Compilers Overview

 

The move to higher-level lanuages

Machine => Assembly => Higher-level languages

=====================================>   more people friendly, easier to develop

 

More human-friendly = Less computer-friendly

 

컴퓨터가 어떻게 high-level 로 작성된 걸 실행시킬 수 있을까?

=> Language translation(additional process) is required

 

 

The major role of language processors (compiler도 이것의 일종)

1. Language translation

• source code => semantically-equivalent target code (assembly / machine lang)

2. Error detection

• translation 과정 중 error를 잡아서 messages를 던짐

 

 

Two repersentative strategies for language processing 

	Compilation	Interpretation
What to translate	An entire source program	One statement of a source program
When to translate	Once before the program runs	Every time when the statement is executed
Translation result	A target program	Target code
Examples	C, C++	Javascript, python

 

 

	Compilation	Interpretation
Runtime Performance	before execution, make target program => no run time overhead => fast	require some run-time overhead => slow
Portability / flexibility	each OS have own language ISA => translation as each in different machine language => portability, flexibility ↓	just install program in any kind of computer => portability, flexibility ↑
Debugging / development	Better when fixing a lot of things (no need to fix it one by one)	Better when fixing a little part of long code  (only check needed part)

 

 

** Variation : Hybrid compilers

• compiler make intermediate program that is easy to traslation into target program
• => reduce run-tiem overhead
• intermediate program can use any type of computer 
• => indepentdent of computer type

 

 

Requirements for designing good compilers

1. Correctness(mandatory)

==> 반드시!! source pgm 을 target pgm으로 semantically equivalent하게 잘 바꿔야함

2. Performance improvement(optional)

3. Reasonable compilation time(optional)

 

 

 

Structure of modern compilers

 

 

Lexical analyzer

• divides the stream of characters into meaningful sequences and produces a set of tokens
• A = B+C  ===> A, =, B, +, C

Syntax analyzer

• creates a tree-like intermediate representation

Semantic analyzer

• checks the source pgm for semantic consistency with the language definition
• ex) type checking/conversion  

Intermediate code generator

• easy to produce and easy to translate into a target machine code
• ex) t1 = inttofloat(C), t2 = t1+B, A = t2

Code optimizer(optional)

• ex)  t1 = inttofloat(C), A = t1+B

Code generator

• maps an intermediate source into the target language