Figuring Framework Essentials/Patterns Audit of Execution Assessment and ISA Plan.


87 views
Uploaded on:
Category: Animals / Pets
Description
Selection of quantitative ways to deal with PC outline taking into account experimental execution perceptions. ... The term Computer building design is now and then wrongly limited ...
Transcripts
Slide 1

Figuring System Fundamentals/Trends + Review of Performance Evaluation and ISA Design Computing Element Choices: Computing Element Programmability Spatial versus Worldly Computing Main Processor Types/Applications General Purpose Processor Generations The Von Neumann Computer Model CPU Organization (Design) Recent Trends in Computer Design/execution Hierarchy of Computer Architecture Computer Architecture Vs. PC Organization Review of Performance Evaluation Review from 550: The CPU Performance Equation Metrics of Computer Performance MIPS Rating MFLOPS Rating Amdahl\'s Law Instruction Set Architecture (ISA) Review from 550: Definition and reason ISA Types and attributes CISC versus RISC A RISC Instruction Set Example: MIPS64 The Role of Compilers in Performance Optimization (Chapters 1,  2)

Slide 2

Computing Element Choices General Purpose Processors (GPPs): Intended for universally useful registering (desktops, servers, bunches..) Application-Specific Processors (ASPs): Processors with ISAs and design highlights custom-made towards particular application spaces E.g Digital Signal Processors (DSPs), Network Processors (NPs), Media Processors, Graphics Processing Units (GPUs), Vector Processors??? ... Co-Processors: An equipment (hardwired) usage of particular calculations with constrained programming interface (expand GPPs or ASPs) Configurable Hardware: Field Programmable Gate Arrays (FPGAs) Configurable exhibit of basic handling components Application Specific Integrated Circuits (ASICs): A custom VLSI equipment answer for a particular computational errand The decision of one or more relies on upon various elements including: - Type and many-sided quality of computational calculation (broadly useful versus Particular) - Desired level of adaptability/ - Performance necessities programmability - Development cost/time - System cost - Power prerequisites - Real-time obliges The primary objective of this course is to study late compositional configuration methods in superior GPPs

Slide 3

Computing Element Choices The principle objective of this course is the investigation of late engineering plan strategies in elite GPPs General Purpose Processors (GPPs): Flexibility Processor = Programmable registering component that runs programs composed utilizing a pre-characterized set of directions Application-Specific Processors (ASPs) Programmability/Configurable Hardware Selection Factors: Co-Processors - Type and intricacy of computational calculations (broadly useful versus Particular) - Desired level of adaptability - Performance - Development cost - System cost - Power prerequisites - Real-time obliges Application Specific Integrated Circuits (ASICs) Specialization , Development cost/time Performance/Chip Area/Watt (Computational Efficiency) Performance

Slide 4

Computes one capacity (e.g. FP-duplicate, divider, DCT) Function characterized at creation time e.g equipment (ASICs) Parameterizable Hardware: Performs restricted "set" of capacities Computing Element Choices: Computing Element Programmability (Hardware) (Processor) Fixed Function: Programmable: Computes "any" calculable capacity (e.g. Processors) Function characterized after creation Instruction Set (ISA) e.g. Co-Processors Processor = Programmable figuring component that runs programs composed utilizing pre-characterized guidelines

Slide 5

Computing Element Choices: Spatial versus Worldly Computing Spatial Temporal (utilizing programming/program running on a processor) (utilizing equipment) Processor Instructions Processor = Programmable figuring component that runs programs composed utilizing a pre-characterized set of guidelines

Slide 6

Main Processor Types/Applications General Purpose Processors (GPPs) - superior. RISC or CISC: Intel P4, IBM Power4, SPARC, PowerPC, MIPS ... Utilized for broadly useful programming Heavy weight OS - Windows, UNIX Workstations, Desktops (PC\'s), Clusters Embedded processors and processor centers e.g: Intel XScale, ARM, 486SX, Hitachi SH7000, NEC V800... Frequently require Digital sign preparing (DSP) support or other application-particular backing (e.g system, media handling) Single project Lightweight, regularly realtime OS or no OS Examples: Cellular telephones, shopper hardware .. (e.g. Compact disc players) Microcontrollers Extremely cost/power touchy Single program Small word size - 8 bit normal Highest volume processors by a wide margin Examples: Control frameworks, Automobiles, toasters, indoor regulators, ... Expanding Cost/Complexity Increasing volume Examples of Application-Specific Processors

