GPU Computing
Get the full GTU study material, English & Gujarati textbook PDFs, syllabus curriculum, and solved previous year question papers.
English Book & Chapters
English Book Coming Soon
We are working hard to add this material. Stay tuned for updates!
Follow on Instagram for UpdatesGTU Previous Year Papers
GPU Computing Syllabus — Unit-wise Content
The official GTU curriculum for GPU Computing (subject code 3720221) as prescribed for CEE · Semester 2 divides the course into 6 units, covering 48 listed topics. The unit-wise breakdown below is read from the GTU syllabus document for this subject code, so you can check what the course actually covers before deciding which papers to attempt.
Download the official syllabus PDF (3720221)Unit 1 Introduction
- GPU Architecture
- Clock speeds
- CPU / GPU comparisons
- Parallel Programming
- CUDA OpenCL / OpenACC
- Kernels Launch parameters
- Thread hierarchy
- Streaming multiprocessors
- 1D/2D/3D thread mapping
- Device properties
- Simple Programs
Unit 2 Memory
- Memory hierarchy
- DRAM / global
- local / shared
- private / local
- Constant Memory
- Parameter Passing
- Arrays and dynamic Memory
- Multi-dimensional Arrays
- Memory Allocation
- Memory copying across devices
- Programs with matrices
- Performance evaluation with different memories
Unit 3 Synchronization
- Memory Consistency
- Barriers (local versus global)
- Memory fence
- Prefix sum, Reduction
- Programs for concurrent Data Structures such as Worklists
- Synchronization across CPU and GPU Functions: Device functions
- Host functions
- Kernels functions
- Using libraries (such as Thrust), and developing libraries
Unit 4 Support
- Debugging GPU Programs
- Profile tools
- Performance aspects Streams: Asynchronous processing
- Overlapped data transfers
- Default Stream
- Synchronization with streams
- Event-based- Synchronization - Overlapping data transfer and kernel execution
- Subject Code: 3720221 w.e.f
Unit 5 Case Studies
- Image Processing
- Graph algorithms
- Deep Learning
Unit 6 Advanced topics
- Dynamic parallelism
- Unified Virtual Memory
- Multi-GPU processing
- Peer access
- Heterogeneous processing
Study on Android App:
