Qualcomm Dragonwing | Snapdragon Platform for IoT/Industrial Embedded

Snapdragon for IoT/Industrial Embedded | What is Dragonwing?

We introduce "Qualcomm Dragonwing," a platform that expands upon Qualcomm Snapdragon, a processor with a proven track record in smartphones, for IoT/industrial embedded devices.

Dragonwing is a platform for edge computing that integrates a variety of cores, including CPU, GPU, DSP, NPU, ISP, and VPU, with low power consumption and high performance. It features high-performance video encoding/decoding, excellent image processing capabilities, and rich connectivity including carrier communication, Wi-Fi, and Bluetooth. In addition, the NPU, which is specialized for AI processing, enables efficient on-device execution of AI models, including image recognition models and large-scale language models (LLMs). This gives it a significant advantage in implementing edge AI where real-time performance and security are required.

At Macnica, we offer a wide range of products to comprehensively support our customers' development, including not only chipsets but also SOMs (System on Modules) and edge computing devices. Furthermore, we collaborate with partner companies to provide optimal solutions for software such as operating systems, middleware, and applications that run on Dragonwing. We will introduce specific devices and propose total solutions tailored to your needs, so please feel free to contact us.

Qualcomm Dragonwing IQ Product Family

Dragonwing IQ9

The ideal choice for mission-critical applications requiring high reliability, such as factory automation, advanced robotics, and autonomous systems.
Go to product page >

Dragonwing IQ8
The ideal choice for factory automation, industrial robotics, machine vision, and AI applications that require large-scale computational processing.

Go to product page >

Dragonwing IQ6
The ideal choice for industrial gateways, controllers, HMI devices, and handheld industrial equipment where cost and reliability are paramount.

Go to product page >

Qualcomm Dragonwing Q Product Family

Dragonwing Q8

The perfect choice for next-generation smart devices that demand immersive multimedia, powerful on-device AI, and enterprise-grade security at the edge.

Go to product page >

Dragonwing Q7

The perfect choice for smart home and enterprise devices that require rich multimedia experiences, real-time on-device AI, and privacy-first edge processing.

Go to product page >

Dragonwing Q6

The ideal choice for advanced IoT applications requiring high-performance, low-power computing and robust edge AI capabilities.

Go to product page >

Dragonwing Q5

A low-power SoC that enables flexible industrial IoT deployments, with expandable AI, camera, and computing capabilities through modular performance packs.

Go to product page >

Technical information page

Qualcomm Dragonwing/Snapdragon performance comparison

I actually tried it with Qualcomm Dragonwing.

SOM / Board Development Partner

Here are our development partners. Each company offers System-on-Modules (SOMs), Single-Board Computers (SBCs), and Development Kits. Click on each manufacturer's logo for more details.

Development tool related information

Qualcomm AI Hub

Qualcomm AI Hub is a platform that makes it easy to deploy AI models, such as computer vision, speech recognition, and large-scale language models, to edge devices on the Qualcomm platform. It provides one-stop optimization, validation, and deployment of custom models, as well as a library of pre-optimized AI models, supporting the rapid development of AI applications.

Click here to learn more about Qualcomm AI Hub.

AI/ML Developer Workflow

Developers can efficiently run models in ONNX, PyTorch, TensorFlow, and TFLite formats on Qualcomm AI hardware {HTP (NPU), GPU, CPU} using the Qualcomm AI SDK.

AI Hardware

- Qualcomm Kryo CPU: A best-in-class CPU that combines high processing performance with excellent power efficiency.

- Qualcomm Hexagon Tensor Processor (HTP): Also known as NPU/DSP/HMX, this processor is designed for low-power, high-performance AI workloads. To achieve optimal performance, pre-trained models must be quantized to a supported precision format.

- Qualcomm Adreno GPU: A GPU that offers an excellent balance of power and performance, making it suitable for running AI workloads. It supports AI acceleration via the OpenCL kernel and can be used to speed up model preprocessing and postprocessing.

AI software

The AI stack consists of a suite of SDKs designed to maximize the performance of AI hardware accelerators. Developers can select the appropriate SDK for their needs and deploy their AI workloads. TFLite models can be run directly using the TFLite Delegate, but other pre-trained models must be converted to a format executable by the selected SDK.

Qualcomm AI/ML Developer Workflow (External Site)

Edge Impulse (Edge AI Development Platform)

Edge Impulse is an edge AI development platform compatible with Qualcomm SoCs. It allows for the entire process, from data collection and model training to device implementation, to be completed within a web browser. Qualcomm SoCs also offer standard support for high-speed inference utilizing the Hexagon NPU. Regardless of AI development experience, users can quickly launch edge AI solutions on the Qualcomm platform.

Edge Impulse HP (external site)

Arduino by Qualcomm

Arduino by Qualcomm is an Arduino series powered by Qualcomm Dragonwing. It surpasses the limitations of traditional microcontroller-based development, offering processing power to enable more complex applications. It seamlessly covers everything from rapid prototyping in the Arduino environment to full-scale development utilizing Linux, all within a single environment.

The Arduino UNO Q is a "dual-brain" development board that integrates a Qualcomm Dragonwing QRB2210 that runs Linux and an STM32 microcontroller for real-time control into a single board.
Product page is here >