Jupiter NX SoM – An NVIDIA Jetson Nano/NX alternative powered by SpacemiT K1 octa-core RISC-V SoC

The Jupiter NX is a system-on-module (SoM) powered by a SpacemIT octa-core 64-bit RISC-V SoC that follows NVIDIA’s Jetson Nano / Xavier NX form factor for compatibility with existing carrier boards.

Designed by Shenzhen Milk-V Technology, the module supports up to 16GB LPDDR4x memory and 32GB eMMC flash, and integrates two gigabit Ethernet PHYs and a WiFi 6 & Bluetooth 5.2 wireless module.

Jupiter NX

Jupiter NX specifications:

  • SoC – SpacemIT K1/M1
    • CPU – 8-core X60 RISC-V (RV64GCVB) processor @ 1.6 GHz
    • GPU – Imagination IMG BXE-2-32 @ 819 MHz with support for OpenGL ES3.2, Vulkan 1.3, OpenCL 3.0; 20 GFLOPS
    • VPU
      • H.265, H.264, VP8, VP9, MPEG4, MPEG2 decoder up to 4K @ 60fps
      • H.265, H.264, VP8, VP9 encoder up to 4K @ 30fps
      • Support simultaneous processing
        • 1080p60 encoding + 1080p60 decoding
        • 1080p30 H.264/H.265 encoding + 4Kp30 H.264/H.265 decoding
    • AI performance – 2.0 TOPS (INT8) through “CPU core fusion”
    • RVA 22 Profile RVV 1.0 compliant
  • System Memory – 2GB, 4GB,  8GB, or 16GB LPDDR4X
  • Storage
    • SPI flash for bootloader
    • Optional 16GB or 32GB eMMC flash
  • Networking
    • 2x Gigabit Ethernet PHY
    • WiFi 6 and Bluetooth 5.2
  • 260-pin SO-DIMM edge connector
    • Storage – MMC Interface
    • Display Interfaces
      • HDMI Interface up to 1920×1440 @ 60FPS
      • 4-lane MIPI DSI
    • Camera interfaces – 4-lane MIPI CSI, 2x 2-lane 2x MIPI CSI
    • Audio – 2x I2S
    • Networking – 2x Gigabit Ethernet
    • USB
      • USB 3.0 (combo with PCIe Gen2 x1)
      • USB 2.0 OTG
      • USB 2.0 host
    • PCIe – 1x 2-lane + 2 x1-lane PCIe Gen2
    • Low-speed I/Os
      • Up to 55x GPIO
      • Up to 3x UART, 2x I2C, 1x SPI
      • Up to 3x PWM
  • Supply Voltage – 5V DC
  • Dimensions – 69.6 x 45 mm (NVIDIA Jetson Nano, TX2 NX, Orin Nano, Orin NX form factor)
  • Temperature Range – -40 to +85°C

RISC-V Jetson Nano alternativeThe company provides optimized Ubuntu and Fedora images for the Jupiter NX module. The Spacengine development tool optimizes AI models to run on the X60 cores with the best possible performance. I could not find documentation specific to the Jupiter NX module per se, but the documentation for the Jupiter Mini-ITX motherboard should offer a starting point since it’s based on the sme SpacemIT K1/M1 RISC-V SoC.

Compared to the original NVIDIA Jetson Nano (4GB), the RISC-V-based Jupiter NX (4GB) is several times cheaper while delivering four times the AI performance per Watt, according to the illustration below. However, readers should note that the TOPS/W advantage might not be true anymore with recent modules like the NVIDIA Jetson Orin Nano.

Jupiter NX vs Jetson Nano price performance
Jupiter NX vs Jetson Nano

There’s no carrier board specific to the Jupitex NX, and instead, we’re told the SoM has been tested with the Jetson Nano Developer Kit carrier board. The company is also working on the cluster-08 cluster carrier board that takes up to eight Jupiter NX or upcoming Megrez NX modules. The latter is based on the ESWIN EIC7700X SoC with 19.95 TOPS of AI performance.

Other carrier boards for the Jetson Nano/NX module might also work, but there are often some small incompatibilities between SoMs and carrier boards, so testing will be required on each base board.

Cluster08 carrier board
Cluster08 carrier board (Coming soon)

The Jupiter NX is available for purchase on Arace for $49.90 for the 4GB/16GB model, and $109.90 for the 16GB/32GB variant. The product page has limited information for now, but additional details might be added over time.

Via Liliputing.

Share this:
FacebookTwitterHacker NewsSlashdotRedditLinkedInPinterestFlipboardMeWeLineEmailShare

Support CNX Software! Donate via cryptocurrencies, become a Patron on Patreon, or purchase goods on Amazon or Aliexpress

Radxa Orion O6 Armv9 mini-ITX motherboard

Leave a Reply

Your email address will not be published. Required fields are marked *

Boardcon CM3588 Rockchip RK3588 System-on-Module designed for AI and IoT applications
Boardcon CM3588 Rockchip RK3588 System-on-Module designed for AI and IoT applications