WOLF-134S
Overview
Slot-Saving Turing RTX5000 Switch Module with 10GigE
The VPX3U-RTX5000E-SWITCH module includes an NVIDIA Quadro Turing RTX5000 GPU for high performance embedded computing (HPEC), a PCIe Gen4 switch, and a 10Gig Ethernet switch, all on one slot saving module.
Adding an RTX5000 GPU to a switch module provides system integrators with a novel way to increase the processing capability of their system without using an additional slot. Data can be sent to/from the GPU via PCIe, providing access to the Turing GPU’s advanced parallel processing and AI inference processing capabilities.
The module includes a 10Gig Ethernet switch, providing fast data movement from the module over the system backplane. An optional fiber to PCIe interface is also available. The default profile for this module is the SOSA switch profile (SLT3-SWH-6F8U-14.4.15). OpenVPX profiles are also supported to provide flexible integration options.
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 pipeline from the GPU die to the wedgelocks.
Block Diagram
The block diagram shows a slot-saving 3U module combining Turing RTX5000, a PCIe Gen4 switch, and a 10GbE Ethernet switch. It highlights the dual-fabric concept: PCIe switching for GPU/peer connectivity and Ethernet switching for network distribution across the VPX backplane, enabling compute and switching within one slot.
Features
WOLF-134S is a rugged 3U VPX module that combines an NVIDIA Quadro Turing TU104 RTX5000 GPU with a PCIe Gen4 switch and a 10 Gigabit Ethernet switch on a single slot-saving switch module. It provides a path to add GPU processing capability without consuming an additional chassis slot, while also enabling fast data movement over the VPX backplane through the integrated Ethernet switch. The default alignment is the SOSA switch slot profile 14.4.15, with OpenVPX profile support available for flexible integration.
GPU Features
- Turing GPU parallel processing:
- 3072 CUDA cores
- CUDA Toolkit 10, CUDA Compute capability 7.5
- OpenCL 1.2, DirectX 12, OpenGL 4.6, OpenGL ES 3.2, Vulkan 1.0
- 384 Turing Cores for advanced AI inference processing
- 16 GB GDDR6 256-bit memory with up to 448 GB/s
- GDDR6 memory provides twice the bandwidth of the previous generation of GDDR5 memory
- HEVC (H.265) and AVC (H.264) Turing NVENC and NVDEC hardware acceleration with up to 8K encode resolution and B-frame support
Connectivity / System Management
- PCIe Gen4 Switch
- 10 Gig Ethernet Switch
- Optional Fiber to PCIe interface
- IPMI Controller for module management
- Windows and Linux drivers
Mechanical / Open Systems Architecture
- High level of ruggedization:
- Rugged Conduction- cooled
- Operating temperature: - 40°C to +85°C
- Vibration (sine wave): 10G peak, 5 - * Shock: 40G peak for conduction- cooled
- Dimensions: 160mm x 100mm x 25.4mm
- Weight: To be Announced
- +12V only power source (as per SOSA requirements)
- ANSI/VITA 48, 65 (VPX- REDI, OpenVPX)
- SOSA aligned with SOSA switch slot profile support
Specifications
Key configuration and performance parameters for WOLF-134S are summarized below for quick comparison. Values shown reflect the standard module configuration; contact WOLF if you need a tailored build or a specific payload profile.
Configuration Guide
The following table defines a series of common order codes for the VPX3U-RTX5000E-SWITCH module. The asterisks denote characters of the part number that are defined based on common configuration options.
Some common configuration options for this module include:
- Default power threshold
- Slot profile
- Variant locked
- Conformal coatings
3U VPX Turing RTX5000, PCIe Gen4, 10GigE, single slot configurations
Contact Sales for the latest Ordering Numbers and available options
Representative ordering numbers
| Ordering Number | Description |
|---|---|
| 134S33-F**0-000VPX3vB0 | 3U VPX, conduction cooled, 1", NVIDIA RTX5000, PCIe Gen4, 10GigE, no fiber interface |
| 134S33-F**1-000VPX3vB0 | 3U VPX, conduction cooled, 1", NVIDIA RTX5000, PCIe Gen4, 10GigE, includes optional fiber interface |
Related Products
- NVIDIA Quadro RTX 5000 (Turing TU104)
- 3072 CUDA Cores, 384 Tensor Cores
- 16 GB GDDR6 VRAM
- PCIe Gen 3
- SOSA Payload, SOSA Legacy
- Module power 50–165 W
- NVIDIA Quadro RTX 5000 (Turing TU104)
- 3072 CUDA Cores, 384 Tensor Cores
- 16 GB GDDR6 VRAM
- PCIe Gen 3
- SOSA Payload
- Module power 50–175 W
- NVIDIA Quadro RTX 5000 (Turing TU104)
- 3072 CUDA Cores, 384 Tensor Cores
- 16 GB GDDR6 VRAM
- PCIe Gen 4
- SOSA Payload
- Module power 60–190 W
FAQ
Choose WOLF-134S when you want to combine system switching and GPU compute in a single slot to save chassis space.
This is typically useful when:
- Slot budget is constrained
- You want a switch module that can also run GPU workloads
- You want to consolidate PCIe and Ethernet switching roles while adding compute
If you prefer strict separation of compute and switching for fault containment or upgrade cadence, discrete modules are often better.
Confirm these early to avoid backplane mismatch:
- Default alignment is the SOSA switch slot profile 14.4.15
- Confirm how 10GbE switching is mapped into the chassis Ethernet planes
- Confirm how the PCIe Gen4 switch is used (GPU feeding vs fan-out to other endpoints)
Switch modules integrate most smoothly when the chassis is explicitly designed for a switch slot role.
It matters when platform constraints require PCIe connectivity over fiber rather than copper (distance, EMI/EMC, or architectural partitioning).
If PCIe connectivity is entirely backplane-local and copper-based, the fiber option is typically not required.
Key ordering-time decisions typically include:
- Switch configuration: Ethernet port mapping and PCIe routing strategy
- Fiber interface included or not
- Power threshold and ruggedization controls
- Program controls: conformal coating and variant locking
Contact Sales for the latest ordering numbers and available configuration options.