In recent years, CUDA® and powerful GPUs have given Desktop PCs, Servers and Super Computers new capabilities. Starting with the NVIDIA® Tegra K1 and X1 it is now possible to bring this capability to embedded devices like Drones, Underwater Robotics, Cars, Mobile Medical Devices, Smart Cameras and even Cube Satellites and many more. The complete design of an embedded system – including the NVIDIA SoC, DDR RAM, power supply & BSP adaptations – is quite complex. Although it is not a big hurdle for those who build hundreds-of-thousands of devices with substantial resources & expertise, it can be prohibitive to design such a complex system for a smaller volume product on a limited development budget. A solution is
Computer on Modules (CoM) – also called
System on Modules (SoM). A CoM encapsulates the complexity of the computing system onto a module, significantly lowering the complexity of a specialized design for a real product.
In this talk, we will discuss current SoM solutions like the NVIDIA Jetson TX1 module or the Toradex
Apalis TK1. You will learn about the architecture of such a system. We will demonstrate a typical development process and you will get a feel for the skill set and resources required to transform a maker board prototype into a real product. In the end, we will look at some applications where such systems are deployed, including drones, robots, cinema cameras and even a 24-module TK1 micro cluster.
Takeaway:
Learn how to bring CUDA to low power devices
Experience how to get from a Proof of Concept to Production Device using a system on module
Get an overview/inspiration from embedded devices incorporating CUDA
Speaker Bio:
Daniel Lang has worked 14 years in the embedded industry - the last 10 years at Toradex, a leading provider of embedded computer modules. At Toradex he helped to grow the Swiss startup into a global brand. Originally from Switzerland, he is now located in Seattle WA where he holds the title of CTO of Toradex Inc.
Event Link: https://www.toradex.com/events/bringing-the-power-of-cuda-to-small-devices
Date & Time: Oct 11, 2016 from 19:00 to 20:00
Venue: Mozilla Offices, Suite 500, 366 Adelaide Street West, Toronto, Ontario M5V 1R7, Canada
