THEBRAIN THATDRIVESIT HOME
NVIDIA's Jetson TX1 to Power New Generation of Autonomous Machines
DECEMBER 2015 NVIDIA's endeavor into the realm of machine learning and the advancement of Deep Neural Network (a.k.a. Deep Learning) has solidified its place in the realm of science and artificial intelligence. That's a huge feat, especially when you consider the company's mainstream reputation as a provider of high-end, gaming-grade graphics cards and workstation-grade graphics solutions.
In addition to that, the company also has a stake in the development of self-learning autonomous vehicle, and has been doing so since the introduction of the Tegra K1 mobile supercomputer back in 2013. Fast forward to today, and the company recently announced the successor to the Tegra K1, the Jetson TX1 module.
The Jetson TX1 is the company's first embedded computer that is designed to process deep neural networks and machine learning with the company's proprietary cuDNN (short of CUDA Deep Neural Network). Specifically, we're talking about a mobile supercomputer module that is able to perform complex tasks like image recognition, conversational speech processing, and real-time navigation via a crowded room.
Specification-wise, the Jetson TX1 is based on NVIDIA's Maxwell GPU architecture, which is made up of 256 cores that has a performance of 1 Teraflops. Supporting is a 64-bit ARM A57 CPU and 4GB of LPDDR4 RAM with speeds of up to 25.6 gigabits per second.
The Jetson TX1 Developer Kit is already available and can be purchased for US$599 (approx. RM2,616), while the module itself will be available to consumers sometime in early 2016 at a price point of US$299 (approx. RM1,306).
Meet the Jetson TX1 module, NVIDIA's new mobile supercomputer.