Scribd Logo
Upload

Welcome to Scribd!

  • Lecture 04 Middle End (CC)

    Uploaded by

    Muhammad Moqeet Azaz
    7 views13 pages
    The document describes a three-pass compiler that uses an intermediate representation (IR) to optimize code in multiple passes before generating machine code. It has three main stages: 1. A front-end that parses source code and generates IR. 2. An optimizer that analyzes and transforms the IR in multiple passes to improve performance. 3. A back-end that takes the optimized IR and generates machine code. The optimizer is the key intermediate stage and aims to reduce running time through transformations like constant propagation, moving computations, specializing based on context, and removing redundant operations. Modern optimizers structure these transformations as a series of optimization passes on the IR.

    Original Description:

    Original Title

    Lecture_04_middle_end(CC)

    Copyright

    © © All Rights Reserved

    Available Formats

    PPT, PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Report this Document
    The document describes a three-pass compiler that uses an intermediate representation (IR) to optimize code in multiple passes before generating machine code. It has three main stages: 1. A front-end that parses source code and generates IR. 2. An optimizer that analyzes and transforms the IR in multiple passes to improve performance. 3. A back-end that takes the optimized IR and generates machine code. The optimizer is the key intermediate stage and aims to reduce running time through transformations like constant propagation, moving computations, specializing based on context, and removing redundant operations. Modern optimizers structure these transformations as a series of optimization passes on the IR.

    Copyright:

    © All Rights Reserved
    Download as PPT, PDF, TXT or read online from Scribd
    Flag for inappropriate content
    7 views13 pages

    Lecture 04 Middle End (CC)

    Uploaded by

    Muhammad Moqeet Azaz
    The document describes a three-pass compiler that uses an intermediate representation (IR) to optimize code in multiple passes before generating machine code. It has three main stages: 1. A front-end that parses source code and generates IR. 2. An optimizer that analyzes and transforms the IR in multiple passes to improve performance. 3. A back-end that takes the optimized IR and generates machine code. The optimizer is the key intermediate stage and aims to reduce running time through transformations like constant propagation, moving computations, specializing based on context, and removing redundant operations. Modern optimizers structure these transformations as a series of optimization passes on the IR.

    Copyright:

    © All Rights Reserved
    Download as PPT, PDF, TXT or read online from Scribd
    Flag for inappropriate content
    of 13

    COMPILER

    CONSTRUCTION
    Lecture 4
    THREE-PASS COMPILER

    Source Front IR Middle IR Back machine


    code End End End code

    errors

    •Intermediate stage for code


    improvement or optimization
    2
    THREE-PASS COMPILER

    Source Front IR Middle IR Back machine


    code End End End code

    errors

    •Analyzes IR and rewrites (or


    transforms) IR
    3
    THREE-PASS COMPILER

    Source Front IR Middle IR Back machine


    code End End End code

    errors

    •Primary goal is to reduce running


    time of the compiled code
    4
    THREE-PASS COMPILER

    Source Front IR Middle IR Back machine


    code End End End code

    errors

    •May also improve space usage,


    power consumption, ...
    5
    THREE-PASS COMPILER
    Source Front IR Middle IR Back machine
    code End End End code

    errors

    • Must preserve “meaning” of the code.


    6
    OPTIMIZER

    IR
    IR Opt IR Opt IR Opt Opt IR
    1 2 3 n

    errors

    •Modern optimizers are structured


    as a series of passes
    7
    OPTIMIZER

    IR
    IR Opt IR Opt IR Opt Opt IR
    1 2 3 n

    errors

    Typical transformations
    •Discover & propagate some
    constant value 8
    OPTIMIZER

    IR
    IR Opt IR Opt IR Opt Opt IR
    1 2 3 n

    errors

    Typical transformations
    •Move a computation to a less
    frequently executed place 9
    OPTIMIZER

    IR
    IR Opt IR Opt IR Opt Opt IR
    1 2 3 n

    errors

    Typical transformations
    •Specialize some computation
    based on context 10
    OPTIMIZER

    IR
    IR Opt IR Opt IR Opt Opt IR
    1 2 3 n

    errors

    Typical transformations
    •Discover a redundant computation
    & remove it 11
    OPTIMIZER

    IR
    IR Opt IR Opt IR Opt Opt IR
    1 2 3 n

    errors
    Typical transformations
    • Remove useless or unreachable code

    12
    HOMEWORK

    What is ad-hoc Lexer?

    13

    You might also like