Slide 7

The Processor Design Space Application particular structures for execution Embedded processors Microprocessors GPPs Real-time limitations Specialized applications Low power/cost requirements Performance is everything & Software rules Performance Microcontrollers The principle objective of this course is the investigation of late compositional outline methods in superior GPPs Cost is everything Chip Area, Power multifaceted nature Processor Cost Processor = Programmable figuring component that runs programs composed utilizing a pre-characterized set of guidelines

Slide 8

General Purpose Processor Generations Classified by innovation: The First Generation , 1946-59: Vacuum Tubes, Relays, Mercury Delay Lines: ENIAC (Electronic Numerical Integrator and Computer): First electronic PC, 18000 vacuum tubes, 1500 transfers, 5000 augmentations/sec (1944). Initially put away program PC: EDSAC (Electronic Delay Storage Automatic Calculator), 1949 . The Second Generation , 1959-64: Discrete Transistors. e.g. IBM Main casings The Third Generation , 1964-75: Small and Medium-Scale Integrated (MSI) Circuits. e.g Main edges (IBM 360) , scaled down PCs (DEC PDP-8, PDP-11). The Fourth Generation , 1975-Present: The Microcomputer. VLSI-based Microprocessors. To start with chip: Intel\'s 4-bit 4004 (2300 transistors), 1970. PC (PCs), portable workstations, PDAs, servers, bunches … Reduced Instruction Set Computer (RISC) 1984 (Microprocessor = VLSI-based Single-chip processor) Common variable among all eras: All objective The Von Neumann Computer Model or worldview

Slide 9

- Control Input Memory (guidelines, information) Datapath registers ALU, transports Output CPU Computer System I/O Devices The Von Neumann Computer Model Partitioning of the programmable figuring motor into segments: Central Processing Unit (CPU): Control Unit (direction interpret , sequencing of operations), Datapath (registers, number juggling and rationale unit, transports). Memory: Instruction and operand stockpiling. Information/Output (I/O) sub-framework: I/O transport, interfaces, gadgets. The put away program idea: Instructions from a direction set are gotten from a typical memory and executed each one in turn AKA Program Counter PC-Based Architecture The Program Counter (PC) focuses to next guideline to be prepared Major CPU Performance Limitation: The Von Neumann figuring model suggests consecutive execution one direction at once Another Performance Limitation: Separation of CPU and memory (The Von Neumann memory bottleneck)

Slide 10

Instruction Fetch Instruction Decode Operand Fetch Execute Result Store Next Instruction Generic CPU Machine Instruction Processing Steps (Implied by The Von Neumann Computer Model) Obtain guideline from system stockpiling The Program Counter (PC) focuses to next direction to be handled Determine required activities and guideline size Locate and get operand information Compute result worth or status Deposit results away for later utilize Determine successor or next direction (i.e Update PC) Major CPU Performance Limitation: The Von Neumann processing model infers successive execution one guideline at once

Slide 11

CPU Organization (Design) Components & their associations required by ISA directions Datapath Design: Capabilities & execution attributes of vital Functional Units (FUs): (e.g., Registers, ALU, Shifters, Logic Units, ...) Ways in which these parts are interconnected (transports associations, multiplexors, and so on.). How data streams between parts. Control Unit Design: Logic and means by which such data stream is controlled. Control and coordination of FUs operation to understand the focused on Instruction Set Architecture to be executed (can either be actualized utilizing a limited state machine or a microprogram). Portrayal of equipment operations with an appropriate dialect, conceivably utilizing Register Transfer Notation (RTN). Control/sequencing of operations of datapath parts to acknowledge ISA directions (From 550)

Slide 12

Recent Trends in Computer Design The cost/execution proportion of registering frameworks have seen an enduring decay because of advances in: Integrated circuit innovation: diminishing component size,  Clock rate enhances generally corresponding to change in  Number of transistors enhances relative to    (or speedier). Engineering upgrades in CPU plan. Chip frameworks specifically reflect IC and engineering change as far as a yearly 35 to 55% change in execution. Low level computing construct has been generally killed and supplanted by different options, for example, C or C++ Standard working Systems (UNIX, Windows) brought down the expense of presenting new models. Rise of RISC structures and RISC-center (x86) models. Reception of quantitative ways to deal with PC plan taking into account experimental execution perceptions. Expanded significance of abusing string level parallelism (TLP) in standard registering frameworks. e.g Multiple processor centers on a solitary chip

Slide 13

1000 Supercomputers 100 Mainframes 10 Minicomputers Microprocessors 1 0.1 1965 1970 1975 1980 1985 1990 1995 2000 Year Processor Performance Trend

Recommended
View more...