LOOKING FOR THE RIGHT
INDUSTRIAL SOLUTION?
Discover embedded, communication and automation solutions tailored to your application.
- Intel Movidius Myriad X Integration:
The EMPA-I101-C1-A71 incorporates the Intel Movidius Myriad X VPU, a high-performance, low-power AI processor optimized for deep learning inference and computer vision workloads at the edge.
- mPCIe Interface:
Designed for easy integration into existing embedded systems, the module connects via a standard Mini PCI Express (mPCIe) slot, providing a compact and standardized form factor.
- Low Power Design:
With a typical thermal design power (TDP) of only 4W, the module is optimized for energy-efficient AI acceleration, making it suitable for fanless and thermally constrained environments.
- AI Framework Support:
Compatible with major deep learning frameworks including TensorFlow, Caffe, and MXNet, enabling developers to deploy pre-trained models seamlessly.
- Intel OpenVINO™ Toolkit Support:
Fully supports Intel’s OpenVINO™ Toolkit, allowing developers to optimize and deploy AI inference models efficiently across Intel hardware platforms.
- Operating System Compatibility:
Supports both Windows 10 (64-bit) and Ubuntu 18.04 (64-bit), ensuring flexibility for various embedded and industrial computing environments.
- Industrial-Grade Reliability:
Built with a 30µ” gold finger connector for enhanced durability and signal integrity, ensuring long-term reliability in industrial applications.
- Regulatory Compliance:
Complies with CE and FCC Class A standards, ensuring electromagnetic compatibility and safety for industrial and commercial use.
- Warranty:
Backed by a 3-year manufacturer warranty, reflecting Innodisk’s commitment to quality and long-term product support.
The EMPA-I101-C1-A71 incorporates the Intel Movidius Myriad X VPU, a high-performance, low-power AI processor optimized for deep learning inference and computer vision workloads at the edge.
- mPCIe Interface:
Designed for easy integration into existing embedded systems, the module connects via a standard Mini PCI Express (mPCIe) slot, providing a compact and standardized form factor.
- Low Power Design:
With a typical thermal design power (TDP) of only 4W, the module is optimized for energy-efficient AI acceleration, making it suitable for fanless and thermally constrained environments.
- AI Framework Support:
Compatible with major deep learning frameworks including TensorFlow, Caffe, and MXNet, enabling developers to deploy pre-trained models seamlessly.
- Intel OpenVINO™ Toolkit Support:
Fully supports Intel’s OpenVINO™ Toolkit, allowing developers to optimize and deploy AI inference models efficiently across Intel hardware platforms.
- Operating System Compatibility:
Supports both Windows 10 (64-bit) and Ubuntu 18.04 (64-bit), ensuring flexibility for various embedded and industrial computing environments.
- Industrial-Grade Reliability:
Built with a 30µ” gold finger connector for enhanced durability and signal integrity, ensuring long-term reliability in industrial applications.
- Regulatory Compliance:
Complies with CE and FCC Class A standards, ensuring electromagnetic compatibility and safety for industrial and commercial use.
- Warranty:
Backed by a 3-year manufacturer warranty, reflecting Innodisk’s commitment to quality and long-term product support.
Comprehensive Product Summary: EMPA-I101-C1-A71 – mPCIe to Intel Movidius Myriad X AI Module
### 1. Product Overview
Product Name: mPCIe to Intel Movidius Myriad X AI Module
Model Number: EMPA-I101-C1-A71
Series: EMPA-I101 Series
Article Number: EMPA-I101-C1-A71
Manufacturer: Innodisk Corporation
Product Type: Mini PCI Express (mPCIe) to Intel Movidius Myriad X AI Accelerator Module
Product Description: The Innodisk EMPA-I101-C1-A71 is a compact, low-power AI acceleration module designed to integrate Intel’s Movidius Myriad X Vision Processing Unit (VPU) into embedded systems via the mPCIe interface. It enables efficient edge AI inference, computer vision, and deep learning applications in space-constrained industrial and embedded environments.
