There are 10 subject sections in the GATE CS syllabus. In a 3-hour paper with 65 questions, only a fraction of those topics actually appear. In The ML Hub's review of recent GATE CS papers (2020–2026), the same topic clusters recur with varying difficulty. This guide maps those clusters for GATE CSE 2027, organised by subject, with the specific sub-topics that carry marks year after year.
The official GATE CS 2027 syllabus PDF from IIT Madras is the authoritative source for scope. Use this guide alongside the GATE CSE subject-wise weightage analysis to prioritise your preparation. Three sections have syllabus changes for 2027: Digital Logic, COA, and Computer Networks — see the GATE CS syllabus page for the full topic list.
Exam Overview: GATE CS 2027, organised by IIT Madras. 100 marks, 65 questions, 3 hours. General Aptitude: 15 marks (fixed). Engineering Mathematics: 13 marks (fixed). Remaining CS sections: 72 marks. Exam dates: 6–7, 13–14, 20–21 February 2027 (source: GATE 2027 Information Brochure, pp.3–4).
How to read the marks below: IIT Madras fixes only General Aptitude (15) and Engineering Mathematics (13); it does not publish marks for individual CS subjects or sub-topics. The subject marks on this page (for example "7.1 avg, range 5–9") are The ML Hub's count of all 23 official GATE CS papers 2014–2026 (1,495 questions), unofficial: the average marks per paper and the lowest–highest marks across the 23 official GATE CS papers, as reproduced on Success Tracker. Each question's marks go to one subject by topic; GA = Q1–10 (Q56–65 in 2017); hashing is counted under Algorithms and disk scheduling under Operating Systems; Discrete Mathematics is counted separately from the rest of Engineering Mathematics; 6 duplicate 2015 questions and 9 old software-engineering questions are excluded. The papers are linked in our GATE CSE previous year question papers guide. Any one paper can differ. The "frequently tested" notes come from our own review of past papers. In short: Programming & Data Structures (10.2 avg) is clearly the largest CS subject, Digital Logic (6.5) and Compiler Design (5.8) the smallest.
Engineering Mathematics: 13 Marks Fixed — Best ROI in the Paper
Engineering Mathematics is the only technical section with a fixed, published mark allocation (13 marks; General Aptitude is also fixed at 15) — source: GATE 2027 question paper pattern. The official pattern allots 13 marks to Engineering Mathematics; classified by topic, maths content averaged 16.4 marks per paper (12–25) in 2014–2026 — exactly 13 in only 3 of 23 papers — because many CS questions are built on discrete maths. In our topic-based count (unofficial), Discrete Mathematics averaged 8.3 marks per paper (3–15) and Linear Algebra, Calculus and Probability together 8.1 (4–11). In our review, the following sub-topics account for most Engineering Mathematics questions:
Discrete Mathematics (8.3 Marks Avg, Range 3–15)
- Graph theory: Connectivity, degree sequences, Euler paths, Hamiltonian paths, graph colouring, bipartite graphs, planar graphs, spanning trees. In our review it is the most frequently tested Engineering Mathematics sub-topic.
- Propositional and first-order logic: Satisfiability, tautology, predicate logic, quantifier scope. Logic questions are almost always 1-mark and reward direct pattern recognition.
- Combinatorics: Counting, pigeonhole principle, recurrence relations (solved by characteristic roots), generating functions. Recurrence relations appear in NAT format — prepare to compute, not just identify.
- Set theory and algebra: Partial orders, equivalence relations, lattices, monoids, groups — present as lighter 1-mark questions.
Linear Algebra, Calculus and Probability together averaged 8.1 marks per paper (range 4–11) in our count; we do not split them further.
Linear Algebra
- Eigenvalues and eigenvectors — NAT computation questions are common.
- System of linear equations: rank, nullity, consistency. LU decomposition.
Probability and Statistics
- Binomial, Poisson, exponential, and normal distributions — mean, variance, and direct application questions.
- Bayes' theorem and conditional probability — frequently tested in recent papers.
- Expected value and variance as NAT questions.
