Stories
Slash Boxes
Comments

SoylentNews is people

posted by martyb on Monday October 05 2020, @09:40PM   Printer-friendly
from the waiting-for-Jane,-Judy,-Elroy,-Rosey,-and-Astro dept.

Nvidia has announced a cheaper version of its $99 Jetson Nano developer kit. The Jetson Nano pairs a quad-core Cortex-A57 ARM CPU with 128 Maxwell GPU cores. The new model has 2 GiB of RAM instead of 4 GiB, drops one of the four USB ports (which may be USB 2.0 instead of 3.0), and drops DisplayPort output.

Elsewhere at NVIDIA's GPU Technology Conference 2020:

NVIDIA Online GTC 2020 Kicks Off Today But No Open-Source Linux Announcement Expected
Quadro No More? NVIDIA Announces Ampere-based RTX A6000 & A40 Video Cards For Pro Visualization
NVIDIA BlueField-2 DPUs Set to Ship In 2021, Roadmaps BlueField-3&4 By 2023


Original Submission

Related Stories

Nvidia Announces Jetson Nano Single-Board Computer 8 comments

Nvidia Announces Jetson Nano Dev Kit & Board: X1 for $99

Today at GTC 2019 Nvidia launched a new member of the Jetson family: The new Jetson Nano. The Jetson family of products represents Nvidia new focus on robotics, AI and autonomous machine applications. A few months back we had the pleasure to have a high level review of the Jetson AGX as well as the Xavier chip that powers it. The biggest concern of the AGX dev kit was its pricing – with retail costs of $2500 ($1299 as part of Nvidia's developer programme), it's massively out of range of most hobbyist users such as our readers.

[...] The Jetson Nano is a full blown single-board-computer in the form of a module. The module form-factor and connector is SO-DIMM and is similar to past Nvidia modules by the company. The goal of the form-factor is to have the most compact form-factor possible, as it is envisioned to be used in a wide variety of applications where a possible customer will design their own connector boards best fit for their design needs.

At the heart of the Nano module we find Nvidia's "Erista" chip which also powered the Tegra X1 in the Nvidia Shield as well as the Nintendo Switch. The variant used in the Nano is a cut-down version though, as the 4 A57 cores only clock up to 1.43GHz and the GPU only has half the cores (128 versus 256 in the full X1) active. The module comes with 4GB of LPDDR4 and a 16GB eMMC module. The Jetson Nano module will be available to interested parties for $129.

$99 without storage.

Related: Nvidia Reveals Jetson Xavier SoC for Robots


Original Submission

This discussion has been archived. No new comments can be posted.
Display Options Threshold/Breakthrough Mark All as Read Mark All as Unread
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
(1)
  • (Score: 0) by Anonymous Coward on Monday October 05 2020, @11:17PM (4 children)

    by Anonymous Coward on Monday October 05 2020, @11:17PM (#1061128)

    No NVDLA so the selling point is 128 CUDA cores, aside from learning the SDK what practically would you do with it?

    • (Score: 2) by JoeMerchant on Tuesday October 06 2020, @12:34AM

      by JoeMerchant (3937) on Tuesday October 06 2020, @12:34AM (#1061149)

      It's a compact "powerful low energy edge" system - you can (probably, with a lot of work) run pretty good facial recognition, or a license plate reader, or similar stuff.

      --
      Україна досі не є частиною Росії Слава Україні🌻 https://news.stanford.edu/2023/02/17/will-russia-ukraine-war-end
    • (Score: 4, Informative) by takyon on Tuesday October 06 2020, @01:07AM (1 child)

      by takyon (881) <takyonNO@SPAMsoylentnews.org> on Tuesday October 06 2020, @01:07AM (#1061151) Journal

      It will run emulators/games better than Raspberry Pi 4. The GPU is the only real selling point, and that 4x Cortex-A57 is getting old. A72 ranges from 16-50% better [tomshardware.com] at the same frequency which might not matter if the GPU is the bottleneck, and the Nintendo Switch also uses A57, at a lower frequency.

      Cutting the RAM in half to 2 GiB isn't going to be good for some machine learning workloads. If you look at the RK3399Pro, that's an ARM chip with a 4 GiB memory limit, but the SoC's NPU can use a separate bank of up to 4 GiB.

      Here's a paper that shows the Jetson Nano crushing Raspberry Pi 4 using a Convolutional Neural Network model:

      10.1109/HORA49412.2020.9152915

      As the dataset size increases, the Jetson TX2 ($400 with 8 GiB RAM) uses up to 6.5 GB in the test, so the Nano and Pi have to tap out. The Pi takes about 5-6x longer than the Nano. There's some problems with the paper but they're Turkish computer scientists so eh, whatever.

      --
      [SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
      • (Score: 2) by JoeMerchant on Tuesday October 06 2020, @02:38AM

        by JoeMerchant (3937) on Tuesday October 06 2020, @02:38AM (#1061161)

        You wouldn't use this to learn/train, but it should be pretty killer for running the nets once they're trained.

        --
        Україна досі не є частиною Росії Слава Україні🌻 https://news.stanford.edu/2023/02/17/will-russia-ukraine-war-end
    • (Score: 2) by SunTzuWarmaster on Tuesday October 06 2020, @06:37PM

      by SunTzuWarmaster (3971) on Tuesday October 06 2020, @06:37PM (#1061347)
      It runs neural networks. So, you know, you bolt it to things that need something like that. Security cameras (with embedded face detection and 'odd behavior' flagging), sound dampeners (dynamically predicting frequencies of sound to nullify), microwave sensors (determine the amount of chicken/fish/whatever based on input and reflected microwaves), etc.
(1)