SYCL* Programming for Accelerated Computing

Course Description

This introductory course explores the challenges of heterogeneous computing. Throughout the 11 modules, students learn how to use the SYCL* programming language and how it can target different types of accelerator devices like CPUs, GPUs, and FPGAs.

  • SYCL is an open standard, cross-architecture and cross-vendor programming solution for heterogeneous, high-performance computing (HPC).
  • SYCL breaks proprietary lock-in and gives you the freedom to choose the best accelerated technology for your project.

Included in this course:

  • 11 modules (Estimated time to complete: 18 hours)
Modules
  • Introduction to oneAPI and SYCL
  • SYCL Program Structure
  • SYCL Unified Shared Memory
  • SYCL Sub Groups
  • SYCL Kernel Reductions
  • SYCL Buffers and Accessors in Depth
  • SYCL Task Scheduling and Data Dependency
  • Intel® oneAPI DPC++ Library (oneDPL)
  • Intel® Advisor - Offload Advisor
  • Intel® VTune™ Profiler
  • Intel® Distribution for GDB* on Intel® Developer Cloud
Get the Assignments and Quizzes

11 modules with slide presentations and quizzes are available as a separate download.

Details

Learning Objectives

After completing this course, students will be able to:

  • Understand the challenges of heterogeneous computing.
  • Write a SYCL program that offloads computation to accelerator devices like CPUs, GPUs, or FPGAs.
  • Perform analysis and profile the SYCL code to find performance bottlenecks.
  • Gain familiarity using Intel® oneAPI tools to find performance and optimization opportunities.

Target Audience

Senior undergraduate and graduate students studying:

  • Computer science
  • Engineering
  • Science and mathematics

Prerequisites

  • C++ programming