GATE CS 2014 Set 3 Q17 — Intermediate Code Generation

The correct option is C.

In a compiler, the compilation process is divided into two main parts:

• Front-end (Source-dependent): Lexical analysis, parsing, and semantic analysis.
• Back-end (Machine-dependent): Code generation and register allocation.

An Intermediate Representation (IR) acts as a bridge between the front-end and the back-end. The primary benefits of using an intermediate code include:

1. Portability and Reusability: If we want to build compilers for $M$ source languages and $N$ target machines, using a common intermediate representation requires only $M$ front-ends and $N$ back-ends (total $M + N$ components) instead of $M \times N$ separate compilers.
2. Machine-Independent Optimization: Code optimization can be performed directly on the intermediate code. Since this optimizer is machine-independent, it can be reused across compilers for different target architectures.

Why other options are incorrect:

• Option A: Parsing and semantic analysis occur before the intermediate code is generated, so intermediate code cannot simplify these phases.
• Option B: Error detection and recovery are primarily the responsibility of the front-end phases (lexical, syntactic, and semantic analyzers).
• Option D: Register allocation is a highly machine-dependent task handled during the final code generation phase, which depends on the physical registers of the target architecture.