Heterogeneous Computing

Class Information

Class room: EECS131
Class time: T6T5T6
Lecturer: Che-Rung Lee
Discussion group iLMS class website

Grading

Class participation: 10%
Four assignments: 40%
1. CUDA Basic: k-nearest neighbor problem. detail
2. Performance Optimization detail
3. Multiple GPU detail
4. OpenCL/OpenACC detail
Final project: 50%

Class Shedule

Week	Topic	Reference
1	Introduction
2	CUDA Basic
3	CUDA Memory
4	CUDA Stream
5	CUDA Architecture
6,7	Performance Optimization	reduce
8	Debugger, Profiler, and Tools
9	Multi-GPU	Multiple GPU
10	Packages and Libraries	OpenACC
11,12	OpenAcc and OpenCL	Matlab GPU
13	Other Accelerators
14	Kepler and CUDA 5	Inside Kepler
15	GPU Precision	GPU Precision
16,17	Project presentations	presentation schedule