The Shield TV product can be used as an Android devkit. There is no current plan to release Linux4Tegra although I’d love to see our engineers spend the effort to release it too. They are focused on our Jetson embedded boards so if Linux is your focus, I suggest you consider one of those instead: https://developer.nvidia.com/jetson-tk1
Any updates on this Nicholas? Did you get L4T running on the console? I’m also looking to do some openCV/deep NN work with robotics on this platform.
I am also looking for a X1 dev kit with L4T. What high speed (100MB/s) interface does the shield support in addition to USB3 and gigabit ethernet?
I am also interested in using the Shield TV as a development platform and have purchased one over the Jetson due to the newer Tegra X1 SOC. Is it at all likely that there will be official support for L4T on the Shield TV, or atleast a build of L4T for the X1 SOC that could be ported over?
Even an AARCH64 build of the GPU drivers/CUDA Toolkit for tegra X1 would useful in porting linux to the system.
We’re pushing for it, but at the moment there’s no new updates regarding Linux OS builds.
I do confirm our interest in Linux4Tegra on X1.
We started with Jetson TK1 but as we are using 4K video extensively, the X1 SOC is far better supporting 4K video encoding / decoding, plus the increased GPU power.
Currently, we are developing on rooted Shield Android TV but we do prefer using GStreamer and other Linux toolchain instead of AOSP.
i agree with them we need a X1 dev board at least 4gb ram or at least shield tv with linux4tegra :)
Is there still no chance of this happening?
I am also eagerly awaiting Linux 4 Tegra on the X1. I really hope they release a kit soon the possibilities on the X1 look awesome for Linux X11 development!
I work as a core dev of the GROMACS molecular simulation package and we are very interested in evaluating the Tegra X1. We have ran on the TK1, but sadly it’s slow and outdated hardware, and represents poorly what the future may (will?) look like.
@Keith & others from NVIDIA: As you can see (and verify if you ask further) the embedded and CS/computational science research community is very interested in the (ex-Denver now Tegra) X1 platform. I urge you to listen to these communities and provide us with the means to enable experimentation. The X1 chip, especially given the CUDA/OpenCL support provides amazing possibilities that are greatly hampered by the lack of a devkit and readily available hardware with Linux&CUDA/OpenCL support.
Given that people have done it, it can’t take a huge amount of effort to ensure that e.g. Ubuntu 15.04 or 15.10 works well on the X1 and that reasonably experienced member of the interested communities can get the device flash with a little effort. All that we’d need is a a simple guide for devs that’s more reliable than stuff currently available on the Internet - and more openness, would not hurt either.
May I second all the above comments. I have especially ordered a Shield TV into the UK for the purposes of exploring its usefulness outside of the the Android ecosystem. This is by far the most interesting “dev board” available out there, because it is fast, and well packaged. I have a very specific financial analysis application in mind which requires Linux. Shield TV is the perfect cost-effective 4k visualization terminal solution for the vector graphics-complex use case I plan, and the target market is in the hundreds to thousands of potential clients in a vertical domain. I believe this device deserves much better than being hobbled by Android/Java, most importantly because it is already well packaged. I can literally purchase these and put them into customer locations, without contracting a maker to produce a custom device based on a dev board. Shield TV + 4k monitor + keyboard + Linux and off we go: advanced visualization for less than 800 USD cost all in.
Android, as an OS, just isn’t as flexible as a development environment. From what I could see, for starters, we’re only missing AArch64 GPU drivers. And I really expect those to be released in a timely fashion.
There are Aarch64 Linux drivers, I believe since the CUDA 6.5 release. The support for 64-bit ARM platforms was drummed up first around ISC14. AFAIK at least one development platform from Cirrascale using APM X-Gene has been available (http://www.cirrascale.com/press/press_RM1905D_ARM64_release.aspx). That’s not to say that that driver is trivial to port to the Tegra X1, but I do assume it should not be too hard.
BTW, there are clear indications  that the devkit does exist internally. So, my guess is that it’s probably a management priority, general interest (or the lack of it), $$$ bias of car companies , etc. the reasons for the lack of public Tegra4Linux X1 release.
Question for Nvidia management.
You put out a Jetson TK1 with much fanfare and full Linux stack more than 18 months ago, now the TX1 has been out for all of 2015 already, without giving us any feedback about our upgrade path. It’s weird because you encourage a community, then you seem to ignore it.
What’s the deal on Linux on the TX1/Shield for developers? Shall we assume that you’ve de-prioritized this? Are you betting on Android only?