Manufacturer Contact Information:
- Headquarters (Taiwan):
5F., No. 237, Sec. 1, Datong Rd., Xizhi Dist., New Taipei City 221, Taiwan
Tel: +886-2-7703-3000
Email: sales@innodisk.com
- Regional Offices:
- USA: usasales@innodisk.com | +1-510-770-9421
- Europe: eusales@innodisk.com | +31-40-3045-400
- Japan: jpsales@innodisk.com | +81-3-6667-0161
- China: sales_cn@innodisk.com | +86-755-2167-3689
Official Website: [www.innodisk.com](http://www.innodisk.com)
Release Date: September 5, 2020
Copyright: © 2019 Innodisk Corporation. All rights reserved. Specifications subject to change without prior notice.
### 2. Key Features
- Intel Movidius Myriad X Integration:
The EMPA-I101-C1-A71 incorporates the Intel Movidius Myriad X VPU, a high-performance, low-power AI processor optimized for deep learning inference and computer vision workloads at the edge.
- mPCIe Interface:
Designed for easy integration into existing embedded systems, the module connects via a standard Mini PCI Express (mPCIe) slot, providing a compact and standardized form factor.
- Low Power Design:
With a typical thermal design power (TDP) of only 4W, the module is optimized for energy-efficient AI acceleration, making it suitable for fanless and thermally constrained environments.
- AI Framework Support:
Compatible with major deep learning frameworks including TensorFlow, Caffe, and MXNet, enabling developers to deploy pre-trained models seamlessly.
- Intel OpenVINO™ Toolkit Support:
Fully supports Intel’s OpenVINO™ Toolkit, allowing developers to optimize and deploy AI inference models efficiently across Intel hardware platforms.
- Operating System Compatibility:
Supports both Windows 10 (64-bit) and Ubuntu 18.04 (64-bit), ensuring flexibility for various embedded and industrial computing environments.
- Industrial-Grade Reliability:
Built with a 30µ” gold finger connector for enhanced durability and signal integrity, ensuring long-term reliability in industrial applications.
- Regulatory Compliance:
Complies with CE and FCC Class A standards, ensuring electromagnetic compatibility and safety for industrial and commercial use.
- Warranty:
Backed by a 3-year manufacturer warranty, reflecting Innodisk’s commitment to quality and long-term product support.
### 3. Technical Specifications
| Parameter | Specification |
|----------------|-------------------|
| Form Factor | Mini PCI Express (mPCIe) |
| Interface | PCI Express 2.0 x1 |
| Processor | Intel Movidius Myriad X Vision Processing Unit (VPU) |
| Thermal Design Power (TDP) | 4W |
| Supported Frameworks | TensorFlow, Caffe, MXNet |
| Supported OS | Windows 10 (64-bit), Ubuntu 18.04 (64-bit) |
| Operating Temperature Range | 0°C to +60°C |
| Dimensions (W x L x H) | With Heatsink: 30 x 50.9 x 35.15 mm; Without Heatsink: 30 x 50.9 x 7.65 mm |
| Weight | With Heatsink: 29.47 g; Without Heatsink: 6.25 g |
| Connector Plating | 30µ” gold finger |
| Certifications | CE, FCC Class A |
| Warranty | 3 years |
| Toolkit Support | Intel OpenVINO™ Toolkit |
| Power Supply | Supplied via mPCIe interface |
| Product Code | EMPA-I101-C1-A71 |
### 4. Functionality
The EMPA-I101-C1-A71 functions as an AI acceleration module that bridges the gap between embedded computing platforms and advanced AI inference capabilities. By leveraging the Intel Movidius Myriad X VPU, it enables real-time computer vision and deep learning inference directly at the edge, reducing latency and dependence on cloud-based processing.
How It Works:
- The module connects to a host system via the mPCIe interface.
- Once installed, it operates as a dedicated AI co-processor, offloading neural network inference tasks from the CPU or GPU.
- Developers can deploy AI models using the Intel OpenVINO™ Toolkit, which optimizes models trained in TensorFlow, Caffe, or MXNet for execution on the Myriad X VPU.
- The module processes visual and sensor data locally, enabling applications such as object detection, facial recognition, motion tracking, and anomaly detection.
Use Cases and Applications:
- Industrial Automation: Real-time defect detection, quality inspection, and predictive maintenance.
- Smart Surveillance: Edge-based video analytics, facial recognition, and behavior analysis.
- Retail Analytics: Customer tracking, heat mapping, and product interaction analysis.
- Robotics: Object recognition, navigation, and autonomous decision-making.
