Compiler Lec1

8/13/2019 Compiler Lec1

1/16

CS346: Compilers

Instructor:Instructor: Dr. Arnab Sarkar

TA:TA: Mr. Satish Kumar

10-Jan-13 CSE346: Compilers, IIT Guwahati 1

Book:Book: Compilers: Principles,Tools and Techniques. Aho,

Lam, Sethi, Ullman


2/16

Lecture #1

Introduction

CSE346: Compilers, IIT Guwahati


3/16


4/16

Compilation Principles The compiler must preserve the meaning of the program

being compiled.

Other issues:

Speed


Space Energy Efficiency

Feedback (Information returned back to the user)

Debugging

Compilation time efficiency


5/16

Front-End Back-EndIntermediate

Representation

Source

code

Target

code

Front-end performs the analysis of the source language:

Breaks u the source code into ieces and im oses a

Structure of a Compiler


grammatical structure.

Using this, creates a generic Intermediate Representation(IR) of the source code.

Checks for syntax / semantics and provides informativemessages back to user in case of errors.

Builds a symbol table to collect and store information aboutsource program.phases?


6/16

Front-End Back-EndIntermediate

Representation

Source

code

Target

code

Back-end does the tar et lan ua e s nthesis:



Builds the desired target program from the IR andinformation in the symbol table.

Why do we need to separate the compilation process intofront-end and back-end phases?


7/16


Front-End

Lexical analysis: Scan the source program and identify logical

pieces of the description.

Syntax analysis (Parsing): Identify how those pieces relate toeach other.


Semantic analysis: Identify the meaning of the overall structure.

IR Generation: Design one possible structure.

Back-End

IR Optimization: Simplify the intended structure.

Target Code Generation: Fabricate the structure.

Target Code Optimization: Improve the resulting structure.


8/16

Lexical Analysis

Reads characters in the source program and groups them intomeaningful sequences called lexemes.

Produces as output a token of the form andpasses it to the next phase syntax analysis

Token-class: The symbol for this token to be used during syntax


analysis Attribute: Points to symbol table entry for this token

E.g.: The tokens for:fin = ini + rate * 60 are:

Symbol Table

1 fin 2 ini

3 rate


9/16

Syntax Analysis (Parsing)

Creates a tree-like IR using the tokens produced by thelexical analyser.

This IR depicts the grammatical structure of the token


s ream.

The IR is usually represented as syntax trees

Interior nodes represent operation

Leaves depict the arguments of the operation


10/16

Syntax Analysis (Parsing)

E.g.: Parse tree for:fin = ini + rate * 60

Tokens:

=


+

*

60


11/16

Semantic Analysis

Collects context (semantic) information from syntax tree and symboltable and checks for semantic consistency with language definition.

Annotates nodes of the tree with the results.

Semantic errors:


Type mismatches, incompatible operands, function called withimproper arguments, undeclared variable, etc.

Eg: int ary[10], x; x = ary * 20;

Type checkers in the semantic analyser may also do automatic typeconversions if permitted by the language specification (Coercions).


12/16

Semantic Analysis

=

+

float

float


*

60

inttofloat

float

float

float


13/16

Intermediate Code Generation

Translate language-specific constructs in the AST into more generalconstructs.

A criterion for the level of generality: it should be straightforwardto generate the target code from the intermediate representationchosen.


-

t1 = inttofloat(60)

t2 = id3 * t1

t3 = id2 + t2

id1 = t3

=

+

*

60

inttofloat


14/16

Code Optimization

Generates streamlined code still in intermediate representation.

A range of optimization techniques may be applied. Eg. Removing unused variables

Suppressing generation of unreachable code segments Constant Folding

Common Sub-expression Elimination

Loop optimization (Removing unmodified statements in a loop)... etc.


Our Examp e:t1 = inttofloat(60) t1 = id3 * 60.0t2 = id3 * t1 id1 = id2 + t1

t3 = id2 + t2

id1 = t3


15/16

Target Code Generation

Map 3-address code into assembly code of the target architecture

Instruction selection: A pattern matching problem.

Register allocation: Each value should be in a register when it isused (but there is only a limited number).

Instruction scheduling: take advantage of multiple functional units.

Our Example A possible translation using two registers:


t1 = id3 * 60.0 MOVF id3, R2

id1 = id2 + t1 MULF #60.0, R2

MOVF id2, R1

ADDF R2, R1

MOVF R1, id1


16/16

Next Lecture

Desi n of the Front-End of a Ver Sim le Com iler


Compiler Lec1

Documents

Transcript of Compiler Lec1