SYCL Work-Group Mapping and GPU Occupancy Calculation

Help Desk

Slides
ALP Graphic Module 1

Software applications achieve top performance when developed to exploit the characteristics of the hardware on which the code will be run. The code should be structured to produce correct answers but in a way that maximizes the hardware’s ability to execute the code.

This workshop focuses on Intel GPU Optimization techniques. The workshop uses the Intel® Developer Cloud to access Intel® GPUs and SYCL programming to demonstrate optimization techniques. These techniques across different Intel GPU architectures all focus on keeping the compute resources busy on the GPU.

The session covers these topics:

  • Choosing the proper work-group and sub-group size to maximize the occupancy of the GPU resources
  • Using the Intel® GPU Occupancy Calculator to determine the most effective work-group size
  • Understanding the impact of different work-group sizes and local memory sizes on GPU utilization, demonstrated using hands-on coding in SYCL