Santa Clara: Designing C++ Hardware

You can run C++ on any computer you want, as long as it pretends it is an ’80s computer. Conveniently most computers pretend to be ’80s computers — with extras, but nothing too radical — because they want to run C++. This contract isn’t written down anywhere, but both sides are absolutely bound by it.

In this talk we’ll walk through the adaptation of the most radical new architecture to run C++ in decades: NVIDIA Volta. The talk's four parts will cover these abstract machine semantics: execution agents, progress guarantees, the object model, and the consistency model.

We will close on new C++ features that make it easier, not harder, for hardware to support C++.

Olivier Giroux has worked on eight GPU and four SM architecture generations released by NVIDIA. Lately, he works to clarify the forms and semantics of valid GPU programs, present and future. He was the programming model lead for the new NVIDIA Volta architecture. He is a member of the ISO C++ committee and is a passionate contributor to C++'s forward progress guarantees and memory model.

This meetup will be hosted at NVIDIA's new global headquarters. You can't miss it - it's the gigantic triangular-shaped building at the intersection of San Tomas and Walsh. There is a parking garage under the building that you can use if you're driving. If you have trouble entering the building, call Bryce Adelstein Lelbach at 510-890-9323.

The talk will begin at 7:00PM, but feel free to come early to socialize and get a tour of the building. We'll be ready for people as early as 6:00PM.