IT-803 (D) RGPV B.Tech
Parallel Computing (PC)
Previous year papers and the official syllabus are free to read, no account needed.
Previous Year Papers
1 paper, latest Jun 2025
Syllabus, notes and important questions for IT-803 (D)
Syllabus units
Topics as the official RGPV syllabus lists them.
Unit 1 Introduction 10 topics
- The need for parallelism
- Forms of parallelism (SISD, SIMD, MISD, MIMD)
- Moore's Law and multi‑core processors
- Fundamentals of parallel computers
- Communication architectures
- Message‑passing architecture
- Data‑parallel architecture
- Dataflow architecture
- Systolic architecture
- Performance issues in parallel systems
Unit 2 Large Cache Design 11 topics
- Shared versus private caches
- Centralized versus distributed shared caches
- Snooping‑based cache‑coherence protocols
- Directory‑based cache‑coherence protocols
- Uniform Cache Access (UCA)
- Non‑Uniform Cache Access (NUCA)
- Distributed NUCA (D‑NUCA) and static NUCA (S‑NUCA)
- Cache inclusion and exclusion policies
- Difference between transactional memory and traditional memory models
- Software Transactional Memory (STM)
- Hardware Transactional Memory (HTM)
Unit 3 Graphics Processing Unit 9 topics
- GPUs as parallel computers
- Architecture of modern GPUs
- Evolution of graphics pipelines
- General‑purpose GPUs (GPGPUs)
- Scalable GPU designs
- Architectural characteristics of future GPU systems
- Implications of GPU technology for users
- Vector addition on GPUs
- Application domains of GPUs
Unit 4 Introduction to Parallel Programming 13 topics
- Parallel programming strategies and mechanisms
- Performance theory for parallel programs
- Parallel programming patterns
- Nesting pattern
- Parallel control pattern
- Parallel data‑management pattern
- Map pattern (scaled vector, Mandelbrot set)
- Reduce pattern and fused map‑reduce
- Scan pattern and fused map‑scan
- Data‑recognition patterns (gather, scatter, pack)
- Stencil and recurrence patterns
- Fork‑join pattern
- Pipeline pattern
Unit 5 Parallel Programming Languages 4 topics
- Distributed memory programming with MPI: trapezoidal rule, I/O handling, derived datatypes, collective communication
- Shared memory programming with Pthreads: conditional variables, read‑write locks, cache handling
- Shared memory programming with OpenMP: parallel for directives, loop scheduling, thread safety
- CUDA programming: CUDA C kernels, thread management, constant memory, events, graphics interoperability, atomics, streams