Figure 1 Computer is every where
Types of Computer:
— Desktop
— Server
— Supercomputer
— Embedded system:
Computer is embedded inside in embedded system e.g. cell phones, fridge, cars (anti-lock
breaking system, climate control, fuel injection system, etc.), etc.
NOTE: the embedded processor is used to do one predetermine task i.e. not available for re-
programming
Characteristic of embedded processor:
▪ Low performance requirement
▪ Stringent cost limitation
▪ Stringent power consumption limitation
▪ Low tolerance for failure
Understanding of computer through abstraction
Abstraction Stack/layer (informal speaking):
— By abstraction we convert a concept into levels such that any abstraction level built upon its
below level.
In this way, below level hides complex mechanism or details and with respect to below level
we look viewing (current) level at required perspective.
In this way, abstraction allows us to look and understand the level that is appropriate to
understand the required concept.
Suppose, given level is not enough we take higher level view and again if it not enough to
understand the targeted concept we will go further up level
▪ The point is: we choose the view level to have right trade-off between understanding
and details
Too much detail blurs our understanding.
Not enough details limit our understanding.
DEF (ISA):
It is an abstract interface between hardware and lowest level of software that contains all the
information necessary to write a machine language program that uses register, memory, I/O etc
Note that this also enables various implementation of hardware which has different cost and
performance to run the same code. For example, AMD and INTEL are two different hardware
implementation of x86/x64 arch (of course some different additional features) but have different cost.
However, they can run same software of course with different performance.
Common Arch:
— x86/x64 — SPARC
— PowerPC — ARM
— MIPS — etc.
Remarks:
— Instruction set (ISA) is the platform/stage supported by hardware, on which application and
system program executes
— Without instruction set you really cannot execute program
— Each instruction in the ISA is actually implemented by hardware.
— To run a software, instruction set should be visible to software. On the other hand, to implement
instruction set, it should be visible to hardware.
This means is the interface on which both hardware and software touches and Thus,
▪ is
[Q]: Is ISA is same for AMD and Intel?
— Yes, the instruction set architecture (ISA) of Intel and AMD processors is largely the same.
— Both companies use the x86 and x86-64 (also known as x64 or AMD64) instruction sets for
their CPUs.
Key Points:
— x86 ISA: Originally developed by Intel for their 16-bit 8086 processor, and later extended to
32-bit with the 80386. This ISA has been adopted and implemented by AMD as well.
— x86-64 ISA: AMD developed the 64-bit extension of the x86 architecture, known as AMD64.
Intel later adopted this ISA with minor differences and refers to it as Intel 64. The two are largely
compatible and can run the same software.
Similarities:
— Both Intel and AMD processors can execute the same x86 and x86-64 instructions.
— Most software compiled for one x86/x86-64 processor can run on the other without
modification.
— Both companies' processors support the same fundamental features and capabilities inherent to
the x86/x86-64 architectures.
Differences:
— While the core ISA is the same, Intel and AMD may implement additional proprietary features
and extensions (e.g., Intel's AVX-512, AMD's XFR).
— Each company optimizes their microarchitectures differently, leading to variations in
performance, power consumption, and efficiency for the same tasks.
— There might be slight differences in the implementation of certain instructions or features that
can affect performance or compatibility in edge cases.
Conclusion:
Overall, for most practical purposes, Intel and AMD processors are compatible with each other at the
ISA level, allowing software designed for x86/x86-64 to run on both.
Basic difference between Computer Organization and Architecture
Computer Organization (CO) Computer Arch (CA)
CO deals with Hardware (machine) CA = CO + ISA
implementation of ISA
The CO designer sees CO (below ISA The layer provided to software
level) in terms of function units (FU) programmer (e.g. register, cache,
(e.g. register unit, ALU, shifter, logical memory etc.) and do not change very
unit, etc.) and their interconnection and much. For instance, INTEL provides
they are concerned with: many different chips like atom, i7, i5 all
Performance characteristics of FUs follows x64 arch but their internals
How FUs are interconnected with bus implementation are different due to
and control logic tradeoff between cost, performance and
How information (data & control) energy.
flows between FU components It talks about how fundamentally
How different FUs are coordinated to machine executes a program But do not
realize instruction in ISA talk about How various FU works or
Tradeoff (cost, performance and their implementation issues it comes
required energy) in the under CO.
implementation of ISA like what It talks about instructions offered by the
would be pipeline depth, cache size, processor and what they do
bus width, ALU width etc.
