Nvidia is hoping to attract machine learning developers with the Jetson TX1, an ARM-based development board powered by the top-end Tegra X1 SoC. The company claims that in certain deep learning tasks that rely on dynamic input and computations—autonomous drones, facial recognition and behavioural analysis, and computer vision—the Jetson TX1 will beat out an Intel Core i7 6700K Skylake CPU in performance.
The TX1’s machine learning prowess—should Nvidia’s benchmarks hold out under real-world testing—is largely down to the GPU, which excels at the kind of parallel processing required by machine learning and deep neural networks. While the graphics performance of the GPU in the TX1 is just shy of the GPU performance of an i7-6700K, it offers far more performance per watt—up to five times according to Nvidia.
The full TX1 kit comes bundled with Ubuntu 14.04 LTS and Linux 4 Tegra, complete with full OpenGL and OpenGL ES support. While that might sound enticing for knocking up your own custom HTPC, the TX1 doesn’t come cheap. The full kit costs $599 (~£450) at retail, or $299 (~£250) to educational institutions. Pre-orders open November 12 in the US, with units shipping on November 16. An international launch will follow later. The standalone module will be available early next year for $299.
SOURCES – Nvidia, Ars Technica UK