<aside> ❗ Headings and text with “▶” can be toggled to expand contained info. You can also see the Table of Contents as a collapsible sidebar on the right →

</aside>

<aside> 🚨 Quizzes 6, 7, and 8 are significantly more difficult than the first 5 quizzes. The quizzes make the assumption that you have understood the lectures and thoroughly read the required papers.

</aside>

Table of Contents

Introduction

• This module covers advanced topics in GPU programming, including handling divergent branches, optimizing GPU registers, and understanding shared memory and register files.
• Students will also develop the skills needed to effectively read and analyze GPU architecture research papers.

Objectives

• Describe the performance issues related to divergent branches and the basic mechanisms for handling them.
• Describe the opportunities for GPU register optimization.
• Explain the differences between shared memory and register files.
• Develop the ability to read GPU architecture papers

Readings

Required

Optional

Notes / Things that stand out

Lesson 1 - Handling Divergent Branches

Divergent Branches

• Warp:
  • A group of threads that are executed together.
• So far, we have assumed at all threads in a warp will execute the same program, following SPMD model.
• But, within a warp, not all threads are necessarily executed.
  • Some threads might work on different parts of the program.
  • E.g. If-Else statements.