MCA-245 Parallel Computing Notes from Unnati Educations cover all 4 units of the GGSIPU MCA Semester 3 Core Elective III paper in 73 pages, with 39 solved numericals and worked examples, 76 short-answer questions with answers and 53 important long questions, mapped to the syllabus for batches admitted from 2025-26.
Written by Prateek Talwar, Founder, Unnati Educations · Reviewed by Ritik Rawat, B.Tech, Senior Full Stack Developer · Checked against the GGSIPU MCA syllabus PDF (syllmcaaff050825900p.pdf, approved 01 Aug 2025) · Updated 14 October 2026 · Editorial policy
Scheme: MCA (Affiliated Institutes), A.S. 2025-26 onwards | Applies to: batch admitted 2025-26 onwards | Verified against: GGSIPU MCA scheme of examination and syllabus PDF (2025-26)
MCA-245 Parallel Computing at a glance: GGSIPU MCA Sem 3 credits, units and exam pattern
MCA-245 is one of eleven Core Elective III options in Semester 3. Values below come from the official GGSIPU MCA scheme and syllabus PDF.
| Code and title | MCA-245 Parallel Computing |
|---|---|
| Course and semester | MCA (2 years), Third Semester |
| Category and credits | Core Elective III (choose any one); 3 credits; L 3, T/P nil |
| Marks | 100: 40 internal + 60 end-term |
| Pass mark | 40 out of 100 (Important Note 9 of the scheme) |
| Applies to | Batch admitted 2025-26 onwards, affiliated institutes |
| End-term pattern | Q1 compulsory, short-answer, whole syllabus, 12 marks; then one of two questions from each unit, 12 marks each (12 + 48 = 60 marks) |
| Units and hours | 4 units: 12 + 10 + 10 + 12 = 44 hours |
| Paired lab | MCA-267 Lab based on Core Elective III (Parallel Computing Lab): 0-2, 1 credit; marks not printed |
Prerequisites printed in the PDF: computer system architecture and an introduction to operating systems. Re-appear rules sit in GGSIPU Ordinance 11. The core papers and every other elective of this term are listed under GGSIPU MCA Semester 3 papers.
MCA-245 Parallel Computing Notes unit-wise: 39 solved numericals and examples across 4 units
These GGSIPU MCA Sem 3 Parallel Computing notes follow Hwang and Grama, so formulas like Amdahl's law, AMAT and wormhole latency always come with a worked number. Unit titles below are the topic headings printed in the official PDF. Notes for every other MCA paper sit under GGSIPU MCA notes.
MCA-245 Unit 1: Theory of Parallelism, Parallel Computer Models, Program and Network Properties
Unit 1 (pages 3 to 21) covers Flynn's and Feng's classifications, UMA, NUMA and COMA, distributed shared memory, Amdahl's and Gustafson's laws, PRAM variants and the VLSI A·T² bound, then Bernstein's conditions, grain packing and static network metrics. It has 8 worked items, 17 short answers, 11 long questions and 8 figures.
| Processors | Speedup 1 / [f + (1 − f)/p] | Efficiency S/p |
|---|---|---|
| 4 | 1 / (0.2 + 0.2) = 2.5 | 62.5% |
| 16 | 1 / (0.2 + 0.05) = 4 | 25% |
| Unlimited | 1/f = 5 (ceiling) | Falls towards 0 |
Quick look: Amdahl's law for a program that is 80% parallel (from page 11)
- Processors
- 4
- Speedup 1 / [f + (1 − f)/p]
- 1 / (0.2 + 0.2) = 2.5
- Efficiency S/p
- 62.5%
- Processors
- 16
- Speedup 1 / [f + (1 − f)/p]
- 1 / (0.2 + 0.05) = 4
- Efficiency S/p
- 25%
- Processors
- Unlimited
- Speedup 1 / [f + (1 − f)/p]
- 1/f = 5 (ceiling)
- Efficiency S/p
- Falls towards 0
MCA-245 Unit 2: Memory Hierarchy Design, Pipelining and ILP
Unit 2 (pages 22 to 40) has two parts. Part A covers effective access time, cache mapping, the 3C miss model, interleaving, paging with a TLB and the Omega network. Part B covers pipeline speedup, hazards, scoreboarding versus Tomasulo, branch prediction, reservation tables with MAL, and superscalar design. It is the most numerical-heavy unit: 12 worked items (10 solved numericals), 19 short answers, 14 long questions and 11 figures.
