September 11, 2019
VMARC SoM Family
VMARC SoM is a series of SoM(System on Module) designed by Vamrs following the SMARC 2.0((Smart Mobility Architecture) standard. The SMARC standard is a versatile small form factor computer Module definition targeting applications that require low power, low costs, and high performance. VMARC SoM stand out with rich choice of computing power, graphics, camera, sound, network and optional wireless interfaces, offering embedded system developers a complete, off-the-shelf, upgradeable, credit-card sized embedded computing core that is ideal for IoT, multimedia, low-power graphics-intensive applications and AI enabled devices.
VMARC RK3399Pro SoM
VMARC RK3399Pro SoM ultilizes Rockchip’s latest AI processor with superior general-purpose computing performance. It equips ARM big.LITTLE architecture, dual-core Cortex-A72 + quad-core Cortex-A53, with technical leadership in overall performance and power consumption; quad-core ARM high-end GPU Mali-T860, integrates multiple bandwidth compression technology, providing overall excellent performance. Its on-chip NPU (Neural Network Processor) offers up to 3.0TOPs computing power.
| Component | Description |
|---|---|
| CPU | Dual-core Cortex-A72 up to 1.8GHz & Quad-core Cortex-A53 up to 1.4GHz |
| GPU | ARM® Mali-T860 MP4 Quad-core GPU |
| NPU | Support 8bit/16bit computing, AI computing power up to 3.0TOPs |
| VPU | Support 4K VP9 and 4K 10bits H265/H264 video decoding, up to 60fps |
| RAM | Optional configuration with the following options: - 3GB LPDRR3(CPU 2GB + NPU 1GB) - 6GB LPDDR3(CPU 4GB + NPU 2GB) - 8GB LPDDR3(CPU 4GB + NPU 4GB) |
| Storage | Optional configuration with the following options: 16GB/32GB/64GB/128GB high speed eMMC |
| Ethernet | Built-in Gigabit Ethernet PHY chip, 10/100/1000Mbps adaptive |
| USB | USB 2.0 HOST x2, USB 3.0 OTG x1 |
| Audio | RK809 Audio Codec, I2S 0/1, HDMI Audio, SPDIF |
| Display | Embed two Video Output Processor, support dual-screen simultaneous/extended display, and can choose to output from the following display interface: - MIPI-DSI×1 - eDP×1 - DP×1 - HDMI × 1 |
| Camera | MIPI CSI x2 |
| Additional | - ADC x3 - GPIO x12 - SDIO x2 - SPI x2 - PWM x3 - UART x3 - I2C x5 |
| Size | 82mm x 50mm |
Ficus2 Carrier Board
Ficus2 Carrier Board is an evaluation board for the VMARC SoM following 96boards Enterprise Edition Specicication. With Ficus2 Carrier Board, you can start prototyping your next AI enabled application immediately: connect a monitor via HDMI, keyboard, and mouse via USB, and you are good to go. The Ficus2 Kits include everything in the box to get started with VMARC SoM. It has the following feature:
| Component | Description |
|---|---|
| Power | - DC 12V/4A, with On/Off switch - ATX 4P 12V Power in |
| Button | Sleep/Resume button with status led |
| Ethernet | - One GbE with PoE support(additional adapter required) - One 10/100Mbit ethernet |
| Wireless | - 802.11 ac wifi, 2.4G&5G with on board antenna(optional uFL antenna) - Bluetooth 5.0 - Optional Mini PCIe 4G module with SIM card slot |
| USB | - USB 3.0 OTG x1 with hardware switch - USB 2.0 HOST x2 |
| PCIe | - One 4 lanes PCIe 2.1 connetor |
| Display | - eDP x1 - 8 lanes LVDS x1 - backlight voltage select - 4 lanes MIPI DSI x1 |
| Audio | - 3.5mm headphone jack - Speaker connector |
| Expansion | 96boards standard 40P Low Speed connector - UART x2 - SPI x1 - I2S x1 - I2C x 4 - GPIO x12 - PWM x1 96boards standard 60P High Speed connector - MIPI CSI x2(4 lanes + 2 lanes) - MIPI DSI x1(4 lanes) - USB HOST x1 - I2C x2 |
| Others | - User LEDs x4, WiFi LED, BT LED - IR receiver - RTC batery connector - 7.4V Li-on battery connector |
| Debug | - on board TTL to USB serial debug - jumper for NPU/CPU serial select |
| Size | 120mm x 160mm |
Software Support
Rockchip officially provides Fedora 17 and Android 8.1 dual boot image for RK3399Pro. A beta version of Debian Stretch is also available. In additional, to help developers access the NPU easily, Rockchip provides the following libraries and tools:
RKNN Toolkit: a python toolkit for AI
RKNN-Toolkit provides for users the development kit of model conversion, inference and performance evaluation based on PC, RK3399Pro, RK1808 hardware. Users can easily implement below features with the provided python interface:
- Model conversion
- Quantization function
- Model inference
- Performance evaluation
- Memory evaluation
- Model pre-compilation
- Model segmentation
- Custom OP
RKNN API: API to access the NPU
The RKNN API is an NPU(Neural Network Unit) acceleration interface based on Linux/Android. It provides a set of application programming interfaces (APIs) that based on NPU hardware acceleration, developers can use this API to develop AI related applications, the API will call the NPU hardware accelerator.
ROCK X SDK: The rapid AI application components
ROCK-X SDK is a set of components libraries for RK3399Pro/RK1808 platform. Developers can build rapid AI applications with ROCK-X SDK API. Currently the following features are provided:
| Classification | Functions |
|---|---|
| Object | Human detection, vehicle detection, object detection |
| Face | Face recognition, face critical features detection |
| Plate | Plate detection, Plate recognition |
| Pose | Body pose estimation, hand pose estimation |
VMARC RK3399Pro and Ficus2 Carrier Board is available at Vamrs Store.