Lectures: TR 3:05-4:25 pm
Location: CCB 17
Instructor: Edmond Chow
E-mail:
Office Hours: Mondays 3-4 pm in KACB 1312
TA: Daniel Henderson
TA E-mail: olen501@gmail.com
TA Office Hours: Fridays 1-3 pm in KACB 1343
Course Description
Design of algorithms and software for high performance computing platforms used in computational science and engineering. Topics include parallelism, locality, machine architectures, and programming.
Students are expected to complete readings prior to each class in order to participate in class discussions and be able to complete short programming assignments in class. Students must be able to bring a laptop computer to class.
If you want to learn how to make programs run really fast, then this course is for you!
Prerequisites
CS 1332, 1372, 2110 or CX 4010. This course requires heavy programming in C/C++. You should have significant programming interest and experience beyond that of the prerequisite courses. Familiarity with using Linux systems is strongly recommended.
Topics
- Aspects of computer architecture and networks
- Parallel algorithms
- Performance modeling
- Parallel scientific applications
- Distributed memory (MPI) programming
- Multithreaded programming (p-threads, OpenMP, etc.)
- Accelerator programming, e.g., Intel Xeon Phi and GPUs
- Other programming models, parallel languages, and tools
Grading
- 80% Mini-Projects (about 6 during the semester)
- 10% Semester-end written assignment (due during exam week)
- 10% Class participation
What You Need to Succeed in this Course
- Desire to learn how to make programs run fast
- Curiosity, for example to investigate performance anomalies
- Engage and participate in class discussions and activities (to do this, you need to keep up with homework and bring a laptop to class)
- Expertise in C or C++ programming
- Familiarity with using the Linux command line
- Not be afraid of matrix operations
Computer Resources
Students in this course will use the jinx cluster, which runs Linux. Accounts will be available at the beginning of the semester. Log into the jinx cluster at jinx-login.cc.gatech.edu using your Georgia Tech user name and password. More information about jinx is available at: https://support.cc.gatech.edu/facilities/instructional-labs/jinx-cluster.
Students in this course will also use Intel Xeon Phi hardware, which is specifically designed for high performance computing.
Textbooks
- Required: Introduction to High Performance Computing for Computational Scientists and Engineers, Georg Hager and Gerhard Wellein, CRC Press, 2010.
- Supplementary: Introduction to Parallel Computing, 2nd edition, A. Grama, A. Gupta, G. Karypis, and V. Kumar, Addison Wesley, 2003. This book is available electronically through the GT library website.
- Supplementary: Parallel Programming and Optimization with Intel Xeon Phi Coprocessors, Colfax International, 2014.
Acknowledgment
We thank the Intel University Program Office for a grant providing hardware used in this course.