| Measure | Working | Result |
|---|---|---|
| Pipelined time | (4 + 99) × 10 ns | 1,030 ns |
| Non-pipelined time | 100 × 4 × 10 ns | 4,000 ns |
| Speedup | 4000 / 1030 | 3.88 |
| Efficiency | 100 / 103 | 97.1% |
Quick look: a 4-stage pipeline, 10 ns clock, 100 tasks (from page 31)
- Measure
- Pipelined time
- Working
- (4 + 99) × 10 ns
- Result
- 1,030 ns
- Measure
- Non-pipelined time
- Working
- 100 × 4 × 10 ns
- Result
- 4,000 ns
- Measure
- Speedup
- Working
- 4000 / 1030
- Result
- 3.88
- Measure
- Efficiency
- Working
- 100 / 103
- Result
- 97.1%
MCA-245 Unit 3: Parallel Architectures and Parallel Algorithms
Unit 3 (pages 41 to 56) has two parts. Part A covers wormhole and E-cube routing, vector chaining, the Cray Y-MP, Hwang's R, L, C multithreading model and latency hiding. Part B covers cost-optimality, Brent's theorem, PRAM prefix sums, Cannon's method, shear sort and bitonic sort. It has 9 worked items, 20 short answers, 14 long questions and 5 figures.
| Distance | Store-and-forward | Wormhole |
|---|---|---|
| 4 hops | 16 × 5 = 80 cycles | 16 + 4 = 20 cycles |
| 10 hops | 16 × 11 = 176 cycles | 16 + 10 = 26 cycles |
Quick look: latency for a 512-bit packet, 32-bit flits, 32 bits per cycle (from page 42)
- Distance
- 4 hops
- Store-and-forward
- 16 × 5 = 80 cycles
- Wormhole
- 16 + 4 = 20 cycles
- Distance
- 10 hops
- Store-and-forward
- 16 × 11 = 176 cycles
- Wormhole
- 16 + 10 = 26 cycles
MCA-245 Unit 4: Parallel Programming Platforms, Parallel Languages and Compilers
Unit 4 (pages 58 to 73) has two parts. Part A shows memory latency dragging a 4 GFLOPS processor to 10 MFLOPS, then covers VLIW, SPMD, MSI coherence, the ts + m·tw cost model and XY routing. Part B covers DOALL and DOACROSS, compiler phases, distance vectors and software pipelining. It has 10 worked items, 20 short answers, 14 long questions and 7 figures.
| Write / read | GCD check | Verdict |
|---|---|---|
A[2i] / A[2i + 1] | gcd 2 does not divide 1 | Independent |
A[4i + 2] / A[6i + 4] | gcd 2 divides 2 | Dependence possible (both reach A[10]) |
A[3i] / A[6i + 2] | gcd 3 does not divide 2 | Independent |
Quick look: GCD dependence test on three loops (from page 69)
- Write / read
A[2i]/A[2i + 1]- GCD check
- gcd 2 does not divide 1
- Verdict
- Independent
- Write / read
A[4i + 2]/A[6i + 4]- GCD check
- gcd 2 divides 2
- Verdict
- Dependence possible (both reach A[10])
- Write / read
A[3i]/A[6i + 2]- GCD check
- gcd 3 does not divide 2
- Verdict
- Independent
Static and dynamic dataflow machines in parallel architectures: excerpt from page 49 of the MCA-245 notes
Excerpt from page 49 (Unit 3, Part A, section 8)
8. Dataflow and Hybrid Architectures
In a dataflow machine an instruction runs as soon as all of its operands, called tokens, have arrived. There is no program counter deciding what comes next. The program is drawn as a dataflow graph in which nodes are operators and arcs carry tokens from one operator to the next.
Take z = (a + b) × (c − d). The addition node needs a and b; the subtraction node needs c and d. Neither depends on the other, so both fire in parallel the moment their inputs arrive. The multiplication node waits only for the two results and then produces z. The parallelism is not written by the programmer; it falls out of the graph itself.
(a) Static dataflow (Dennis, MIT)
- At most one token may sit on an arc at any time.
- A node fires only when every input arc holds a token and its output arc is empty, so acknowledge signals have to travel back along the graph.
- The design is simple, but different loop iterations or recursive calls cannot run at the same time, which limits parallelism.
(b) Dynamic dataflow (tagged-token)
- Many tokens may wait on the same arc, and each one carries a tag made of the context, the iteration number and the instruction address.
- A matching unit pairs tokens that share a tag; once a pair is complete, the instruction fires.
- Because tags keep iterations apart, separate loop iterations and procedure calls can run in parallel. Examples are the MIT Tagged-Token Dataflow Architecture and the Manchester Dataflow Machine.
- The weak point is matching: an associative search for equal tags is slow and costly in hardware.
