Heterogeneous processors that combine different types of processors such as CPUs, GPUs, and accelerators are emerging in all domains of computing. The state-of-the-art of these heterogeneous processors is advancing faster than ever – they are starting to expose a shared memory interface to enable direct communication across these devices with the promise of energy efficiency and orders of magnitude of performance improvement. Consequently, the interactions between the devices with respective memory models result in a heterogeneous weak memory concurrency model.

In this talk, I will present an overview of my research in this direction and share my learning.