- Healthcare Devices: Medical imaging analysis and patient monitoring.
- Transportation: Driver monitoring, traffic analysis, and smart vehicle systems.
Compatibility:
The EMPA-I101-C1-A71 is compatible with embedded systems, industrial PCs, and edge computing devices equipped with an mPCIe slot and running supported operating systems (Windows 10 64-bit or Ubuntu 18.04 64-bit).
### 5. Components & Accessories
Included Items:
- EMPA-I101-C1-A71 module (mPCIe to Intel Movidius Myriad X AI module)
- Integrated heatsink (for EMPA-I101-C1 variant)
- Documentation and installation guide (digital or printed, depending on shipment)
Optional Accessories:
- Passive or active cooling solutions (depending on system thermal design)
- Adapter boards or risers for non-mPCIe systems
- Additional mounting hardware for custom enclosures
Connectors:
- Standard mPCIe edge connector with 30µ” gold plating for enhanced conductivity and durability.
- Power and data transmission are handled through the mPCIe interface, eliminating the need for external power connectors.
### 6. Installation & Setup
System Requirements:
- Host system with an available mPCIe slot (PCIe 2.0 x1 interface)
- Supported operating system: Windows 10 (64-bit) or Ubuntu 18.04 (64-bit)
- Intel OpenVINO™ Toolkit installed (available from Intel’s official documentation site)
Installation Steps:
1. Power off the host system and disconnect it from the power source.
2. Insert the EMPA-I101-C1-A71 module into the available mPCIe slot, ensuring proper alignment of the gold fingers.
3. Secure the module using the mounting screw provided by the host system or chassis.
4. If using the version with a heatsink, ensure adequate airflow or passive cooling within the enclosure.
5. Power on the system and verify device recognition in the BIOS or operating system.
6. Install the Intel OpenVINO™ Toolkit and configure the environment for AI model deployment.
7. Load and run AI inference models using supported frameworks (TensorFlow, Caffe, MXNet).
Mounting Options:
- Direct mPCIe slot installation on embedded boards or industrial PCs.
- Optional riser or adapter for systems with non-standard slot orientations.
Configuration:
- Software configuration is performed through the OpenVINO™ environment.
- Model optimization and deployment are managed via the toolkit’s Model Optimizer and Inference Engine components.
### 7. Performance & Capabilities
Processing Power:
The Intel Movidius Myriad X VPU integrates a Neural Compute Engine (NCE) capable of delivering high-performance deep learning inference at low power consumption. It supports multiple neural network layers and parallel processing for real-time AI workloads.
Speed and Efficiency:
- Optimized for low-latency inference on edge devices.
- Typical power consumption of 4W ensures efficient thermal management.
- Capable of handling multiple video streams or sensor inputs simultaneously, depending on model complexity.
Capacity and Scalability:
- Supports deployment of multiple AI models concurrently.
- Scalable architecture allows integration of multiple modules for higher aggregate inference throughput.
Thermal Management:
- Available in two configurations: with or without heatsink.
- The heatsink-equipped version (EMPA-I101-C1) provides enhanced thermal dissipation for sustained performance under continuous workloads.
Limitations:
- Operating temperature limited to 0°C ~ +60°C; not suitable for extreme temperature environments without additional thermal management.
- Requires host system support for mPCIe interface and compatible OS.
- AI performance dependent on model complexity and optimization level within OpenVINO™.
### 8. Standards & Compliance
Regulatory Certifications:
- CE (Conformité Européenne): Ensures compliance with European safety, health, and environmental protection standards.
- FCC Class A: Certified for electromagnetic compatibility in industrial and commercial environments.
Industry Standards:
- PCI Express 2.0 x1: Compliant with PCI-SIG specifications for data communication and interoperability.
- mPCIe Form Factor: Conforms to Mini PCI Express mechanical and electrical standards.
- Intel OpenVINO™ Compatibility: Fully compatible with Intel’s AI software ecosystem for model optimization and deployment.
Environmental Compliance:
- RoHS-compliant materials and manufacturing processes.
- Designed for industrial-grade reliability and long lifecycle support.
### 9. Additional Information
Warranty:
- 3-year limited warranty covering manufacturing defects and component reliability.
- Extended support options available through Innodisk’s regional service centers.
