In this Free OnDemand curriculum you will learn how to migrate GPU‑accelerated applications from CUDA to AMD ROCm™ using HIP, by understanding RDNA™ architecture and SIMT execution, mapping CUDA to HIP, optimizing memory and kernels, and applying practical workflows for porting, debugging, profiling, validation, and performance optimization on AMD iGPUs.
Skill's You'll Learn:
- Understand AMD RDNA™ GPU architecture and the ROCm™ software stack
- Develop and run GPU applications using the HIP programming model
- Migrate CUDA applications to HIP by mapping APIs and execution models
- Debug, profile, and optimize GPU workloads using ROCm tools
