Sycl Academy LogoSycl Academy Logo
listShow Lessons List
  1. What is SYCL
  2. Enqueueing a Kernel
  3. Managing Data
  4. Handling Errors
  5. Device Discovery
  6. Data Parallelism
  7. Introduction to USM
  8. Using USM
  9. Asynchronous Execution
  10. Data and Dependencies
  11. In Order Queue
  12. Advanced Data Flow
  13. Multiple Devices
  14. ND Range Kernels
  15. Image Convolution
  16. Coalesced Global Memory
  17. Vectors
  18. Local Memory Tiling
  19. Further Optimisations
  20. Matrix Transpose
  21. More SYCL Features
  22. Functors
Lesson
Source
Solution
GitHub Iconedit

Functors

In this exercise you will learn how to use functors instead of lambda functions as kernels.

1.) Use functors

Now that global memory access is coalesced and optimized using vec, you can further optimize by using a linear buffer. Given that the blur filter is symmetrical, the same result can be achieved by applying the same linear filter to rows and columns. Instead of using a lambda function for the kernel, a functor, which can be reused, is preferable.

To implement this, you need to create a functor class that accepts an input accessor, an output accessor, an accessor containing the linear filter, and a parameter that specifies the direction in which the filter will be applied.

Keep in mind that in order to apply the linear filter twice, an intermediate buffer of the same size as the input buffer is required to hold the result of the first pass.

2.) Refactor the kernel function

The kernel function must consider both the direction in which the filter needs to be applied and the necessity to store the intermediate image while preserving the halo. It's important to note that the halo is not required for the final image.

Build and execution hints

For DevCloud via JupiterLab follow these instructions.

For DPC++: instructions.

For AdaptiveCpp: instructions.