Technical Support:
- Global support network with dedicated regional offices in Taiwan, USA, Europe, Japan, and China.
- Access to online documentation, driver downloads, and software resources via [www.innodisk.com](http://www.innodisk.com).
Maintenance:
- No active maintenance required beyond standard system cleaning and thermal management.
- Firmware and software updates provided through Intel’s OpenVINO™ platform and Innodisk’s support channels.
Software Updates:
- Users are advised to download the latest version of the Intel OpenVINO™ Toolkit from Intel’s official documentation site:
[https://docs.openvinotoolkit.org/latest/index.html](https://docs.openvinotoolkit.org/latest/index.html)
Product Lifecycle and Availability:
- Designed for long-term availability in industrial and embedded markets.
- Backward compatibility with existing mPCIe-based systems ensures ease of integration and upgrade.
Order Information:
- Model: EMPA-I101-C1
- Description: mPCIe to One Intel Movidius Myriad X AI Module, with Heatsink
- Ordering Code: EMPA-I101-C1-A71
### Summary
The Innodisk EMPA-I101-C1-A71 is a compact, energy-efficient AI acceleration module that brings Intel Movidius Myriad X VPU capabilities to embedded and industrial systems via the mPCIe interface. It delivers powerful edge AI inference performance while maintaining low power consumption and a small footprint. With support for major AI frameworks, Intel’s OpenVINO™ Toolkit, and both Windows and Linux operating systems, it provides a flexible and scalable solution for deploying AI at the edge. Its industrial-grade design, CE/FCC compliance, and 3-year warranty make it a reliable choice for applications in automation, surveillance, robotics, and smart infrastructure.
### 1. Product Overview
Product Name: mPCIe to Intel Movidius Myriad X AI Module
Model Number: EMPA-I101-C1-A71
Series: EMPA-I101 Series
Article Number: EMPA-I101-C1-A71
Manufacturer: Innodisk Corporation
Product Type: Mini PCI Express (mPCIe) to Intel Movidius Myriad X AI Accelerator Module
Product Description: The Innodisk EMPA-I101-C1-A71 is a compact, low-power AI acceleration module designed to integrate Intel’s Movidius Myriad X Vision Processing Unit (VPU) into embedded systems via the mPCIe interface. It enables efficient edge AI inference, computer vision, and deep learning applications in space-constrained industrial and embedded environments.
Manufacturer Contact Information:
- Headquarters (Taiwan):
5F., No. 237, Sec. 1, Datong Rd., Xizhi Dist., New Taipei City 221, Taiwan
Tel: +886-2-7703-3000
Email: sales@innodisk.com
- Regional Offices:
- USA: usasales@innodisk.com | +1-510-770-9421
- Europe: eusales@innodisk.com | +31-40-3045-400
- Japan: jpsales@innodisk.com | +81-3-6667-0161
- China: sales_cn@innodisk.com | +86-755-2167-3689
Official Website: [www.innodisk.com](http://www.innodisk.com)
Release Date: September 5, 2020
Copyright: © 2019 Innodisk Corporation. All rights reserved. Specifications subject to change without prior notice.
### 2. Key Features
- Intel Movidius Myriad X Integration:
The EMPA-I101-C1-A71 incorporates the Intel Movidius Myriad X VPU, a high-performance, low-power AI processor optimized for deep learning inference and computer vision workloads at the edge.
- mPCIe Interface:
Designed for easy integration into existing embedded systems, the module connects via a standard Mini PCI Express (mPCIe) slot, providing a compact and standardized form factor.
- Low Power Design:
With a typical thermal design power (TDP) of only 4W, the module is optimized for energy-efficient AI acceleration, making it suitable for fanless and thermally constrained environments.
- AI Framework Support:
Compatible with major deep learning frameworks including TensorFlow, Caffe, and MXNet, enabling developers to deploy pre-trained models seamlessly.
- Intel OpenVINO™ Toolkit Support:
Fully supports Intel’s OpenVINO™ Toolkit, allowing developers to optimize and deploy AI inference models efficiently across Intel hardware platforms.
- Operating System Compatibility:
Supports both Windows 10 (64-bit) and Ubuntu 18.04 (64-bit), ensuring flexibility for various embedded and industrial computing environments.
- Industrial-Grade Reliability:
Built with a 30µ” gold finger connector for enhanced durability and signal integrity, ensuring long-term reliability in industrial applications.
