Advanced Parallel Programming

CISC 879 - Advanced Parallel Programming

Professor	John Cavazos
Class Time	WF, 2:30pm-3:45pm
Room Number	Smith Hall 102A
Office Hours	by appointment
Held in	Smith Hall 412
Course Number	CISC 879

For two decades, software developers enjoyed an increase in application performance due to a doubling of transistors in processers about every two years, in accordance with Moore's Law. Much of this performance gain resulted from increasing clock speeds. However, due to temperature concerns, this is no longer possible. The industry now agree that the future of architecture designs lies in multicores, i.e., processors with simpler cores running at lower frequencies. As a consequence, all computer systems today, from embedded devices to high-end servers, are being built with multicore processors. Thus software developers can no longer sit idly by ("the La-Z Boy Era") and wait for application performance to improve. In fact, application performance is likely to degrade given future generations of multicores with ever more simple cores.

Although researchers in industry and academia are exploring many different multicore hardware design choices, most agree that software for execution on multicore processors is the major unsolved problem. Unlike earlier generations of hardware evolution, this shift will have a major impact on how software is designed and developed. Developers will have to learn how to properly design their applications to utilize multicore parallelism.

The first part of the course will consist of discussions about current multicore architecutures (specifically, GPUs) and parallel programming models (specifically, OpenCL with a lecture or two on CUDA and HMPP). The second part of the course will involve discussing several research papers (chosen by students) pertaining to multicores. Each student will be asked to present one paper to the class. The third and last part of the course will consist of presentations by students of programming projects. Students will have a variety of multicore architectures available for their class projects, including several systems accelerated with GPUs.

Lecture (Tenative Outline)	Slides	Papers / Resources / Notes
2/8 The Multicore Revolution	Slides (PDF)	Read this paper. AND watch the Single-Threaded vs. Multi-Threaded video.
2/10 A General Discussion on Parallelism	Slides (PDF)
2/15 Patterns for Parallel Programming I	Slides (PDF)	Read this paper.
2/17 Patterns for Parallel Programming II	Slides (PDF)
2/22 Introduction to OpenCL (Part I)	Slides (PDF)	Code from lecture here.
2/24 Introduction to OpenCL (Part II)	Slides (PDF)
2/29 Introduction to OpenCL (Part III)	Slides (PDF)
3/2 A Code Merging Optimization Technique for GPU	Slides (PDF)
3/7 Project 1 Overview	Robbie's Slides (PDF) Tristan's Slides (PDF)
3/9 Loop Optimizations (Part I)	Slides (PDF)
3/14 Loop Optimizations (Part II)	Slides (PDF)
3/16 Introduction to HMPP	Scott's Slides (PDF)
3/21 Autotuning HMPP	Slides (PDF)
3/23 OpenACC and How to Read, Write, and Present a Paper	OpenACC (PDF) Presenting a Paper (PDF)
4/4 Student Paper Presentations	Scott Grauer-Gray Angelos	Scott
4/6 Student Paper Presentations	Hristo Asenov Robbie Searles	Hristo Robbie
4/11 Student Paper Presentations	Vinar Amrutia Lifan Xu	Vinar Lifan
4/13 Student Paper Presentations	Shriya Chintada Santosh Nagaral	Shriya Santosh
4/18 Student Paper Presentations	Archana Garikapati Yanan Liu	Archana Yanan
4/20 Student Paper Presentations	Sameer Kulkarni Stephen Sigwart	Sameer Stephen
4/25 Student Paper Presentations	Tristan Vanderbruggen Ben Perry	Tristan Ben
4/27 Student Paper Presentations	Eunjung Park Angelos	EJ Angelos
Class Resources: Class Syllabus Instructions to Use ACAD system Projects: Most of your grade will come from two projects which will total 80% of your grade. Both projects are mandatory. Each student is encouraged to form or find a group. I would like to have groups of 2 developers. Plan on checkpointing your project with the instructor (showing your progress) at various stages during your project. You must checkpoint at least once for the first project and at least twice for the second part of the project. These checkpoints will be part of your grade. You are not required to stay with the same group for both projects. Check the syllabus for more details. Project 1 Handout (Extended Due Date: 3/9) Templates to use for Project Report Project 2 Handout (Due Date: 5/21) Basic Project Plan
Student Presentation Information Your presentation should be on a paper pertaining to the use of GPUs used for general purpose computation (also known as GPGPU), although I will accept presentations on papers related to other cutting parallel architectures and/or programming environments. Note: You should send me an email with the title of the paper you will present at least two weeks before you present so that no two people present the same paper! Tentative Schedule for Talks Template for Talks You can find GPGPU papers at the following locations. GPGPU website GPGPU workshop Another GPGPU website (Look under "Features Stories and Papers") ACM Digital Library (Search for GPGPU) Google Scholar Search for GPGPU