watch time: 3 minutes
Hardware’s in a really disruptive period right now, as we: (1) get more — but not faster — transistors (it’s the ‘end’ of Dennard scaling); and (2) move computing into sensors and other small devices in the Internet of Things. It’s all leading to specialization in hardware, from FPGAs to specialized processors and combinations.
So what does this mean for software? Well, we’ll need better system design and programming abstractions. But we’ll also need to embrace making these interfaces much more programmable — and by more people, with a programming model where you don’t have to be a PhD in computer science to use it.
Kathyrn McKinley is a principal researcher at Microsoft (and former professor at the University of Texas at Austin) who focuses on future architectures and on improving the programmability, power, accuracy, and more of their practical applications. This video was recorded as part of the a16z Academic Roundtable 2014.