HSA is a system architecture design for modern computing, integrating CPUs, GPUs, and other accelerators into a single shared memory system. To enable low-latency dispatch between different agents (devices) HSA provides architected queues, which enable low-power hardware implementations along slide software variants. In this chapter HSA a queues are introduced, details of the programmer submission, job scheduling, and integration with HSA's memory consistency model are described.
Gaster, B., Howes, L., & Hower, D. (2015). HSA queuing model. In W. Hwu (Ed.), Heterogeneous System Architecture: Practical Applications for IndustryMorgan Kaufmann