Programming, Data Structures and Algorithms: ~18 Marks (Avg)
This is the largest block of CS marks in our count: Programming & Data Structures alone averages 10.2 marks per paper (range 5–17), the biggest single CS subject, and Algorithms adds 7.9 (range 3–12), for a combined average of about 18.1. Sub-topics below are listed by how often they recur in our review; we do not estimate marks for individual sub-topics.
Graph Algorithms
- BFS and DFS: Traversal order, edge classification, connected components, cycle detection, topological sort. Among the most frequently tested Algorithms topics in our review.
- Shortest paths: Dijkstra's (with heap complexity), Bellman-Ford (negative weights, cycle detection), Floyd-Warshall. Numerical NAT questions are common.
- Minimum spanning trees: Kruskal's and Prim's — given a graph, compute MST weight.
Dynamic Programming
- LCS, LIS, 0/1 Knapsack, Matrix Chain Multiplication, Edit Distance. Write the recurrence, identify time complexity.
Sorting and Complexity
- Merge sort, quick sort (worst-case), heap sort, counting/radix sort. Comparisons made, worst-case scenarios.
- Master theorem for recurrences — mandatory preparation.
Trees and Heaps
- AVL trees (rotation types, height after insertion), B-trees, binary heaps (heapify, build-heap complexity). Classic NAT territory.
Hashing
- Open addressing (linear/quadratic probing, double hashing), chaining, load factor. Probe sequence as NAT questions.
Theory of Computation: 7.1 Marks Avg (3–10)
Regular Languages
- DFA/NFA construction, subset construction, DFA minimisation (table-filling).
- Regular expressions, Kleene's theorem, pumping lemma for regular languages, closure properties.
Context-Free Languages
- CFG design, ambiguous grammars, PDA construction.
- Pumping lemma for CFLs — identifying non-context-free languages — a recurring question type.
Turing Machines and Decidability
- Halting problem, Rice's theorem, recognizable vs decidable languages. Reductions.
- Note: P vs NP and NP-completeness are not listed in the GATE CS 2027 syllabus (TOC ends at "Turing machines and undecidability"), so do not spend time on them.
Operating Systems: 7.6 Marks Avg (5–10) — Numerical Questions Dominate
CPU Scheduling
- FCFS, SJF (preemptive/non-preemptive), Round Robin, Priority Scheduling — compute average waiting time, turnaround time, CPU utilisation. Pure calculation; practice with process tables.
Memory Management
- Page replacement: FIFO (Belady's anomaly), LRU, Optimal — given a reference string and frame count, compute page faults. A recurring NAT question type.
- Effective memory access time with TLB hit ratio. Two-level page table size calculations.
Synchronisation and Deadlock
- Semaphores, producer-consumer, dining philosophers — standard patterns.
- Banker's algorithm — resource matrix, safety algorithm. Classic 2-mark NAT question.
File Systems and I/O
- Inode-based file system: direct/indirect blocks — max file size or disk access count. Disk scheduling (SSTF, SCAN, C-SCAN) — compute head movement.
Computer Networks: 8.2 Marks Avg (6–10) — Revised Syllabus for 2027
Compared with the GATE CS 2026 syllabus, the 2027 syllabus no longer lists OSI/TCP-IP stack naming, framing, Ethernet bridging, flooding and shortest-path as named routing protocols, ARP, DHCP, ICMP, UDP, SMTP, FTP or email. Our 8.2-mark average comes from 2014–2026 papers, so the reduced 2027 syllabus may support somewhat less. Focus only on what remains in the official 2027 syllabus:
- Routing: Distance-vector routing (Bellman-Ford, count-to-infinity, split horizon), link-state routing (Dijkstra applied to routing). Compute routing tables after iterations — NAT question.
- IPv4: CIDR, subnetting, NAT, IP fragmentation (MTU, fragment offset). Numerical subnet calculations are mandatory preparation.
- TCP: Sliding window (Go-Back-N, Selective Repeat — throughput calculations), slow start, congestion avoidance (AIMD), fast retransmit. Trace congestion window evolution — standard 2-mark question.
- Data Link: CRC computation, CSMA/CD, error detection methods.
