CSCI 50.01 LT #1

someone kill me pls

Computer Architecture

the attributes of a system that are visible to a programmer (ex. in Discord, users can join servers, send messages, edit their profile or in a computer it is the function to multiply)

Computer Organization

the operational units and their interconnections that realize the architectural specifications (ex. in Discord, programmers can access Database or in a computer it is the hardware)

ISA

Instruction Set Architecture (defines: Instruction formats, opcodes, registers, and memory. Effect of executing instructions on registers and memory. Algorithm for controlling instruction execution)

*This term is interchangeable with Computer Architecture

Structure

The way in which components are interrelated

Function

Operation of each individual component as part of the structure

Four basic functions that a computer can perform

(1) Data Processing; (2) Data Storage; (3) Data Movement; (4) Control

Data Processing

Range of processing requirements is broad. There are few fundamental methods or types.

Data Storage

The method of saving and retrieving information in a computer system, which can include various typs of storage. The computer must temporarily store the pieces of data being worked on; hence need for short-term ______

Data Movement

Computer’s operation environment consists of devices that are either sources or destinations of data, this require some form of _____.

Input Output (I/O)

The process by which data are received from or delivered to a device directly connected to the computer. Moves data between the computer and its external enviornment

Data Communications

The process by which data are moved over longer distances, to or from a remote device

Control

____ unit manages the computer’s resources and directs the performance of its functional parts in response to instructions

Four main components of a computer

(1) Central Processing Unit (CPU); (2) Main Memory; (3) I/O; (4) System Interconnection

Central Processing Unit

Controls the operation of the computer and performs its data processing functions

*Formerly called the processor

Main Memory

Stores short-term data

System Interconnection

Mechanism that provides for communciation among the CPU, main memory, and I/O

System bus

A common example for system interconnection. It consists of a number of conducting wires to which all other components are attached to.

Four main components of a CPU

(1) Control Unit; (2) Arithmetic and Logic Unit; (3) Registers; (4) CPU Interconnection

Control Unit

Directs the operation of the CPU and, by extension, the computer

Arithmetic and Logic Unit

Performs the computer’s data processing functions

Registers

Provides storage internal to the CPU (NOT THE MAIN MEMORY)

CPU Interconnection

Mechanism that provides for communication among the control unit, ALU, and registers

Multicore Computer

When all processing units all reside on a single chip, the term ____ is used.

Core

Each processing unit is called a ____, which is equivalent to a CPU on a single CPU system.

Processor

Computer component that interprets and executes instructions.

Multicore Processor

A processor contains multiple cores

Information

Knowledge that reduces uncertainty

True

Information can be represented by a number

Amount of Information

Smallest maximum length of number needed to represent that information

A = log_10(N/M)

Amount of information in digits

A = log_2(N/M)

Amount of information in bits

CPU Trends

(1) Transistor count still rising; (2) Clock speed flattening sharply; (3) Power reaching its limits; (4) instructions per clock tick

Heat

Limitations in CPU advancements are due to ____. The ability of a microprocessor to remove ____ or dissipate power is ther limiter for CMOS scaling today

Moore’s Law

the number of transistors on a microchip doubles approximately every TWO YEARS, leading to increased computing power and decreased costs.

Propogation Delay (tpd)

The upper bound between new valid inputs and new valid outputs, or how long it takes for the entire system to “settle” after inputs are changed

Contimation Delay (tcd)

The lower bound between invalid inputs and invalid outputs, or how long it takes for one invalid signal to mess up the result

Feature Size

It is the distance between the source and drain on a MOS transistor in this case, means a better transistor

Transistor’s Performance

Tpd is largerly dependent on ____.

True

A device’s tpd CANNOT go below tcd

Highest tpd and lowest tcd

The actual tpd and tcd of a device depends on __________ across all input combinations, over all input-output paths

Critical Path

Route with the highest tpd based on the assumption that delays are additive.

