From Graphics To Self Driving Cars
NVIDIA has announced it has developed a new autopilot chip for Tesla’s self driving cars. This autopilot chip is purportedly so advanced it is ahead of the market by 5 years. The advances signal a new era for the graphics hardware company. With the rise of GPU computing, general purpose programming that uses graphics hardware, companies like NVIDIA are shifting their focus towards AI. For Tesla and its aspirations of dominating the automotive market these advances are more than welcome. Recently Tesla decided to put the Drive PX2 supercomputer into every single new car it releases.
Solving The AI Problem
The decision to use the technology of a graphics company to power a self driving car may seem strange, but NVIDIA and its counterparts have spent the last decade and a half working on a completely new approach to solving the sort of issues Tesla is having with it’s self driving car. The hardware needed to create 3D graphics is necessarily very powerful and often more capable than CPUs. Using this hardware for something outside of 3D graphics opens up a world of opportunity. Computer scientists and engineers can use this hardware to create more advanced and faster software. Tesla’s goal of self driving cars is heavily depended on GPU computing.
Ahead Of Its Time
The use of GPU computing in a self driving car is unprecedented. Until recently GPU computing was an experimental approach to solving complex scientific algorithms. Developing GPU computing solutions for commercial issues is so new that GPU computing options weren’t available to the general public until 2008. Nonetheless Elon Musk has invested deeply into NVIDIA’s ideas. Tesla has even cut its relationship with its former autopilot chip manufacturer Mobileye for reasons related to AI development. Tesla’s commitment to NVIDIA’s Drive PX2 computer could give it an advantage over its competition in 2017 and perhaps for many years to come.