The VPX3U-AD5000E-CX7-OPTICAL (WOLF-153W) combines an NVIDIA RTX 5000E (AD103) embedded GPU. Along with an NVIDIA ConnectX-7 SmartNIC to provide a compact 3U VPX solution for GPU-accelerated computing and high-speed optical networking.
The integrated GPU features 16 GB of ECC GDDR6 memory and supports CUDA Toolkit 12, CUDA-X AI, CUDA-X HPEC libraries. As well as, OpenCL 3.0, Vulkan 1.2 and DirectX 12 Ultimate, enabling developers to deploy GPU-accelerated applications. Such as across AI, computer vision, sensor processing and high-performance computing workloads.
Integrated ConnectX-7 networking provides PCIe Gen4 connectivity together with dual 40/100GBASE optical fabric interfaces. Therefore, supporting GPUDirect RDMA and RoCE for efficient, low-latency transfer of large datasets between systems.
The WOLF-153W operates within a configurable 100 W to 140 W power envelope. In addition, it is available in a rugged conduction-cooled design for operation in harsh environments. The module aligns with SOSA Payload Profile 14.6.11-0, making it suitable for integration into open systems architectures.
Key Features
- Rugged 3U VPX NVIDIA RTX 5000E embedded GPU module
- NVIDIA RTX 5000E (AD103) GPU
- NVIDIA ConnectX-7 SmartNIC
- Dual 40/100GBASE optical fabric interfaces (P2A & P2B)
- PCIe Gen4 backplane connectivity
- 16 GB ECC GDDR6 memory
- Up to 9,728 CUDA Cores
- 304 fourth-generation Tensor Cores
- 76 third-generation RT Cores
- CUDA Toolkit 12 and CUDA-X AI support
- GPUDirect RDMA and RoCE support
- PCIe configurable switch
- NVIDIA ASAP2 switching offload
- On-board IPMI system management
- Block-level hardware encryption with dedicated user keys
- Linux and Windows driver support
- Conduction-cooled rugged design
- SOSA Payload Profile 14.6.11-0 compliant
- Operating temperature from -40°C to +85°C
Applications
The WOLF-153W is designed for embedded platforms requiring GPU acceleration and high-bandwidth optical networking, including:
- High Performance Embedded Computing (HPEC)
- Artificial Intelligence (AI)
- GPU-accelerated computing
- Sensor processing
- High-speed data movement
- Embedded image and video processing
- Defence mission systems
- Aerospace embedded computing
- Open architecture SOSA platforms
Frequently Asked Questions
What GPU is used in the WOLF-153W?
The WOLF-153W integrates a single NVIDIA RTX 5000E (AD103) embedded GPU with 16 GB of ECC GDDR6 memory.
Does the WOLF-153W support high-speed optical networking?
Yes. The module incorporates an NVIDIA ConnectX-7 SmartNIC with dual 40/100GBASE optical fabric interfaces. Therefore, supporting GPUDirect RDMA and RoCE for low-latency data transfer.
Is the module compatible with SOSA architectures?
Yes. The WOLF-153W aligns with SOSA Payload Slot Profile 14.6.11-0 using dual optical fabric interfaces on P2A and P2B.
Why Choose the WOLF-153W?
Finally, the WOLF-153W combines GPU acceleration, high-speed optical networking. In addition, 3U VPX packaging into a single embedded computing module. Achieved by integrating an NVIDIA RTX 5000E GPU with ConnectX-7 SmartNIC technology. Therefore, providing a compact platform for applications. Especially those requiring GPU processing, low-latency networking and deployment within SOSA-aligned embedded systems.
Contact the Metromatics team to discuss the WOLF-153W 3U VPX NVIDIA RTX 5000E module.
