If you walked into the lobby of the Fairmont San Jose hotel this last week, you'd be confronted with this machine and a video of it in use. It was created by Adam and Jamie of Mythbuster's fame for Nvidia, and it illustrates paralellism quite dramatically. It's a mega paintball cannon that paints the Mona Lisa in one single shot. Think ink jet printer on a big scale. As far as geek stuff goes, this is pretty darn cool.
That's probably a good word for the whole conference. As far as tech conferences go, the Nvidia conference was pretty-darn-cool. The key note speaker on the last day was the CTO of Industrial Light and Magic, with follow on detailed talk by one of their key geeks. ILM are the guys who do CG work on practically every film released these days. They gave a big presentation including lots of full screen video, showing exactly how they use parallel GPUs to save thousands of hours of their rendering farms. They demoed pouring absinthe and worked that into the talk in a non-awkward way (I was impressed). They showed us parts of movies that hadn't been released yet. Those parts will appear as blacked out in the youtube video, and the exclusivity was flattering. Finally, they couldn't resist doing some audience participation--making everyone in the audience run a cellular automata program to demonstrate wave and point sources and their parallelism. Talk about taking Life to the next level. Using humans to simulate a cellular automata is something that even the most stuffy people around me seemed to find amusing. If you ever get the chance for an ILM presented key note speech, my advice is--don't you dare miss it.
That's probably a good word for the whole conference. As far as tech conferences go, the Nvidia conference was pretty-darn-cool. The key note speaker on the last day was the CTO of Industrial Light and Magic, with follow on detailed talk by one of their key geeks. ILM are the guys who do CG work on practically every film released these days. They gave a big presentation including lots of full screen video, showing exactly how they use parallel GPUs to save thousands of hours of their rendering farms. They demoed pouring absinthe and worked that into the talk in a non-awkward way (I was impressed). They showed us parts of movies that hadn't been released yet. Those parts will appear as blacked out in the youtube video, and the exclusivity was flattering. Finally, they couldn't resist doing some audience participation--making everyone in the audience run a cellular automata program to demonstrate wave and point sources and their parallelism. Talk about taking Life to the next level. Using humans to simulate a cellular automata is something that even the most stuffy people around me seemed to find amusing. If you ever get the chance for an ILM presented key note speech, my advice is--don't you dare miss it.
All in all, a very interesting and fun conference to go to if you ever get the chance. My presentation had a lot less explosions than your average GPU session since I was talking about automotive graphics. Avoiding explosions is usually pretty fundamental in automotive. Instead, I explained the QNX composition manager. How it uses the OpenKODE API to merge all kinds of graphics output into one seamless display, and how we work with Nvidia to enable those types of applications. Yeah, no explosions. Given that the sessions around me were showing how to create particle simulations to simulate flames or volumetric simulations of fluids, I certainly felt a little jealous at their cool presentations. I'd like to think that the people who attended my session learned about the cutting edge of in-vehicle graphics, even if it wasn't on fire. Or being bombarded by paint balls.
Now that I know the competition, I'll make sure to up my game if I talk again next year.
Great to see QNX and NVidia finally work together. I'm sure many will appreciate the support for NVidia GPUs through QNX.
ReplyDelete