CS 759: High Performance Computing for Applications in Engineering

Fall 2026 · UW–Madison

Cross-listed as ECE/ME/EMA/CS 759. The course explores how to choose and implement parallel computing approaches for engineering workloads, with hands-on programming and a final project.

Topics

• Computer architecture, memory, and performance
• GPU programming with CUDA
• Shared-memory parallelism with OpenMP
• Distributed-memory parallelism with MPI
• Matching applications to parallel hardware and programming models

Notes

• Assignment 1
• Assignment 2: Scan, Convolution, and Matrix Multiplication — due October 2, 2026.
• Assignment 3: OpenMP Matrix Multiplication, Convolution, and Merge Sort — due October 16, 2026, at 23:59.
• Assignment 4: N-body Simulation and OpenMP Scheduling — due October 30, 2026, at 23:59.
• Assignment 5: CUDA Kernels and Vector Scaling — due November 13, 2026, at 23:59.
• Assignments
• Final project — kickoff, proposal, presentation, and suggested directions.
• Lecture notes
• From Code to Machine Instructions
• Instruction-Level and Thread-Level Parallelism
• Time and Memory Hierarchy
• Cache
• Towards Parallel Computing
• C++ Threads
• OpenMP Basics
• OpenMP Loop Scheduling
• OpenMP Tasking and Variable Scoping
• OpenMP Synchronization and Reduction
• N-body Simulation
• OpenMP Performance, False Sharing, and NUMA
• GPU Architecture and CUDA Introduction
• CUDA Kernel Execution Configuration
• CUDA Kernel Scheduling and Thread Divergence
• Bonus: Memory Bandwidth and Virtual Memory
• Syllabus and course details
• Textbooks

Resources

• Assignment code repository — checked out locally in assignments/ as a nested submodule.
• Canvas
• Piazza via Canvas