Info

• Lectures : Monday 9-11 am, UG40, Computer Science
• Office Hour: Thursday 3 pm, UG35, Computer Science. Please email for appointments at any other time.
• Lab: Thursday 1-3 pm, Computer Lab UG04 (use the machines marked CUDA)
• Official module page

Resources

This module uses information from courses by Prof. Wen-mei Hwu and others at UIUC and by Prof. Mike Giles at Oxford.

Assessment

Examination 80% + (2 x 10% assignments OR 1 x 20% microproject).

Lectures and Labs

• Revision slides
• slides
• Week 10:The landscape of parallel programming
• slides
• Week 9: N-body problems
• slides
• nbody source, stand-alone tarball
• Assessed exercise 2: modify the example N-body code to simulate flocking behaviour ("boids"). Deadline: Thursday 31 March, in lab.
• Week 8: Parallel sorting
• slides
• Week 7: Case study: memory transpose
• slides
• Week 6: CUDA memory
• Assessed exercise 1
• slides
• Mark Harris's parallel reduce optimisation slides
• Week 5: CUDA threads
• slides
• source code
• lab handout
• Week 4: Basics of CUDA programming
• source code
• lab handout
• Week 3: GPU and FPGA programming for the masses (invited lecture by Prof. Satnam Singh, Microsoft Research)
• paper
• Week 2: The architecture of a GPU
• slides
• lab handout
• Week 1: A critical look at CPU-based architectures
• slides