Dual NVIDIA Tesla Pascal P6, 12.4 TFLOPS, HPC Mode and GRID vGPU
- Dual NVIDIA Tesla P6, 12.4 TFLOPS
- 4096 CUDA core GPGPU Engine
- High Performance Compute (HPC) mode
- NVIDIA GRID™ vGPU virtualization with up to 32 vGPU instances
- Operating power configurable hard cap: 80 – 180W
The VPX6U-TESLA-P6-DUAL-HPC board uses advanced NVIDIA® Tesla® PascalTM 16nm GPU technology. This rugged Pascal-based board runs in NVIDIA’s HPC mode which exposes a very large 32GB memory base address register (BAR) for direct access to the GPU memory from the CPU and other PCIe devices.
The Tesla P6 includes support for NVIDIA GRIDTM, the industry’s most advanced technology for sharing true virtual GPU (GRID vGPU™) hardware acceleration between multiple users. This technology ensures complete application compatibility, which means features and experience are the same as they would be on a physical device.
The rugged VPX6U-TESLA-P6-DUAL-HPC board includes air-cooled and conduction cooled options. For additional options contact WOLF to discuss MCOTS and custom design services.
- 32 GB GDDR5 memory with NVIDIA GPUDirect™ DMA technology
- Error Correcting Code (ECC) to protect against memory errors
- Pascal GPGPU parallel processing:
- 4096 CUDA® cores
- CUDA Toolkit 9, CUDA Compute version 6.1
- OpenCL™ 1.2, DirectX® 12, OpenGL 4.5, Vulkan 1.0
- Memory width: 256-bit to each GPU
- Maximum memory bandwidth: 192 GB/s to each GPU
- PCIe x16 Gen3
- NVENC/NVDEC accelerator for HEVC (H.265) and AVC (H.264) hardware encode/decode
- Windows and Linux drivers
- High level of ruggedization:
- Rugged air-cooled or conduction-cooled
- Operating temperature: -40° to +71°C
- Vibration (sine wave): 10G peak, 5 - 2000Hz
- Shock: 40G peak for conduction-cooled, 30G peak for air-cooled
- Front I/O and Rear I/O configurations
- Supported VPX configurations:
- VPX-REDI (ANSI/VITA 48.x)
- OpenVPX (ANSI/VITA 65)
Open the VPX6U-TESLA-P6-HPC datasheet: includes block diagram, order codes, etc.