- Regulatory Compliance:
Complies with CE and FCC Class A standards, ensuring electromagnetic compatibility and safety for industrial and commercial use.
- Warranty:
Backed by a 3-year manufacturer warranty, reflecting Innodisk’s commitment to quality and long-term product support.
### 3. Technical Specifications
| Parameter | Specification |
|----------------|-------------------|
| Form Factor | Mini PCI Express (mPCIe) |
| Interface | PCI Express 2.0 x1 |
| Processor | Intel Movidius Myriad X Vision Processing Unit (VPU) |
| Thermal Design Power (TDP) | 4W |
| Supported Frameworks | TensorFlow, Caffe, MXNet |
| Supported OS | Windows 10 (64-bit), Ubuntu 18.04 (64-bit) |
| Operating Temperature Range | 0°C to +60°C |
| Dimensions (W x L x H) | With Heatsink: 30 x 50.9 x 35.15 mm; Without Heatsink: 30 x 50.9 x 7.65 mm |
| Weight | With Heatsink: 29.47 g; Without Heatsink: 6.25 g |
| Connector Plating | 30µ” gold finger |
| Certifications | CE, FCC Class A |
| Warranty | 3 years |
| Toolkit Support | Intel OpenVINO™ Toolkit |
| Power Supply | Supplied via mPCIe interface |
| Product Code | EMPA-I101-C1-A71 |
### 4. Functionality
The EMPA-I101-C1-A71 functions as an AI acceleration module that bridges the gap between embedded computing platforms and advanced AI inference capabilities. By leveraging the Intel Movidius Myriad X VPU, it enables real-time computer vision and deep learning inference directly at the edge, reducing latency and dependence on cloud-based processing.
How It Works:
- The module connects to a host system via the mPCIe interface.
- Once installed, it operates as a dedicated AI co-processor, offloading neural network inference tasks from the CPU or GPU.
- Developers can deploy AI models using the Intel OpenVINO™ Toolkit, which optimizes models trained in TensorFlow, Caffe, or MXNet for execution on the Myriad X VPU.
- The module processes visual and sensor data locally, enabling applications such as object detection, facial recognition, motion tracking, and anomaly detection.
Use Cases and Applications:
- Industrial Automation: Real-time defect detection, quality inspection, and predictive maintenance.
- Smart Surveillance: Edge-based video analytics, facial recognition, and behavior analysis.
- Retail Analytics: Customer tracking, heat mapping, and product interaction analysis.
- Robotics: Object recognition, navigation, and autonomous decision-making.
- Healthcare Devices: Medical imaging analysis and patient monitoring.
- Transportation: Driver monitoring, traffic analysis, and smart vehicle systems.
Compatibility:
The EMPA-I101-C1-A71 is compatible with embedded systems, industrial PCs, and edge computing devices equipped with an mPCIe slot and running supported operating systems (Windows 10 64-bit or Ubuntu 18.04 64-bit).
### 5. Components & Accessories
Included Items:
- EMPA-I101-C1-A71 module (mPCIe to Intel Movidius Myriad X AI module)
- Integrated heatsink (for EMPA-I101-C1 variant)
- Documentation and installation guide (digital or printed, depending on shipment)
Optional Accessories:
- Passive or active cooling solutions (depending on system thermal design)
- Adapter boards or risers for non-mPCIe systems
- Additional mounting hardware for custom enclosures
Connectors:
- Standard mPCIe edge connector with 30µ” gold plating for enhanced conductivity and durability.
- Power and data transmission are handled through the mPCIe interface, eliminating the need for external power connectors.
### 6. Installation & Setup
System Requirements:
- Host system with an available mPCIe slot (PCIe 2.0 x1 interface)
- Supported operating system: Windows 10 (64-bit) or Ubuntu 18.04 (64-bit)
- Intel OpenVINO™ Toolkit installed (available from Intel’s official documentation site)
Installation Steps:
1. Power off the host system and disconnect it from the power source.
2. Insert the EMPA-I101-C1-A71 module into the available mPCIe slot, ensuring proper alignment of the gold fingers.
3. Secure the module using the mounting screw provided by the host system or chassis.
4. If using the version with a heatsink, ensure adequate airflow or passive cooling within the enclosure.
5. Power on the system and verify device recognition in the BIOS or operating system.
6. Install the Intel OpenVINO™ Toolkit and configure the environment for AI model deployment.
7. Load and run AI inference models using supported frameworks (TensorFlow, Caffe, MXNet).
Mounting Options:
- Direct mPCIe slot installation on embedded boards or industrial PCs.
