By Piyush Wairale (M.Tech, IIT Madras) · Former Instructor, IIT Madras BS Degree Programme · Mentor to 20,000+ GATE aspirants · Updated November 2026

GATE CSE 2027 Compiler Design Syllabus (Official, IIT Madras)

Section 7 of the CS paper: Lexical analysis, parsing, syntax-directed translation. Runtime environments. Intermediate code generation. Local optimisation. Data flow analyses: constant propagation, liveness analysis, common subexpression elimination.

Full paper structure: GATE CS 2027 syllabus and exam pattern.

Topic-wise Weightage and Priority

Lexical Analysis

The lexer turns characters into tokens (keywords, identifiers, constants, operators, punctuation, string literals) using regular expressions and finite automata, discarding whitespace and comments. It follows the longest-match rule (so >= is one token, not two). A whole string literal like “x=%d” is a single token. Lexical errors are things like illegal characters; a missing semicolon is a syntax error, not a lexical one.

Parsing: Top-down and Bottom-up

Top-down (LL(1)): needs a grammar with no left recursion and no common prefixes (left factor first). A grammar is LL(1) iff for each A → α | β, FIRST(α) and FIRST(β) are disjoint, and if β ⇒* ε then FIRST(α) and FOLLOW(A) are disjoint. No ambiguous grammar is LL(1) or LR(k).

Bottom-up (shift-reduce): LR(0) ⊂ SLR(1) ⊂ LALR(1) ⊂ CLR(1) in power. LR(0), SLR(1) and LALR(1) parsers have the same number of states; CLR(1) can have more. Merging CLR(1) states into LALR(1) can introduce reduce-reduce conflicts but never new shift-reduce conflicts. Operator precedence parsing handles a restricted class of expression grammars.

Syntax-Directed Translation

Synthesised attributes come from children; inherited attributes come from the parent or left siblings. An S-attributed SDD uses only synthesised attributes and suits bottom-up parsing; an L-attributed SDD allows inherited attributes from the left and suits top-down parsing. Every S-attributed SDD is L-attributed. GATE often gives an SDT with print actions and asks for the output on an input string: build the parse tree, then do a careful post-order (or the specified) traversal.

Runtime Environments

Each procedure call pushes an activation record holding parameters, return address, saved registers, local variables, and control and access links. Static scoping resolves a free variable using the program text (the enclosing block where the function is defined); dynamic scoping uses the most recent active binding on the call stack. Recursion requires stack allocation; static allocation cannot support it.

Intermediate Code, Local Optimisation and Data Flow Analysis

Intermediate forms: three-address code (quadruples, triples), syntax trees and DAGs. A DAG for a basic block shares common subexpressions, so its node count is the number of distinct leaves plus distinct operations. Local optimisations within a basic block include constant folding, copy propagation, dead code elimination and common subexpression elimination. Liveness analysis is a backward analysis: a variable is live at a point if its current value may be used later before being redefined (IN = USE ∪ (OUT − DEF)). Constant propagation is a forward analysis that replaces variables with known constant values.

Five GATE-Style Problems, Solved

Q1 (NAT). How many tokens are in: printf(“x=%d”, x);
Solution: printf, (, “x=%d”, the comma, x, ), ; = 7.

Q2 (MCQ). For E → TE’, E’ → +TE’ | ε, T → id, find FIRST(E’) and FOLLOW(E’).
Solution: FIRST(E’) = {+, ε}, FOLLOW(E’) = FOLLOW(E) = {$}.

Q3 (NAT). How many nodes does the DAG for a + a*(b − c) + (b − c)*d have?
Solution: Leaves a, b, c, d (4). Operations: b−c (shared), a*(b−c), a + that, (b−c)*d, and the final + (5). Total 9.

Q4 (MCQ). In the block a = b + c; d = a + 1; b = d * 2; return b; which variables are live on entry?
Solution: b and c are used before being redefined: {b, c}.

Q5 (NAT). After constant propagation and folding of x = 4; y = x * 2; z = y + x; what value is assigned to z?
Solution: y = 8, so z = 8 + 4 = 12.

Common Mistakes to Avoid

Splitting a string literal into several tokens when counting.

Adding ε to FOLLOW sets: FOLLOW never contains ε.

Assuming LALR has more states than SLR: they have the same number; CLR can have more.

Doing liveness forward instead of backward.

Mixing static and dynamic scoping when tracing a program’s output.

2 to 3 Week Preparation Strategy

Week 1: Lexical analysis, grammars, FIRST/FOLLOW and LL(1). Week 2: LR(0), SLR, CLR and LALR item sets; SDT. Week 3: Runtime environments, intermediate code, DAGs and data flow analysis, then the Compiler Design tests in the GATE CSE 2027 Test Series. Revise Theory of Computation first; the automata and grammar concepts carry over.

FAQs: GATE CSE 2027 Compiler Design

GATE CSE 2027: All Subject-wise Guides

1. Engineering Mathematics · 2. Digital Logic · 3. Computer Organization & Architecture · 4. Programming & Data Structures · 5. Algorithms · 6. Theory of Computation · 7. Compiler Design · 8. Operating System · 9. Databases (DBMS) · 10. Computer Networks

Full paper overview: GATE CS 2027 Syllabus, Exam Pattern & Cutoff. Practise all ten subjects in one place with the GATE CSE 2027 Test Series.