- 課程代碼
- 11110CS542200
- 課程名稱
- 平行程式Parallel Programming
- 學分
- 3
- 學期
- 1111
- 單位
- 資工系 (CS)
- 班級
- 碩士班
- 修課人數
- 67 人
- 老師
-
- 助教
-
陳煜翔資工系 (CS)
王楷翔資工系 (CS)
歐川銘資工系 (CS)
蔡濶光資工系 (CS)
黃 晨資工系 (CS) - 備註
- 1. The course lectures will be offered online: linkBut the first class on 9/12 will also be in the classroom: delta building 1062. The lecture will be given in English, but you can find the previous recorded lectures in Chinese here: 2021Fall, 2018Fall3. Steps for applying extra enrollment(1) Attend the first class to hear about the course syllabus(3) Receive the confirmation letter by 9/15(4) If you are accepted, follow the school instructions to submit your enrollment request on "Academic Information Systems" (校資系統)
- 課程簡介
- This course will introduce the basic concepts of parallel computing systems, and teach programming languages, including Pthread, OpenMP for multi-core systems, MPI for cluster computing, CUDA for GPU, and MapRduce, Spark for distributed computing. Course students are required to complete 4 program assignments, and a final project.
- 課程大綱
Lectures:
Part I: Introduction
- Introduction to Parallel Computers
- Introduction to Parallel Computing
Part II: Parallel Programming
- Message-Passing Programming (MPI)
- Shared Memory Programming (Pthread and OpenMP)
Part III: Parallel Computing Techniques
- Embarrassingly Parallel Computations
- Partitioning and Divide-and-Conquer Strategies
- Pipelined Computations
- Synchronous Computations
- Load Balancing and Termination Detection
Part IV: GPU Programming
- Heterogeneous computing
- CUDA programming model
- GPU Architecture & Multi-GPU
- Advanced CUDA Programming & Optimization
Part V: Distributed Programming
- Big Data frameworks: MapReduce, Spark
- Distributed Deep Learning
- 教科書
1. Parallel Programming – Techniques and applications Using Networked
Workstations and Parallel Computers, Barry Wilkinson and Michael Allen, Prentice Hall, 1999.
2. Parallel Programming in C with MPI and OpenMP, Michael J. Quinn, McGraw-Hill, 2003.
- 成績說明
Programming assignment: 100%
(1) MPI: Even-old Sort
(2) OpenMP + MPI: Mandelbort
(3) CUDA: All-Pair-Shortest-Path
(4) MPI+Pthread: MapReduce
(5) Final Project
- QR Code
- 放大