(c) Explicit Token Store (ETS)
ETS removes the associative search. Each activation of a code block is given its own frame in ordinary memory, and the two tokens meant for an instruction meet at a fixed slot inside that frame. Finding the partner becomes a direct memory access instead of a search. The MIT and Motorola Monsoon machine used this idea.
In an exam answer, draw the small graph for z first, then set static and dynamic dataflow side by side on three points: tokens per arc, how matching is done, and how much parallelism each allows. Page 50 of the notes continues with hybrid dataflow and von Neumann designs and a comparison table.
How to write MCA-245 answers: numerical steps, labelled block diagrams and comparison tables
Long questions carry 12 marks each, so structure beats length. The exam tips in the notes come down to four habits:
- Formula, then assumption. Before Amdahl's law, write that it assumes a fixed workload, then show the 1/f ceiling (page 12).
- Tables for compare questions, such as scoreboarding against Tomasulo.
- Draw what is named. Page 35 asks for the 2-bit predictor state diagram, plus the line that it mispredicts a loop branch once per loop, not twice.
- Dry runs for sorting networks. Write every bitonic stage, then finish with the depth and comparator formulas (page 54).
Worked example: branch penalty (from page 35)
Given: 20% branches, 60% of them taken, 3-cycle penalty.
- Without prediction: CPI = 1 + 0.2 × 0.6 × 3 = 1.36, a 36% slowdown.
- With a 90%-accurate predictor: CPI = 1 + 0.2 × 0.1 × 3 = 1.06.
Practise this format on GGSIPU MCA important questions.
MCA-245 exam corner: 76 short-answer and 53 long-answer questions by unit
The notes label short answers 2.5 marks and long answers 12.5 marks (older pattern); the 2025-26 paper sets 12 marks per long question.
| Unit | Short-answer (answered) | Long-answer (listed) | Numericals and examples |
|---|---|---|---|
| Unit 1 | 17 | 11 | 8 |
| Unit 2 | 19 | 14 | 12 |
| Unit 3 | 20 | 14 | 9 |
| Unit 4 | 20 | 14 | 10 |
| Total | 76 | 53 | 39 |
MCA-245 notes: questions and worked items per unit
- Unit
- Unit 1
- Short-answer (answered)
- 17
- Long-answer (listed)
- 11
- Numericals and examples
- 8
- Unit
- Unit 2
- Short-answer (answered)
- 19
- Long-answer (listed)
- 14
- Numericals and examples
- 12
- Unit
- Unit 3
- Short-answer (answered)
- 20
- Long-answer (listed)
- 14
- Numericals and examples
- 9
- Unit
- Unit 4
- Short-answer (answered)
- 20
- Long-answer (listed)
- 14
- Numericals and examples
- 10
- Unit
- Total
- Short-answer (answered)
- 76
- Long-answer (listed)
- 53
- Numericals and examples
- 39
The MCA-245 notes answer 76 short-answer questions and list 53 long-answer questions across 4 units, plus 39 solved numericals and worked examples.
One question per unit, with answer pointers:
- Unit 1: Is the parallel relation transitive? Pointer: no; it is commutative only.
- Unit 2: What is Belady's anomaly? Pointer: under FIFO, more frames can mean more page faults.
- Unit 3: Depth of a bitonic sorter for 16 inputs? Pointer: k = 4, depth k(k + 1)/2 = 10 levels, 80 comparators.
- Unit 4: DOALL versus DOACROSS? Pointer: DOACROSS iterations carry dependences and run synchronised; DOALL ones do not.
More practice sits in the MCA-245 important questions.
MCA-245 and the GGSIPU MCA Sem 3 papers it connects to
Paired lab: MCA-267, Lab based on Core Elective III (Parallel Computing Lab); its entry is on the GGSIPU MCA syllabus page. The PDF gives it four outcomes but no practical list, so these notes are theory-only.
- Units 3 and 4 latency and communication cost → lab CO1.
- Unit 1 SIMD/MIMD and Unit 4 message passing → lab CO3 on paradigms.
