New Metrics and Models for a Post-ISA Era: Managing Complexity and Scaling Performance in Heterogeneous Parallelism and Internet-of-Things

Pushed by both application and technology trends, today's computer systems employ unprecedented levels of heterogeneity, parallelism, and complexity as they seek to extend performance scaling and support new application domains. From datacenters to Internet-of-Things (IoT) devices, these scaling gains come at the expense of degraded hardware-software abstraction layers, increased complexity at the hardware-software interface, and increased challenges for software reliability, interoperability, and performance portability This talk will explore how new metrics, models, and analysis techniques can be effective in this "Post-ISA" era of shifting abstractions. The talk will cover hardware and software design opportunities, methods for formal verification, and a look into the implications on technologies like IoT.