wants is in the l2 cache (although it is
not 4x as likely), meaning that the p2 will have the advantage sometimes, and the
celeron will have an advantage at others depending on the kind of software you are
using. It isnt all quite as simple as that, but you get the picture. In most cases, it works
out pretty much the same.