Fan in

It is related to increased number of inputs for a single gate which means higher RC delay values

Fan out

It refers the increased number of connections from an output means higher RC delay values

Delays

____ dictate a processor’s speed and frame rate on a given hardware setup

Modern Processor Capabilities

(1) Pipelining; (2) Branch Prediction; (3) Superscalar Execution; (4) Data Flow Analysis; (5) Speculative Execution

Pipelining

Enables a processor to work simultaneously on multiple instructions by performing a different phase for each instruction at the same time (Like an assembly line). Overlapping execution of multiple instructions, speed up execution of instructions

False

Pipelining helps with latency of a single task.

Instruction Cycle

1. Fetch the instruction from memory

2. Decode the instruction

3. Operand Fetch to get necessary operands

4. Execute the instruction

5. Store the result in the appropriate location

Non-pipelined

Only one unit does both fetching and execution

Pipelined

Higher latency per instruction. Generally higher instruction throughput and lower total latency

Branch Prediction

the processor looks ahead in the instruction code fetched from memory and predicts which groups of instructions are likely to be processed next

Superscalar execution

the ability to issue more than one instruction in every processor clock cycle. Uses multiple parallel pipelines (not necessarily multiple cores)

Data Flow Analysis

the processor checks which instructions are dependent on each other's results or data to create an optimized schedule of instructions

Speculatively Execute

Using branch prediction and data flow analysis, some processors _______ instructions ahead of their actual appearance in the program execution

Cache

Relatively small, fast memory unit (much faster than main memory) designed to hold recently accessed data, speeding up subsequent access to that data

True

Memory speed is much lower than processor speed.

Increasing Processor Speed

(1) Increase hardware speed of processor; (2) Increase the size and speed of caches between the processor and main memory; (3) Make changes to the processor organization and architecture that increase the effective speed of instruction exectuion

Threats to Moore’s Law

(1) Power; (2) RC Delay; (3) Memory Latency and Throughout

Many Integrated Core (MIC)

A series of microarchitectures that integrated many physical cores onto a single integrated circuit.

Multicore

____  processors provide the potential to increase performance without increasing the clock rate

General Purpose Computing on GPUs

GPGPU stands for?

Amdahl's Law

The performance improvement to be gained from using some faster mode of execution is limited by the fraction of the time the faster mode can be used.

Speed Up

\frac{PERFORMANCE_{WITH-ENHANCEMENT}}{PERFORMANCE_{WITHOUT-ENHANCEMENT}}\frac{}{_{}}

Speed Up

=\frac{EXECUTIONTIME_{WITHOUT-ENHANCEMENT}}{EXECUTIONTIME_{WITH-ENHANCEMENT}\left._{}\right.}

Speed Up

Formula to determine how much faster a task will run when using the computer with enhancement/s as opposed to the original, unenhanced computer

Fraction Enhanced <= 1

The fraction of computation time in the original machine that can be converted to take advantage of the enhancement. 1 means the entire time, anything less is that percent of the time

Speed Up Enhanced > 1

The improvement gained by the enhanced execution mode. This particular speed up is ONLY for the time specified by FRACTION_Enhanced

Exec Time Percent

\left(1-FRACTION_{ENHANCED}\right)+\frac{FRACTION_{ENHANCED}}{SpeedUP_{ENHANCED}}

Exec Time New

EXECTIME_{OLD}\cdot EXECTIME_{PERCENT}

Speed Up Overall

\frac{EXECTIME_{OLD}}{EXECTIME_{NEW}}

Speed Up Overall

\frac{EXECTIME_{OLD}}{EXECTIME_{OLD}\cdot EXECTIME_{PERCENT}}

SpeedUp_Overall

\frac{1}{EXECTIME_{PERCENT}}

Clock Speed

Rate of pulses known as _____ or clock rate

Clock Cycle

One increment/pulse is a ____ or clock tick

Clock Period

Time between pulses is the ____ or cycle time

True

Clock Speed is NOT Arbitrary

MIPS Rate

Millions of Instructions per second. A common measure of processor performance.

MFLOPS

Floating-point performance is expressed as millions of floating point operations per second

MFLOPS Rate

Number of executed floating-point operations in a program/Execution time * 10^6.

Little’s Law

L = λW

Throughput

The amount of material or items passing through a system or process.