Intermediate Code Generation

Intermediate code generation is a phase between the front end (lexical, syntax, semantic analysis) and the back end (code optimization, code generation). The intermediate code is an abstract representation independent of the source or target machine.

Parser → Static Checker → Intermediate Code Generator → Code Generator

Advantages of Intermediate Representation

• Makes the compiler retargetable: one front end, multiple back ends.
• Optimization is easier on IR than on machine code.
• Simplifies the overall compiler design.

Forms of Intermediate Representation

1. Graphical Representation

Two forms: Syntax Trees and DAGs.

• Syntax Tree: depicts the complete hierarchical structure of the source program. Every node represents an operation or operand.
• DAG (Directed Acyclic Graph): compact form of syntax tree. Common subexpressions share nodes — eliminates redundancy.

2. Postfix Notation

An operator follows its operands. No parentheses needed — unambiguous due to fixed operator arity.

Expression:  a = b * -c + b * -c

Postfix:     a  b  c  uminus  *  b  c  uminus  *  +  assign

3. Three-Address Code (TAC)

Each statement has the form:  x := y op z  — at most one operator on the right side. Compiler-generated temporaries are used.