NVIDIA is pleased to announce Jetson AGX Xavier, the latest addition to the Jetson platform with greater than 10x the energy efficiency and more than 20x the performance of it’s predecessor, Jetson TX2. It’s an AI computer for autonomous machines, delivering the performance of a GPU workstation in an embedded module under 30W. With multiple operating modes at 10W, 15W, and 30W, Jetson Xavier has a peak performance of more than an incredible 30 TOPS (teraops) of mixed-precision FP32/FP16/INT8 performance. Please refer to the FAQ for further information.
US, Canada, UK, France, Germany, Italy, Poland, Spain, and China
Additionally, orders can be placed through distributors for:
Japan, Europe, Israel, and Turkey (other regions to follow soon)
Initial orders will be fulfilled on a first-come first-served basis within an estimated six-week time period beginning early September 2018.
The price of the devkit for Registered Developers is $1,299 (limit 1 unit per Registered Developer, prices vary by region). Anyone can qualify for the special price by registering an account at developer.nvidia.com. Additional units are available in quantities 1-10 at $2,499 and quantities >10 are $1,599.
…that would be astonishing…that amount of throughput would exceed any desktop PC currently out…current PCs only go to gen3. I’ve not seen anything other than demo technology for this so far, so that would mean this Jetson will probably be the first practical device to support this (someone else might come out with gen4 before then…we’ll have to wait and see).
I am very excited about this. I am wondering whether VisionWorks will be supported in this board or not. Could you please let us know about the plans of NVIDIA to support VisionWorks and other SDKs that has been previously supported in TX2 and TX1 in this new platform?
can the interface provided manage at least 8 of them? So I’d put 2 Xavier in parallel to manage them all. Maybe I might use the PCIs, using two 4 ports ethernet PCI cards…
Xavier boasts (2x) 4Kp60 | HEVC video decoding. 4k is too little for 12MP…can I split the 8 cameras video stream in 2, and encode the upper and lower part of the images in 2 separate videos, leveraging the HW encoder?
Can it encode 16 video streams (2x8) at the same time? The camera stream is at 64fps. I might lower it to 60 so that it fits Xavier’s 60Hz limit.
Can you clarify how many USB ports are exposed on the Xavier’s DevKit? Looking at videos from Computex, it looks like there’re two USB type C connectors on the same side as 40pin header, and there’s another USB type C connector on the opposite side? Does it mean it has only three USB connectors? Of course, there’s a PCIe slot which can be used by various PCIe USB cards… Also, where is M.2 Key M connector located? Is it sandwiched between Xavier module and DevKit carrier board?
I am curious about the SPDI connector which looks like a micro-USB connector. Does this actually use USB (e.g., serial UART) for audio data, or is it perhaps customized and passes a balanced AES3 signal directly through (meaning an alternate connector versus what is usually scene, but still equivalent to the larger pro studio connector)? For reference on the normal AES3 style connector:
[url]AES3 - Wikipedia
[url]AES3 - Wikipedia
At around 2:23 into the video he shows the micro-USB connector, and this is what I heard:
“…and this micro-USB connector is actually an SPDI serial connection.”
So…he was really saying “FTDI” instead of “SPDI”…I should have known! Makes much more sense. On the other hand, I was rather excited that special support might be added for audio. This is one of my original interests, but limitations on the TK1/TX1/TX2 i2s channels limited options. Incidentally though, is there more than one i2s or other capability to go beyond two channel digital audio?