[Q] How hardware and software effects performance?
[1] [Algorithm]: Determines #source level statements require to run a particular task.
[2] [Language/compiler/Arch]: Determines #machine instruction require to run present algorithm
[3] [Processor/Memory]: These specification determines how fast each instruction can execute
[4] [I/O and #cores]: Specifics overall system performance to execute particular task
Basic functional unit of Computer
Computer system (as shown above) can be broadly classified into 5 functional units
— [I/O unit]: Any program receive information from input device like keyboard and produces
output in output-display device
— The received information 1st store in memory for processing
— Required processing or computation via ALU unit
— Data and instruction binary pattern is transferred from src (MEM) to dest (processor) through
interconnection network bus
— Finally, all above mentioned actions are co-ordinated by control unit of the processor.
— Speed of the computer system is defined by the architecture of the processor being used.
Another simplified view of computer
More details diagram on processor (or cpu)
The CPU (or simply processor) is referred as the brain of a computer system. As shown in above
figure, CPU consists of many subsystems (or we can say functional units) like the Control Unit (CU),
the Arithmetic Logic Unit (ALU), and the Registers.
— PC: Program counter a special register that contains next instruction to fetch
— IR: Instruction register hold currently executing instruction. IR directs control unit to control
various components of processor to execute current instruction
— GPR: A register file that contains 32-64 registers to hold operand/data on which operation will
perform.
— ALU: This components is used to perform A & L operations on the operand stored in
GPR/Memory/Immediate (instruction itself).
The ALU contains electronic circuits necessary to perform arithmetic and logical operations.
The arithmetic operations are ADD, SUBSTRACT, MULTIPLY etc.
The logical operations include COMPARE, SHIFT, ROTATE, AND, OR, etc
— The control unit analyses each instruction in the program and sends the relevant signals to all
other units – ALU, Memory, Input unit and Output unit
How computer program executes?
— A computer program consists of both instructions and data. The program is fed into the
computer through the input unit and stored in the memory.
— In order to execute the program, the instructions have to be fetched from memory one by one to
IR (instruction register).
— After an instruction is fetched, the control unit decodes the instruction from the content of IR
(instruction register)..
— According to the instruction, the control unit issues control signals to other units to execute the
instruction.
— After an instruction is executed, the result of the instruction is stored in memory or register.
What is instruction cycle?
— Instruction are fetched and executed by the control unit one by one.
— The sequences involved for the fetch of one instruction and its execution are known as
instruction cycle
— The processor manages the execution of instruction cycle.
Memory management in a computer system
Memory Hierarchy
The memory hierarchy in a computer system refers to the structured arrangement of different types of
memory based on speed, cost, and size. This arrangement exists in all computers. This hierarchy is
necessary because it balances performance and cost — fast memory is expensive and limited in size,
while large memory is slower but more affordable. By organizing memory in levels, the system can
keep frequently accessed data in faster memory and move less-used data to slower storage, improving
processing speed and efficiency.
— Remarks
Different types of memory associated with a given PC
Each has different capacity and different access time
In any system these memory are arranged in a hierarchy. Lower levels are correspondence
with clossness with processor and higher one are far from processor
At the top of the hierarchy are small, fast, and expensive memory units like registers and
cache, followed by main memory (RAM), and then larger, slower, and cheaper storage like
hard drives (HDD/SSD) and external storage.
Component wise diagram of Computer
Above diagram show:
— The fastest bus in above system-diagram is the one which is inside CPU (if exist), then the next
fastest bus is the one that connects .
— Slow peripherals (e.g. USB, serial etc.) are connected via Multi-I/O-controller to system bus
(fast bus but slower then above mentioned busses).
— System chipset is also known as North bridge
