You are on page 1of 17

COMPUTER SYSTEM AND

ARCHITECTURE

(MODULE CODE- AICT001-3-2)

CPU, MEMORY AND ADVANCED


ARCHITECTURE
Introduction
 The report emphasizes on two things-
 General Architecture i.e. CPU and Memory
 Advanced architecture “why it is developed, how it
works, how it differs from the Von Neumann
architecture, which CPU vendor uses it etc.”
Microprocessors
 Microprocessors has input, output and process.
 Input and Output are a series of voltage used to

control external devices.


Different Trends include-
 Increasing number of Cores

 Awareness and improvements in energy efficiency

 Include High Speed Cache and Buses

 Reducing the size of the Processor

 Continuing advances in virtualization


CPU and Memory

 For Better performance, Faster Program Execution,


maximum throughput compatibility is required.
Cache
 Cache Coherence – each CPU is having its own
copy of data in its cache
 Bus Snooping
 Directory Based approach
 Onboard Cache – on the motherboard or off chip
 Write Back Cache- two approaches to write data
back in main memory.
 Implementation – Direct Mapped Cache, Associate
Mapping, Set Associative Mapping
Virtual Memory
 Illusion of Memory
 Implemented using two schemes-
 Demand Paging – valid-invalid bit method, 1 for page
present and o for absent (page fault). The they are
loaded into frames.
 Demand Segmentation- each row of lookup table has a

starting address of the logical memory, size of the


block and starting address for physical memory
However both operates the same.
Demand Paging
Problem of Fragmentation
 Fixed Memory Partitioning
 Variable Memory Partitioning
 Both leads to creation of small holes when the
processes are swapped in.
Memory Hierarchy
Role of Registers
 Registers are defined to perform specific purposes such
as MAR, MDR, Accumulator etc.
 Used as calculator to perform arithmetic calculations
such as Accumulator and are used to hold address of
instructions such as Address Registers.
 Special purpose register are used to store the status of
the program. It consist of program counter, stack
pointer and status registers.
 It has a higher transfer rate and the minimum access
time as it is directly attached to the CPU.
Addressing Modes
 Simplifies the memory references, adds
convenience and flexibility to have modes of
addressing, and it allows a large range of
addressable memory while using a reasonable
number of bits.
 Help writing programs that uses loops and all.
 For Example- Indexed Addressing. Now a day’s
computer allows programmer accessible registers
that manipulate data directly between registers.
Numerical
ADVANCED
ARCHITECTURE

COMPUTER SYSTEM AND


ARCHITECTURE
Multicore
 Doing several operations at a single time increases
efficiency.
 Processor reads and executes fixed and variable
length instructions.
 Reason for Development-
 Increasing CPU performance
 Applications demanded thread level parallelism
 Increased available space.
How it Works
 Concept of multi-threaded programs or software.
Multithreaded software
 Operating systems that have the ability to split
tasks and commands into a set of separate
workloads that could then be processed and run
simultaneously on each of the cores present.
 Vendors –
 Intel and AMD “AMD Phenom II X2, Intel Core Duo,
and AMD Phenom II X 6, Intel Core i7 Extreme
Edition 980 X etc.”
Different from Von Neumann Architecture

 The Von Neumann architecture is concerned with


the sequential processing of instructions whereas in
multicore several instructions can be processed
concurrently.
 Consist of one CPU but multicore consist of more
than one CPU.
Success/Failure
 Success-
 Low Heat Dispatch
 Thread level parallelism
 Failure
 Applications are not capable of splitting tasks

You might also like