Welcome and Introduction

• Markus Rampp (MPCDF)

Presentation AMD GPU Roadmap

Overview of MI series programming model

OpenMP Target, lectures with exercises

MPI on AMD GPUs

Wrap-up

HIP and ROCm

HIPifying CUDA code

Optimizing HIP code

Mixing OpenMP and HIP code

OpenMP - Fortran and other tips and tricks

Wrap-up

Debugging with Rocgdb

Timeline profiling with Rocprof and Omnitrace

Kernel profiling with Rocprof and Omniperf

Other resources and wrap-up