Generally, GPU clusters can be hosted on the cloud as well as at the edge. Even though cloud clusters can provide very high processing power, sometimes they have drawbacks such as high latency, low data security, and low system reliability. So, the most ideal solution is to have an edge GPU cluster to tackle these issues. However, they are very expensive and complex to get started. So, we are excited to bring you Jetson Mate Cluster Advanced which can provide a powerful edge GPU cluster solution at a much more affordable price, while being very easy to get started with the provided guides. Also, this solution comes in a much more compact form factor when compared to other GPU clusters in the market.
Jetson Mate Cluster Advanced is an all-in-one solution for your GPU cluster and server applications. It includes a Jetson Mate carrier board, a case, a cooling fan, and most of all, four Jetson Xavier™ NX SoMs with 6 TFLOPS AI performance each.
Highlighted Features
- Utilizes the power of four Jetson Xavier NX with 6 TFLOPS each
- Pre-installed SoMs on the Jetson Mate carrier board make it very convenient to get started
- All-in-one solution with a case, cooling fan, and SoMs in a compact form factor
- Easy to build a GPU cluster with easy-to-follow guides provided
- Onboard 5-port Gigabit switch for the SoMs to communicate with each other
- One master node with three worker nodes with the ability to power on/off the worker nodes separately
Applications
- GPU Server
- Home Lab
- GPU Cluster
Specifications
NVIDIA Jetson Xavier NX Module
| Specification | Details |
|---|---|
| AI Performance | 21 TOPS (INT8) |
| GPU | NVIDIA Volta architecture with 384 NVIDIA CUDA® cores and 48 Tensor cores |
| Max GPU Frequency | 1100 MHz |
| CPU | 6-core NVIDIA Carmel ARM®v8.2 64-bit CPU 6 MB L2 + 4 MB L3 |
| Max CPU Frequency | 1900 MHz |
| Memory | 8 GB 128-bit LPDDR4x @ 51.2GB/s |
| Storage | 16 GB eMMC 5.1 |
| Power | 10 W | 15 W |
| PCIe | 1 x1 (PCIe Gen3) + 1 x4 (PCIe Gen4), total 144 GT/s * |
| CSI Camera | Up to 6 cameras (24 via virtual channels) |
| 14 lanes (3x4 or 6x2) MIPI CSI-2 | |
| D-PHY 1.2 (up to 30 Gbps) | |
| Video Encode | 2x 4K @ 30 | 6x 1080p @ 60 | 14x 1080p @ 30 (H.265/H.264) |
| Video Decode | 2x 4K @ 60 | 4x 4K @ 30 | 12x 1080p @ 60 | 32x 1080p @ 30 (H.265) 2x 4K @ 30 | 6x 1080p @ 60 | 16x 1080p @ 30 (H.264) |
| Display | 2 multi-mode DP 1.4/eDP 1.4/HDMI 2.0 |
| DL Accelerator | 2x NVDLA Engines |
| DLA Max Frequency | 1100 MHz |
| Vision Accelerator | 7-Way VLIW Vision Processor |
| Networking | 10/100/1000 Base-T Ethernet |
| USB | 1xUSB 3.1 (10 Gbps) | 3xUSB 2.0 |
| Other IOs | 1x SDIO / 2x SPI / 3x UART / 2x I2S / 4x I2C / 1x CAN / GPIOs |
| Dimensions | 45 mm x 69.6 mm |
| Connector | 60-pin SO-DIMM connector |
Hardware Overview
Figure 1: Device Schematics
Figure 2: USB Mounting Guide