- Application Layer: DNS (resolution process, records) and HTTP (request-response, persistent vs non-persistent connections) only.
Databases: 7.1 Marks Avg (5–9) — Reliable and Approachable
- SQL: SELECT with JOINs, GROUP BY, HAVING, nested subqueries, aggregation. Given a schema, trace query results. The most approachable DB topic.
- Normalisation: 1NF, 2NF, 3NF, BCNF. Identify functional dependencies, candidate keys, and violations. Normalisation is one of the most consistently tested DBMS topics in our review.
- Relational algebra: σ, π, ⋈, UNION, DIFFERENCE — translate between SQL and relational algebra.
- B+ trees: Insertion, deletion, split/merge — given a tree order, compute structure after operations.
- Transaction concurrency: Serializability (conflict and view), 2PL, MVCC, isolation levels. GATE tests whether a given schedule is serialisable — practice the precedence graph method.
Compiler Design: 5.8 Marks Avg (3–8)
- LL(1) and LR parsing: FIRST and FOLLOW sets, LL(1) parse table construction, SLR/LALR/LR(1) items. Given a grammar, construct the parse table or identify conflicts — appears as 2-mark questions.
- Lexical analysis: Token classification, regex for tokens, DFA for a lexer.
- Syntax-directed translation: Attribute grammars, synthesised and inherited attributes, L-attributed and S-attributed definitions.
- Data flow analysis: Live variable analysis, reaching definitions, available expressions. These appear in the context of code optimisation questions.
Computer Organisation and Architecture: 7.8 Marks Avg (4–13) — Restructured for 2027
The 2027 syllabus adds design of the control unit (hardwired and microprogrammed) and ALU design explicitly, and lists "memory interfacing and hierarchy: performance, cache memory mapping" — the 2026 wording "cache, main memory and secondary storage" is gone (sources: 2027 and 2026 CS syllabi).
- Instruction pipelining: Pipeline stages, throughput, speedup, hazards (data, control, structural), forwarding. Pipeline timing diagrams are standard 2-mark NAT questions.
- Cache memory: Direct-mapped, fully associative, set-associative — cache hit ratio, average memory access time, tag/index/offset field sizes. Cache mapping calculations are a recurring question type.
- Control unit: Hardwired control design, microprogrammed control (micro-operations, microinstruction format) — new explicit addition for 2027.
- ALU and arithmetic: Fixed-point and floating-point representation, addition, multiplication (Booth's algorithm), division — numerical questions.
Digital Logic: 6.5 Marks Avg (3–9) — More Explicit in 2027
Digital Logic is one of the two smallest CS subjects in our count, at 6.5 marks per paper on average (range 3–9), but it is not small enough to skip. The 2027 syllabus now explicitly names algebraic minimisation, K-map minimisation, and tabular (Quine-McCluskey) minimisation as separate learning objectives.
- K-map minimisation: 3-variable and 4-variable K-maps, SOP and POS forms, essential prime implicants. These are consistent 1-mark questions.
- Sequential circuits: Flip-flop types (SR, JK, D, T), state tables, state diagrams, counters, shift registers. Given a circuit diagram, determine the next state — standard question type.
- Combinational circuits: Multiplexers, demultiplexers, decoders, encoders, adders (full adder, ripple-carry, carry-lookahead).
- Number systems: Fixed-point and floating-point arithmetic — signed representation, overflow detection.
General Aptitude: 15 Marks Fixed — Do Not Neglect
General Aptitude is fixed at 15 marks in every GATE paper: 5 one-mark and 5 two-mark questions (source: GATE 2027 Information Brochure, p.24). Many candidates deprioritise it because it feels "non-technical", but every GA mark counts exactly as much as a technical mark: moving from 8 to 12 in GA adds 4 marks to your total with the same technical performance.
- Quantitative aptitude: Data interpretation (graphs, tables, maps), ratios, percentages, powers, exponents and logarithms, permutations and combinations, series, mensuration and geometry, elementary statistics and probability.
- Verbal aptitude: Basic grammar, vocabulary (words, idioms and phrases in context), reading comprehension and narrative sequencing.
- Analytical aptitude: Deduction and induction, analogy, numerical relations and reasoning.
- Spatial aptitude: Transformation of shapes (translation, rotation, scaling, mirroring), assembling and grouping, paper folding, cutting and patterns in 2 and 3 dimensions.
- Topic list source: GATE 2027 General Aptitude syllabus (IIT Madras).
Topic Priority Matrix for GATE CSE 2027
| Subject | Marks per paper: avg (range); ML Hub count unless fixed | Must-Do Topics |
|---|---|---|
| Engineering Mathematics | 13 (official); topic count: Discrete 8.3 (3–15) + LA/Calculus/Probability 8.1 (4–11) | Graph theory, Bayes' theorem, recurrence relations, eigenvalues |
| Programming & DS + Algorithms | 18.1 avg: PDS 10.2 (5–17), Algorithms 7.9 (3–12) | C tracing, BFS/DFS, shortest paths, MST, DP, AVL/B-trees, Master theorem |
| Computer Networks | 8.2 (6–10) | Distance-vector routing, CIDR subnetting, TCP congestion control, CRC |
| COA | 7.8 (4–13) | Pipeline hazards, cache mapping calculations, control unit design |
| Operating Systems | 7.6 (5–10) | CPU scheduling numericals, page replacement, Banker's algorithm, inode |
| Databases | 7.1 (5–9) | SQL with JOINs, BCNF normalisation, B+ tree ops, serializability |
| TOC | 7.1 (3–10) | DFA minimisation, pumping lemma (regular + CFL), Rice's theorem |
| Digital Logic | 6.5 (3–9) | K-map minimisation, flip-flop state tables, combinational circuit design |
| Compiler Design | 5.9 (3–8) | FIRST/FOLLOW sets, LL(1)/LR parse table, data flow analysis |
| General Aptitude | 15 (fixed) | Verbal, quantitative, analytical and spatial aptitude |
Source: fixed marks for Engineering Mathematics and General Aptitude from the GATE 2027 Information Brochure, p.24 (Section 7.3, Table 8). All other marks are The ML Hub's count of all 23 official GATE CS papers 2014–2026 (1,495 questions), unofficial.
Key Takeaways
- Engineering Mathematics (13 marks, officially fixed) and Programming, Data Structures and Algorithms (about 18 marks on average in our count) together account for roughly a third of the paper — anchor your preparation here.
- Beyond that, the CS subjects are fairly even: Computer Networks, COA and OS at about 7.6–8.2 marks, Databases, TOC and Digital Logic at 6.5–7.1, and Compiler Design the smallest at 5.8 — so cover every subject rather than betting on one.
- The Computer Networks syllabus is considerably reduced for 2027. Do not prepare removed topics (OSI model stack naming, ARP, DHCP, UDP, SMTP, FTP, Ethernet bridging, shortest-path/flooding routing).
- COA for 2027 now explicitly includes hardwired and microprogrammed control-unit design — add this to your COA preparation if you prepared for 2026.
- General Aptitude is 15 fixed marks. In our experience, a few weeks of targeted PYQ practice raises GA accuracy substantially; do not treat it as optional.
- NAT questions in OS (scheduling, page replacement, Banker's) and Networks (subnetting, throughput) carry 1 or 2 marks with no negative marking — high priority for score improvement.
Frequently Asked Questions
Which is the most important topic in GATE CSE 2027?
Engineering Mathematics (especially graph theory and probability) is the single most important block because it is fixed at 13 marks in every CS paper. Classified by topic, maths content has averaged even more (16.4 marks per paper in our count), because many CS questions are built on discrete maths. After that, Programming & Data Structures is the largest CS subject in our count (10.2 marks per paper on average), and together with Algorithms (BFS/DFS, shortest paths, DP; 7.9) it forms a block of about 18 marks. The remaining CS subjects average between 5.8 (Compiler Design) and 8.2 (Computer Networks).
What topics were removed from GATE CS Networks syllabus for 2027?
Comparing the 2027 syllabus with the GATE CS 2026 syllabus, the following topics are no longer listed in the GATE CS 2027 Computer Networks syllabus: OSI model stack naming, TCP/IP stack as a named model, framing (Data Link Layer), Ethernet bridging, shortest-path and flooding routing, generic "IP addressing" (only IPv4/CIDR/NAT remains), ARP, DHCP, ICMP, UDP as a separately named protocol, and SMTP/FTP/Email from the Application Layer. DNS and HTTP are the only Application Layer topics retained. Verify against the official 2027 syllabus PDF.
Is Compiler Design important for GATE CSE 2027?
Yes. Compiler Design averages 5.8 marks per paper in our count (range 3–8), the smallest CS subject but still worth securing, and has a well-defined question pattern: FIRST/FOLLOW sets, LL(1)/LR parse table construction, and data flow analysis. These are learnable, repeatable question types, so solving several years of Compiler Design PYQs is high-value practice.
Which topics give the most NAT questions in GATE CSE?
NAT (Numerical Answer Type) questions — which have no negative marking — appear most frequently in: CPU scheduling (waiting time, turnaround time), page replacement (page fault count), cache memory (hit ratio, AMAT), graph algorithms (MST weight, shortest path distance), subnetting (number of hosts, subnet ranges), TCP throughput calculations, and recurrence relation solutions. Practising these as pure computation exercises is essential.
How should I prepare TOC for GATE CSE 2027?
Use Michael Sipser's "Introduction to the Theory of Computation" as the primary reference. Practice DFA minimisation via the table-filling method and pumping lemma proofs for both regular and context-free languages. For decidability, focus on Rice's theorem applications and Turing machine reductions. Solve all previous year GATE TOC questions from 2019 onwards — TOC questions are highly pattern-consistent across years. For structured topic coverage, the ML HUB GATE CS course covers TOC with live classes and all GATE CS PYQs (2014–2026) with solutions on Success Tracker.
What is the most effective way to use this important topics list?
Cross-reference this list against the subject-wise weightage guide to identify where your current preparation gaps carry the highest mark cost. Then study each must-do topic from a standard reference, solve all available GATE PYQs on that specific topic, and take a subject-wise mock test to verify retention. The goal is not breadth — it is depth on the recurring topic clusters listed above.
Next in This Series
Continue with GATE CSE 2027 Study Plan: Month-by-Month Timetable.
Know the Topics — Now Build the Depth
Identifying the right topics is step one. Building the problem-solving depth to score on 2-mark NAT questions under time pressure is the actual work. The ML HUB GATE CS Course covers all 10 GATE CS subjects — taught by Jay Bansal (AIR 2, GATE CS 2019) and Sriniwas Paliwal (AIR 90, GATE CS 2019), both IIT Bombay alumni — with live classes (recorded for replay), topic-wise tests, full-length grand mocks, and all GATE CS PYQs (2014–2026) with solutions on Success Tracker. Explore the course at themlhub.ai/gate-cs-course.
Sources
Official sources checked on 25 October 2026. Anything marked as an estimate or as The ML Hub's analysis is our own and is not an official figure.
- GATE 2027 Information Brochure — IIT Madras (Organising Institute, GATE 2027), Section 7.3, Table 8 (Engineering Mathematics 13 marks in CS): https://gate2027ib.iitm.ac.in/GATE2027-IB.pdf
- GATE 2027 Question Paper Pattern — IIT Madras: https://gate2027.iitm.ac.in/question_paper_pattern
- GATE 2027 CS (Computer Science and Information Technology) syllabus — IIT Madras: https://gate2027.iitm.ac.in/static/doc/GATE2027_Syllabus/CS_GATE2027_Syllabus.pdf
- GATE 2026 CS syllabus — IIT Guwahati (Organising Institute, GATE 2026): https://gate2026.iitg.ac.in/doc/GATE2026_Syllabus/CS_2026_Syllabus.pdf
- GATE 2027 General Aptitude syllabus — IIT Madras: https://gate2027.iitm.ac.in/static/doc/GATE2027_Syllabus/GA_GATE2027_Syllabus.pdf
- Official GATE CS question papers, 2014–2026 — IIT Madras download page (basis of The ML Hub's count of subject marks, via the reproductions on Success Tracker): https://gate2027.iitm.ac.in/download