SOFTWARE OPTIMIZATION TECHNIQUES - Software Design for Six-Sigma: A Roadmap for Excellence - page 451

In general, the closer the memory is to the CPU, the more expensive it tends to be.
The main memory holds temporary data and programs for execution by the CPU.
Cache memory is a type of memory designed to provide the most frequently and
recently used instructions and data for the processor, and it can be accessed at
rates many times faster than the main memory can. 25 The processor first looks at
cache memory to find needed data and instructions. There are two levels of cache
memory—internal cache memory and external cache memory. Internal cache memory
is called level 1 and is located inside the CPU chip. Internal cache memory ranges
from 1KB to 32KB. External cache memory is called level 2 and is located on the
system board between the processor and RAM. It is SRAM memory, which can
provide much more speed than main memory.
The registers are for temporary storage for the current instructions, address of the
next instruction, and storage for the intermediate results of execution and are not
a part of main memory. They are under the direction of the control unit to accept,
store, and transfer data and instructions and perform at a very high speed. Earlier
models of computers, such as the Intel 286, had eight general-purpose registers. Some
types of registers have special assignments such as the accumulator register, which
holds the results of execution; the address register, which keeps address of the next
instruction; the storage register, which temporarily keeps instruction from memory
and general-purpose registers, which are used for operations.
The part of the system that manages memory is called the memory manager. Mem-
ory management primarily deals with space multiplexing (Sobh & Tibrewal, 2006).
Spooling enables the transfer of a process while another process is in execution. The
job of the memory manager is to keep track of which parts of memory are in use and
which parts are not, to allocate memory to processes when they need it and to deallo-
cate it when they are done, and to manage swapping between main memory and disc
when the main memory is not big enough to hold all the processes. However, the three
disadvantages related to memory management are synchronization, redundancy, and
fragmentation. Memory fragmentation does not affect memory utilization; however,
it can degrade a system's response, which gives the impression of an overloaded
memory.
Spooling allows the transfer of one or more processes while another process is in
execution. When trying to transfer a very big process, it is possible that the transfer
time exceeds the combined execution time of the processes in the RAM and results
in the CPU being idle, which was the problem for which spooling was invented.
This problem is termed as the synchronization problem. The combined size of all
processes is usually much bigger than the RAM size, and for this reason, processes are
swapped in and out continuously. The issue regarding this is the transfer of the entire
process when only part of the code is executed in a given time slot. This problem
is termed as the redundancy problem. Fragmentation is when free memory space
is broken into pieces as processes are loaded and removed from memory. External
fragmentation exists when enough total memory space exists to satisfy a request, but
it is not continuous.
25 http://www.bsu.edu/classes/nasseh/cs276/module2.html