GATE CSE Subject-Wise Weightage 2009–2026

GATE CSE subject-wise weightage has been fairly stable since 2010. In analyst compilations for 2009–2026, Programming and Data Structures plus Algorithms average about 18 marks a paper, Operating System, COA, Computer Networks and Theory of Computation sit at roughly 8–9 marks each, DBMS is near 7, and Compiler Design and Digital Logic are near 5.

Every subject figure below is approximate. The numbers come from third-party paper analyses, not from IIT Madras or any GATE organizing institute.

Key takeaways

• Only two blocks are fixed by the official pattern: General Aptitude (15 marks) and Engineering Mathematics (about 13 marks). The remaining 72 or so marks are spread across core CS subjects by the paper setters.
• Programming and DS plus Algorithms is the biggest core block, about 18 marks a paper, but the split between the two moves every year.
• COA has been the most reliably high medium-weight subject since 2020. Compiler Design and Discrete Mathematics have also gained weight since 2018.
• Analysts disagree by 2–3 marks in some years because borderline questions get filed under different subjects. Read trends, not single numbers.
• Low weight is not low value. Digital Logic and Compiler Design give some of the best marks per hour of study.

What is officially fixed and what is not

The GATE 2027 CS paper is a 3-hour computer-based test with 65 questions for 100 marks. General Aptitude has 10 questions worth 15 marks, Engineering Mathematics carries about 13 marks, and the core CS subjects carry about 72 marks. That is the only official split, and you should confirm it at gate2027.iitm.ac.in.

How those 72 marks divide between Operating System, DBMS, Networks and the rest is decided paper by paper and never published. So every subject-wise weightage table you see online, including this one, is reconstructed by counting questions in past papers. The GATE CSE exam pattern and marking scheme covers question types and negative marking.

Where the numbers come from

The averages below come from a year-by-year series built from several public analyses: gateexam.info for 2009–2019, GeeksforGeeks for 2018–2023, 2025 and 2026, Careers360 for 2020–2023, GO Classes for 2023–2025 and PW for 2026 shift 2. Where a year had several sets or shifts, the value is the average across them.

Three cautions apply:

• Classification differences. A recurrence relation can be counted under Discrete Mathematics or Algorithms. A C code fragment asking for time complexity can be counted under Programming and DS or Algorithms.
• Merged rows. For 2014 and 2015 the source merges Data Structures and Algorithms, and the split between Discrete Mathematics and the rest of mathematics is not available for every year.
• One-off papers. The 2009 paper had no General Aptitude section, and its very high Operating System (16) and Algorithms (14) figures reflect that one paper, not a trend.

Summary table: average marks per subject

Approximate analyst compilation, marks per 100-mark paper. "Papers with data" counts the years in which the subject could be separated.

This article gives the averages and ranges. The full year-by-year table for every subject from 2009 to 2026, with per-set notes and the source used for each year, is in the GATE CSE 2027 book, which then goes below subject level to a ranked Top 100 concepts.

Five patterns in the weightage data

1. PDS and Algorithms behave as one block

Together they average about 18 marks, the largest core block. The split swings: 2021 gave about 6 marks to PDS and 11 to Algorithms, while 2026 shift 1 gave 9 and 10. Prepare only one of them well and a swing year hurts. Treat them as a single unit.

2. A stable middle tier

Operating System, COA, Computer Networks and Theory of Computation each sit at 8–9 marks on average. COA has been the most dependable recently, with 11, 10, 9, 12, 9, 8 and 9 marks across 2020–2026. DBMS runs at 7–8 marks in the recent period, with 2023 (5 marks) the one weak year.

3. Compiler Design has doubled

Compiler Design moved from 2–4 marks in 2009–2013 to 4–8 marks in recent papers, and data-flow analysis questions have appeared since the 2021 syllabus revision. Digital Logic stays at 3–6 marks with occasional spikes, such as 8 in 2011 and 9 in 2019.

4. Mathematics has settled near 13

Mathematics was overweight in 2014–2017, reaching 30 marks in one 2014 set. Since 2018 it has settled close to the official 13 marks, split roughly 5–10 marks of Discrete Mathematics and 4–8 of Linear Algebra, Calculus and Probability.

5. Single papers are noisy

The 2014 and 2015 sets show the widest variance in the archive: Digital Logic ranged from 2 to 7 marks and COA from 3 to 8 across sets. One paper tells you little. A nine-year average tells you much more.

Weight bands for 2027

Based on these averages, subjects fall into three bands. This is an analysis of past papers, not a forecast of exact marks.

Computer Networks deserves a note. The 2027 syllabus removed many older theory-recall topics, so if the subject keeps its usual share of marks, the retained computation topics such as TCP congestion control and CIDR subnetting are likely to carry more of it.

Low weight is not low priority

A 5-mark subject answered fully moves your score as much as a 10-mark subject answered halfway. Digital Logic and Compiler Design are the most formula-stable subjects in the paper and yield the best marks per hour of preparation. Digital Logic important topics and traps shows what that looks like in practice.

What rose and what fell after 2018

Comparing 2009–2017 averages with 2018–2026 averages:

• Rising: COA (7.7 to 9.2), Compiler Design (3.8 to 5.4) and Discrete Mathematics (6.0 to 7.8).
• Falling: Linear Algebra, Calculus and Probability as a block (13.5 to 6.1, driven by heavy 2016–2017 papers), and slightly, Operating System (9.3 to 8.3) and Algorithms (9.0 to 8.1).
• Gone: old-syllabus areas such as Software Engineering and Web Technologies, absent since 2016.

Inside subjects, recent papers leaned on specific computations: pipelining and cache numericals in COA, multi-level paging with a TLB and inode questions in Operating System, TCP congestion control and subnetting in Networks, and non-standard dynamic programming in Algorithms. The most repeated topics in GATE CSE ranks the concepts behind these marks.

How to use weightage in your plan

1. Allocate time roughly in proportion to marks, then adjust for difficulty. Algorithms and Theory of Computation are the subjects most consistently rated difficult in analyst reviews; General Aptitude and Digital Logic the easiest.
2. Never drop a subject. Even the lowest-weight subjects average close to 5 marks a paper.
3. Prepare borderline topics once. Recurrences and code complexity appear under different subject labels. Learn them once and expect them under either heading.
4. Remember the format mix. In 2021–2026 papers, roughly half the questions were MCQ, a quarter NAT and about a fifth MSQ. NAT and MSQ carry no negative marking, so about half the marks can be attempted without that risk.
5. Go below subject level. Weightage tells you how much time a subject deserves, not which topics inside it matter. Work topic-wise through past papers as explained in how to use GATE CSE previous year papers, then fit the work into a week-by-week GATE CSE 2027 preparation strategy.