Android on TX2

A long time ago, Android L was released for the TK1 boards. Are their any plans to bring a modern version of Android (7.1 Nougat for example) to the TX2 module? Being able to run a modern version of Android on the Jetson boards provides great value to many developers.

At this time, there are no plans to release nor support Android on Jetson. Typically dealing with device drivers and software customization is more challenging on Android than on Ubuntu Linux.

What specific use case does Linux4Tegra not serve for you?

We can use any software: multimedia, office, navigation and etc with Android system. It’s best choise for multimedia car systems. Creating navigation software for ubuntu is more dificultly as for android.

What Linaro says, plus creating a GUI that works with touch-input is way easier on Android.

It would be nice to leverage Google’s Project Tango on the TX2.

The Jetson is ideally suited for 3D SLAM.

I’ll add my vote for Android support on TX2. There are Android SDK’s that I’d like to use.

As someone who is trying to incorporate deep learning, and GPU computing into shipping products I have to say it’s been pretty frustrating.

I’m in a weird spot by being absolutely in awe of what it can do in terms of performance/watt, but at the same time there are all these road blocks.

It doesn’t support Android despite the fact that the NVidia shield uses Android (if I’m not mistaken). The lack of Android support doesn’t really even impact anything I do. I actually prefer straight up Linux, but it’s extremely hard to sell the higher ups on it when the software developers roll their eyes at the lack of support for Windows IOT, or Android.

It doesn’t support booting fully (uboot, kernel, filesystem, etc) from the SD card. So I can’t have the same level of redundancy as the product I’m trying to transition to the Jetson TX2.

I can’t get a good gauge on whether NVidia is really even serious about the embedded market. For a higher end product we need to use Win10 IOT Enterprise and x86, but we’d like to incorporate some of the functionality that GPU computing offers (along with reusing CUDA code that the lower end product has). So this generally means finding something along the lines of a GTX 1030 MXM module, and there are companies that sell them. But, when talking to NVidia they are basically “we don’t support using the GTX1030 in Enterprise applications”. How else am I supposed to add Deep learning, and GPU computing to an x86 platform? Plus the driver from NVidia simply says “this OS is not supported”.

The market is completely ripe for NVidia to gobble up market share, but NVidia won’t put themselves into a position of doing so.