- Unit 3 parallel algorithms → lab CO4.
| Group | Papers |
|---|---|
| Core | MCA-201 Design and Analysis of Algorithms; MCA-203 Artificial Intelligence and Machine Learning |
| Core Elective II | MCA-205 Statistics and Data Analytics; 207 Enterprise Computing with JAVA; 209 Natural Language Processing; 211 Computer Graphics; 213 Wireless Sensor Networks; 215 Software Project Management; 217 Advanced Computer Architecture; 219 Distributed Systems; 221 Applied Cryptography; 223 Cloud Computing; 225 e-Business Systems |
| Core Elective III alternatives | MCA-227 Web Intelligence and Big Data; 229 Flutter and Dart; 231 Service Oriented Architecture; 233 Multimedia Technologies; 235 Internet of Things; 237 Soft Computing; 239 Software Quality Management; 241 Digital Image Processing; 243 Compiler Design; 247 Numerical and Scientific Computing |
| Open Elective II | MCA-249 Research Methodology; 251 Operational Research; 253 Cyber Security and Cyber Laws; 255 e-Content Development |
| Labs, project, NUES | MCA-261 DAA Lab; 263 AI and ML Lab; 265 Lab based on Core Elective II; 267 Lab based on Core Elective III; 269 Minor Project III; 271 Entrepreneurship Mindset; 273 Professional Proficiency III |
Other MCA Semester 3 papers (A.S. 2025-26 scheme)
- Group
- Core
- Papers
- MCA-201 Design and Analysis of Algorithms; MCA-203 Artificial Intelligence and Machine Learning
- Group
- Core Elective II
- Papers
- MCA-205 Statistics and Data Analytics; 207 Enterprise Computing with JAVA; 209 Natural Language Processing; 211 Computer Graphics; 213 Wireless Sensor Networks; 215 Software Project Management; 217 Advanced Computer Architecture; 219 Distributed Systems; 221 Applied Cryptography; 223 Cloud Computing; 225 e-Business Systems
- Group
- Core Elective III alternatives
- Papers
- MCA-227 Web Intelligence and Big Data; 229 Flutter and Dart; 231 Service Oriented Architecture; 233 Multimedia Technologies; 235 Internet of Things; 237 Soft Computing; 239 Software Quality Management; 241 Digital Image Processing; 243 Compiler Design; 247 Numerical and Scientific Computing
- Group
- Open Elective II
- Papers
- MCA-249 Research Methodology; 251 Operational Research; 253 Cyber Security and Cyber Laws; 255 e-Content Development
- Group
- Labs, project, NUES
- Papers
- MCA-261 DAA Lab; 263 AI and ML Lab; 265 Lab based on Core Elective II; 267 Lab based on Core Elective III; 269 Minor Project III; 271 Entrepreneurship Mindset; 273 Professional Proficiency III
Who writes and checks the MCA-245 Parallel Computing notes?
- Prateek Talwar, author: Founder, Unnati Educations; Associate Company Secretary (ACS), ICSI, membership no. 58704, since 15 May 2019; B.Com, University of Delhi (School of Open Learning), 2012. Author profile
- Ritik Rawat, subject reviewer: Senior Full Stack Developer, Unnati Educations; B.Tech. He checked the MCA-245 formulas and numericals.
Mapped to the MCA scheme approved on 01 August 2025; 2026-27 edition, 73 pages, 31 figures, updated 14 October 2026. See our editorial policy.
The same team prepares GGSIPU important questions for every course.
Do the MCA-245 notes cover the full GGSIPU syllabus?
Yes. Every topic listed in Units I to IV of the 2025-26 MCA-245 syllabus has its own section in the notes; no gap found.
| Official unit | Official topics | Where in the notes | Status |
|---|---|---|---|
| Unit I | Theory of Parallelism; Parallel computer models | Sections 1-2, pp. 4-14 | Covered |
| Unit I | Program and Network Properties | Section 3, pp. 14-19 | Covered |
| Unit II | Memory Hierarchy Design | Part A, pp. 23-30 | Covered |
| Unit II | Pipelining and ILP | Part B, pp. 31-38 | Covered |
| Unit III | Parallel Architectures (incl. latency hiding, dataflow and hybrid) | Part A, pp. 42-50 | Covered |
| Unit III | Parallel Algorithms (PRAM, mesh, hypercube, combinational circuit) | Part B, pp. 50-54 | Covered |
| Unit IV | Parallel Programming Platforms | Part A, pp. 59-66 | Covered |
| Unit IV | Parallel languages and compilers | Part B, pp. 66-70 | Covered |
MCA-245 official syllabus against the notes
- Official unit
- Unit I
- Official topics
- Theory of Parallelism; Parallel computer models
- Where in the notes
- Sections 1-2, pp. 4-14
- Status
- Covered
- Official unit
- Unit I
- Official topics
- Program and Network Properties
- Where in the notes
- Section 3, pp. 14-19
- Status
- Covered
- Official unit
- Unit II
- Official topics
- Memory Hierarchy Design
- Where in the notes
- Part A, pp. 23-30
- Status
- Covered
- Official unit
- Unit II
- Official topics
- Pipelining and ILP
- Where in the notes
- Part B, pp. 31-38
- Status
- Covered
- Official unit
- Unit III
- Official topics
- Parallel Architectures (incl. latency hiding, dataflow and hybrid)
- Where in the notes
- Part A, pp. 42-50
- Status
- Covered
- Official unit
- Unit III
- Official topics
- Parallel Algorithms (PRAM, mesh, hypercube, combinational circuit)
- Where in the notes
- Part B, pp. 50-54
- Status
- Covered
- Official unit
- Unit IV
- Official topics
- Parallel Programming Platforms
- Where in the notes
- Part A, pp. 59-66
- Status
- Covered
- Official unit
- Unit IV
- Official topics
- Parallel languages and compilers
- Where in the notes
- Part B, pp. 66-70
- Status
- Covered
Verify it yourself in the official GGSIPU MCA syllabus, affiliated institutes, A.S. 2025-26 onwards (MCA-245 is on PDF pages 157 to 159).
