Advanced Topics

[lucaparisi91]

This is meant to be mostly an overview of what can be done using advanced features, more than going into the details.
No base material available.

• Custom mappers ( Create mappers for data structures #1 )
  • mapping of a class containing pointers
• Custom memory allocators ( openmp allocators #4 )
  • Memory allocations allocation per thread ( firstprivate) vs shared variables ( may live in shared memory/global memory/local memory vs being passed as a kernel argument )
  • Pinned memory allocation on CPU
  • shared memory allocations
• Concurrency
  • submit kernels from multiple threads (Streams with openmp #3). Demonstrates using different cudaStreams/hipStreams.
  • use openmp asynchronously ( stretch ) (asynchronous offload #7). Demonstrates overlapping memory transfers and execution or multiple small kernels which might be difficult to merge in one bigger kernel.
• Interoperability ( Calling cuda/rocm libraries from openmp #2 )
  • Dementrate how to use cuda with a variable mapped from openmp and how to use a variable allocated from cuda in openmp.
  • example of using cuFFT ( or any cuda/rocm numerical library ) together with openmp

[lucaparisi91]

Most of the content is there. I need to add the concurrency topics to the slides.