NVIDIA Jetson AGX Orin, ConnectX-7, HPC, Payload Profile
- Embedded Ampere GPU with 2048 CUDA cores & 64 Gen3 Tensor cores
- Embedded 12-core NVIDIA Cortex ARM64 CPU, 2.2GHz
- 64 GB LPDDR5 256-bit memory with up to 205 GB/s
- ConnectX-7, up to 100GbE Ethernet, PCIe Gen5
- Module power: configurable from 60W - 100W
The VPX3U-ORIN-CX7-HPC module provides the data processing capability needed for demanding C5ISR applications, providing a secure compute node that provides advanced AI and HPC processing capabilities, high data transfer rates, and the cyber security features required to ensure data is being protected. This autonomous SOSA aligned module includes an NVIDIA Jetson AGX Orin, and an NVIDIA ConnectX-7 SmartNIC.
The NVIDIA Orin’s embedded Ampere GPU provides the CUDA cores and Tensor cores for data processing, deep learning inference, machine vision, audio processing, and video encoding/decoding. The 2048 CUDA cores provide FP and Int processing, while the 64 Gen3 Tensor cores provide the underlying architecture required for an efficient inference engine that can achieve up to 275 TOPS (INT8) of deep learning inference computing.
The NVIDIA ConnectX-7 SmartNIC provides secure, high-speed network data transfer and a configurable PCIe Gen5 switch. ConnectX-7 is ideal to meet the high data transfer and security requirements for C5ISR tasks. The ConnectX-7 also provides support for RDMA over Converged Ethernet (RoCE) and support for NVIDIA GPUDirect, enabling the fastest method for transferring data across the network to the GPU.
Unlocking the best performance requires the best cooling capability. WOLF’s advanced cooling technology is designed to move heat using a low weight, high efficiency path from the hot chip’s die to the wedgelocks.
WOLF-14TZ 3U VPX Module
This product is currently undergoing development and/or verification by the WOLF engineering team.
Contact WOLF for the latest update on product availabilty or to discuss your MCOTS requirements.
Additional AGX Orin Features
- DisplayPort Multi-Stream Transport (MST) with support for two display streams
- 2x Deep Learning Accelerator (DLA) v2 engines for inference operations
- Vision Accelerator engine for 7-way VLIW Vision Processor v2
- Dedicated programmable audio processor
- 2x HEVC (H.265) and AVC (H.264) NVENC and NVDEC with up to 4K-UHD encode resolution
- 1000 BASE-T Ethernet
- CUDA® 11, OpenGL® 4.6, OpenGL ES 3.2, Vulkan™ 1.0
- Flash Storage: 64 GB eMMC 5.1 with support for ECC
- USB 3.2 (up to 480 Mb/s); optional front panel port
Connectivity / System Management
- Storage: NVMe 1TB
- PCI Express Gen5 Switch with x16/x32
- Backplane Ethernet with 10/40/100 GBASE-KR4 and 10 GBASE-KR data planes; supports GPUDirect RDMA
- Block-level hardware encryption and the use of dedicated encryption keys per user
- Switching is offloaded from the CPU and run on the ConnectX hardware with NVIDIA ASAP2 technology
- On-board IPMI controller for system management
- Windows, Linux or ARM root complex
Mechanical / Open Systems Architecture
- High level of ruggedization:
- Operating temperature: -40° to +85°C
- Vibration (sine wave): 10G peak, 5 - 2000Hz
- Shock: 40G peak
- Dimensions: 160mm x 100mm x 25.4mm
- Weight: TBD
- ANSI/VITA 48, 65 (VPX REDI, OpenVPX)
- SOSA Aligned Payload slot profile 14.6.11-0 (with P2 depopulated) or 14.6.13 0 (with P2B depopulated)
Open the product datasheet - includes block diagram, order codes, etc.