High Computing Power AI Mainboard from Firefly

The Octa-Core AI mainboard from Firefly has a high processing capacity and is driven by the SOPHON AI processor BM1684X. This mainboard supports up to 16GB of RAM. INT8 computational power of up to 32TOPS.

Features

A New Generation Processor: This mainboard is powered by the new generation SOPHON AI processor BM1684X, which has octa-core ARM Cortex-A53, a frequency of up to 2.3GHz, and a 12nm lithography process. It has up to 32Tops (INT8) of computing power, and 16TFLOPS (FP16/BF16) of computing power,  It works with popular programming frameworks and can be used in a lot of cloud and edge applications for AI inference.

Powerful multi-channel video AI performance: Up to 32 channels of 1080P H.264/H.265 video decoding can be done on the mainboard. It can process and analyze over 16 channels of HD video at the same time, which makes it great for AI applications like face detection and license plate recognition on live video streams.

Complete software and hardware: With full software and hardware, it’s easy to do artificial intelligence inference for both cloud and edge applications. All of them speed up the development of edge applications like face recognition, video structuring, abnormal alarm, equipment inspection, situation prediction, etc.

The one-stop toolkit is convenient and efficient: The SOPHON SDK one-stop deep learning development toolkit includes a number of software tools.

Specifications

• SOC: SOPHON BM1684X
• CPU: High-performance octa-core ARM A53, 12nm lithography process, frequency up to 2.3GHz.
• TPU: Built-in tensor computing module TPU, computing power up to: 32TOPS (INT8), 16TFLOPS (FP16/BF16), 2TFLOPS (FP32)
  Support mainstream programming frameworks: TensorFlow / Caffe / PyTorch / Paddle / ONNX / MXNet / Tengine / DarkNet.
• VPU: 32-channel H.265/H.264 1080p@25fps video decoding 32-channel 1080P@25fps HD video processing(decoding + AI analysis)
  Support 12-channel H.265/H.264 1080p@25fps video encoding
• RAM: 6GB/12GB/16GB LPDDR4/LPDDR4X
• Storage: 32GB/64GB/128GB eMMC
  -128MB SPI Flash
  -1 x M.2 SATA3.0 SSD (2242)
  -1 x TF card slot
• Ethernet: 2 x 1000Mbps Ethernet
• Wireless Network: 2.4GHz/5GHz dual-band WiFi, 802.11a/b/g/n/ac, capable of 4G LTE expansion (Mini PCIe), 5G (M.2)
• Video Output: 1 x HDMI（1080p@30fps）
• Audio Output: 1 x HDMI audio output
• USB: 2 x USB3.0 (current limit: 1A)
  2 x USB2.0 (current limit: 500mA)
• Power: DC 12V/5A (DC5.5×2.5mm)
• Other Interfaces:  1 x RS232, 1 x RS485, 1 x Debug（3P-2.0mm）1 x FAN（12V，4P-1.25mm）
• OS: Linux
• Dimension: 149mm × 97mm × 37.45mm
• Weight: 180g
• Environment: Operating Temperature: -20℃～60℃
  -Storage Temperature: -20℃～70℃
  -Operating humidity: 10%～90% (non-condensing)

Applications

Several popular application programming frameworks such as open SDK, a one-stop AI development toolkit, and a full, user-friendly toolchain are all provided. These features enable quick deployment of algorithm applications and inexpensive algorithm migration. This mainboard can be used for intelligent security, intelligent transportation, intelligent parks, intelligent retail, intelligent finance, smart cities, industries, energy, and more.

Please follow and like us: