- Browse Categories
- YQFH Drone
- DJI Drone
- Other Drone
- Drone Payloads
- Robot
- Pipeline Detection Instrument
- 3D Laser Scanner
- Industrial Measurement Devices
- Weather Station Series
- Optical Survey Instruments
- GNSS RTK & CORS System
- BDS System Products
- Visual Displacement Monitoring System
- UWB Indoor Positioning System
- Hydrology Products
- Others
- Spectrometer
- Wet Blast Machine
- SDI Soil Tester
- Pengjin Smart Flying Drone Patrol and Inspection System
- Supercomputing
- Module
- Viscometer
- Electric Hygro-thermograph
- Echo Sounder
- Processing Software
- Keson Measuring Tools
Atlas 200I A2 acceleration module, Atlas 200I DK A2 AI Developer Kit Acceleration Module
The Atlas 200I A2 acceleration module can accelerate AI applications such as target recognition and image classification on edge-end devices, widely used in edge-end AI scenarios such as smart edge devices, robots, drones, and intelligent industrial control.
1. Core Specifications and Parameters
• Processor (NPU): Integrated Ascend 310 series (Ascend 310B) AI processor
• AI Computing Power: Available in versions offering either 20 TOPS INT8 (10 TFLOPS FP16) or 8 TOPS INT8 (4 TFLOPS FP16)
• CPU: 4-core ARM processor (1.6 GHz for the 20 TOPS version; 1.0 GHz for the 8 TOPS version)
• Memory (RAM): LPDDR4X; available in 12GB / 8GB / 4GB (for the 20 TOPS version) or 4GB (for the 8 TOPS version); supports ECC
• Audio/Video Codec: Supports H.264 / H.265 hardware encoding and decoding
• Power Consumption: Approx. 25W (20 TOPS version) / 21W (8 TOPS version)
• Operating Temperature: -20°C to +103°C
• Physical Dimensions: 82 x 60 x 7 mm (MXM form factor, suitable for embedded integration)
2. Key Application Scenarios
• Edge AI and Intelligent Terminals: Object recognition and image classification for applications such as intelligent surveillance, industrial control, smart robotics, drones, and intelligent transportation systems.
• Video and Data Analytics: Data stream analysis and real-time AI inference on video servers.
• Development and Education: Used in conjunction with the Atlas 200I DK A2 developer kit, it is frequently employed for university-level instruction, AI algorithm validation, and the rapid deployment of edge applications (supports the Huawei Ascend CANN software stack).
3. Technical Specifications
|
Products |
Atlas 200I A2（20 TOPS） |
Atlas 200I A2（8 TOPS） |
|
AI computing power |
20 TOPS INT8， 10 TFLOPS FP16 |
8 TOPS INT8， 4 TFLOPS FP16 |
|
Memory specifications |
LPDDR4X, 12 / 8 / 4 GB, total bandwidth 51.2 / 34.1 / 34.1 GB/s, supports ECC |
LPDDR4X. 4GB, total bandwidth 25.6GB/s, supports ECC |
|
CPU computing power |
4core * 1.6GHz |
4core * 1.0GHz |
|
Encoding and decoding capabilities |
Supports H.264 / H.265 hardware decoding, 40 channels of 1080P at 30FPS, 4 channels at 4K (3840 x 2160) at 75FPS |
Supports H.264 / H.265 hardware decoding, 20 channels of 1080P 30FPS, 2 channels of 4K (3840 x 2160) 75FPS |
|
High-speed interface |
High speed: 8 lanes, supports SGMII (1.25Gbps), 1000BASE-R (1.25Gbps), USB3.0 (5Gbps), SATA 3.0 (6Gbps), backward compatible with SATA 2.0 (3Gbps) and SATA 1.0 (1.5Gbps), PCIe Gen3 (8Gbps), and PCIe Gen2 (5Gbps) and PCIe downcompatible Gen1 (2.5Gbps) is flexibly configurable |
|
|
Low-speed interface |
UART * 5 / I2C * 4 / SPI * 2 / CAN * 4 |
|
|
Typical power consumption |
25W |
21W |
|
Operating environment temperature |
-20℃～+103℃（-4℉～+217.4℉） |
|
|
Structural dimensions |
Uses MXM connectors: 82mm (length) * 60mm (width) * 7mm (height) |
|
- Other Products