Every course in GGSIPU notes is mapped to its official PDF the same way.
MCA-245 Parallel Computing notes FAQs
How do I get the full MCA-245 notes?
Send a WhatsApp message to 9899436384 using the Get Full Notes button on this page. The complete MCA-245 notes come only as a printed book; they are not sold as a PDF. Delivery takes 5 to 7 days at most, depending on your location. The sample PDF above shows six real pages first.
What is the passing mark for MCA-245 Parallel Computing?
MCA-245 is marked out of 100: 40 internal and 60 end-term. The MCA scheme of examination states pass marks as 40%, which is 40 out of 100 for every paper. Rules on end-term minimums, re-appear and evaluation are in GGSIPU Ordinance 11 on ipu.ac.in, which the MCA scheme names as the ordinance governing the programme.
Is MCA-245 compulsory in MCA Semester 3?
No. MCA-245 is one of eleven options in the Core Elective III slot, and you choose one. The scheme says an elective needs at least 20 students for classes; with fewer, it may still be offered, but you complete it through an approved MOOC and still sit the institute's continuous evaluation. Ask your institute which options it runs.
Which MCA-245 unit has the most numericals?
Unit 2 on memory hierarchy and pipelining. It has 10 solved numericals and 2 worked examples, covering effective access time, cache address bits, AMAT, page faults under FIFO, LRU and OPT, pipeline speedup, branch CPI, collision vectors with MAL, and superscalar timing. Unit 4 comes next with 6 solved numericals, mainly on memory performance and message cost.
Do I need MCA-217 Advanced Computer Architecture before MCA-245?
No. The MCA-245 prerequisites are computer system architecture and an introduction to operating systems. MCA-217 sits in the separate Core Elective II slot, so you can take both in the same term. They share the Hwang and Jotwani text, so Flynn's taxonomy, interconnection networks, memory interleaving and pipelines overlap, and revision for one helps the other.
Is there a lab with MCA-245 Parallel Computing?
Yes. MCA-267, Lab based on Core Elective III, runs as the Parallel Computing Lab for students who take MCA-245, with 2 practical hours a week and 1 credit. The official PDF gives its learning objectives and four course outcomes but no list of practicals, so these notes cover only the MCA-245 theory paper and its units.
Get the full MCA-245 Parallel Computing Notes on WhatsApp
The full MCA-245 Parallel Computing IPU notes come as a printed book, not a PDF. These are Parallel Computing notes in English with block diagrams, ordered on WhatsApp.
Pages and worked items per unit in the printed MCA-245 book
- Part
- Cover and index
- Pages
- 1-2 (2)
- Numericals and examples
- –
- Part
- Unit 1
- Pages
- 3-21 (19)
- Numericals and examples
- 8
- Part
- Unit 2
- Pages
- 22-40 (19)
- Numericals and examples
- 12
- Part
- Unit 3
- Pages
- 41-56 (16)
- Numericals and examples
- 9
- Part
- Unit 4
- Pages
- 58-73 (16)
- Numericals and examples
- 10
- Part
- Total
- Pages
- 73
- Numericals and examples
- 39
Delivery: within 5 to 7 days at most; the exact time may vary with your location.
Get the full MCA-245 notes on WhatsApp (4 units, 73-page printed book)
Unnati Educations is an independent education provider based in Vikaspuri, Delhi. We are not affiliated with, authorised by, or endorsed by Guru Gobind Singh Indraprastha University (GGSIPU). 'GGSIPU' and 'IPU' are used only to describe the syllabus our notes follow. For official information, visit ipu.ac.in.
Ready for the full notes?
Discuss coverage and delivery with our team.