- Optional riser or adapter for systems with non-standard slot orientations.
Configuration:
- Software configuration is performed through the OpenVINO™ environment.
- Model optimization and deployment are managed via the toolkit’s Model Optimizer and Inference Engine components.
### 7. Performance & Capabilities
Processing Power:
The Intel Movidius Myriad X VPU integrates a Neural Compute Engine (NCE) capable of delivering high-performance deep learning inference at low power consumption. It supports multiple neural network layers and parallel processing for real-time AI workloads.
Speed and Efficiency:
- Optimized for low-latency inference on edge devices.
- Typical power consumption of 4W ensures efficient thermal management.
- Capable of handling multiple video streams or sensor inputs simultaneously, depending on model complexity.
Capacity and Scalability:
- Supports deployment of multiple AI models concurrently.
- Scalable architecture allows integration of multiple modules for higher aggregate inference throughput.
Thermal Management:
- Available in two configurations: with or without heatsink.
- The heatsink-equipped version (EMPA-I101-C1) provides enhanced thermal dissipation for sustained performance under continuous workloads.
Limitations:
- Operating temperature limited to 0°C ~ +60°C; not suitable for extreme temperature environments without additional thermal management.
- Requires host system support for mPCIe interface and compatible OS.
- AI performance dependent on model complexity and optimization level within OpenVINO™.
### 8. Standards & Compliance
Regulatory Certifications:
- CE (Conformité Européenne): Ensures compliance with European safety, health, and environmental protection standards.
- FCC Class A: Certified for electromagnetic compatibility in industrial and commercial environments.
Industry Standards:
- PCI Express 2.0 x1: Compliant with PCI-SIG specifications for data communication and interoperability.
- mPCIe Form Factor: Conforms to Mini PCI Express mechanical and electrical standards.
- Intel OpenVINO™ Compatibility: Fully compatible with Intel’s AI software ecosystem for model optimization and deployment.
Environmental Compliance:
- RoHS-compliant materials and manufacturing processes.
- Designed for industrial-grade reliability and long lifecycle support.
### 9. Additional Information
Warranty:
- 3-year limited warranty covering manufacturing defects and component reliability.
- Extended support options available through Innodisk’s regional service centers.
Technical Support:
- Global support network with dedicated regional offices in Taiwan, USA, Europe, Japan, and China.
- Access to online documentation, driver downloads, and software resources via [www.innodisk.com](http://www.innodisk.com).
Maintenance:
- No active maintenance required beyond standard system cleaning and thermal management.
- Firmware and software updates provided through Intel’s OpenVINO™ platform and Innodisk’s support channels.
Software Updates:
- Users are advised to download the latest version of the Intel OpenVINO™ Toolkit from Intel’s official documentation site:
[https://docs.openvinotoolkit.org/latest/index.html](https://docs.openvinotoolkit.org/latest/index.html)
Product Lifecycle and Availability:
- Designed for long-term availability in industrial and embedded markets.
- Backward compatibility with existing mPCIe-based systems ensures ease of integration and upgrade.
Order Information:
- Model: EMPA-I101-C1
- Description: mPCIe to One Intel Movidius Myriad X AI Module, with Heatsink
- Ordering Code: EMPA-I101-C1-A71
### Summary
The Innodisk EMPA-I101-C1-A71 is a compact, energy-efficient AI acceleration module that brings Intel Movidius Myriad X VPU capabilities to embedded and industrial systems via the mPCIe interface. It delivers powerful edge AI inference performance while maintaining low power consumption and a small footprint. With support for major AI frameworks, Intel’s OpenVINO™ Toolkit, and both Windows and Linux operating systems, it provides a flexible and scalable solution for deploying AI at the edge. Its industrial-grade design, CE/FCC compliance, and 3-year warranty make it a reliable choice for applications in automation, surveillance, robotics, and smart infrastructure.
-
EMPA-I101_Datasheet.pdf
PDF Catalog
No accessories available.