Introduction
Memory Cache is additional working memory, which the CPU
can access easier (and faster) than main memory. Having a memory cache can speed
up processing in some cases.
Today’s Pentium computers have so much memory; the
memory addresses are by necessity fairly long. Since the processor must keep
track of every memory location, quite a bit of memory and processing time is
used up just keeping track of what data is where.
To improve performance, computer designers began
providing the CPU with supplementary “scratch pad” memory. This supplemental
memory allows CPU to store immediate results and other data in a special
“cache” memory chip, one that has short addresses.
Today’s Pentium chips come
with some cache built into the chip itself. This is known as Level 1 (or L1)
cache. But if you will be using complex programs, it might help to have
additional cache memory. Most today’s motherboards allow for an additional
cache memory, known as Level 2 (or L2) cache.
Having 256 kilobytes of Level 2
cache can significantly enhance the speed with which the Pentium computer
executes complex programs. Having 512 kilobytes might offer even faster
processing. But beyond 512k, the memory addresses begin to get long again,
defeating the purpose of having cache memory in the first place!
Most of today’s software can
be run adequately without Level 2 cache, but software is becoming so advanced
that having 256k or 512k of L2 cache will provide noticeable improvement in
performance.
Cache memory is generally very
cheap, at least up to about 256k. Often, computer systems will come with L2
cache at little or no extra charge.
Usage of Memory Cache (or Cache Store)
Since one instruction may have
several accesses to memory the speed of memory access is an important
determinant of the speed of the processor. The cache store is small store with
very fast access time situated between the memory and the processor. It works on
the assumption that for most of the time processing requires data and
instructions, which are close to each other in main memory. Areas of main memory
surrounding the current machine instruction are loaded into the cache store,
thus data and instructions required by the processor are then accessed by the
cache store.
Reference :
