Graphics Processor | NVIDIA Quadro RTX 5000 GPU (TU104 Turing Architecture) or NVIDIA Quadro RTX 3000 GPU (TU106 Turing Architecture) Supporting DirectX 12, OpenGL 4.6 and Vulkan 1.0 |
Interface | 6U VPX Form Factor 16 Lane PCI Express 3.0, 2.0, 1.0 SOSA aligned profile: SLT6-PAY-4F2Q1H4U1T1S1S1TU2U2T1H-10.6.4 |
Graphics Memory | RTX 5000 GPU: 16 GB GDDR6 256-bit Memory Interface 448 GB/s Memory Bandwidth RTX 3000 GPU: 6 GB GDDR6 192-bit Memory Interface 336 GB/s Memory Bandwidth |
Video Outputs Front Panel |
per GPU: Two DisplayPort++ (4K UHD) (DisplayPort can be converted to DVI or VGA with adapters) |
GPGPU Capabilities | RTX 5000 GPU: 3072 CUDA Cores. 384 Tensor Cores. 48 RT Cores. Up to 9.49 TFLOPS FP32 Single Floating Point Performance RTX 3000 GPU: 1920 CUDA Cores. 30 RT Cores. 240 Tensor Cores. Up to 5.3 TFLOPS FP32 Single Floating Point Performance NVIDIA RTX Platofrm: Supports CUDA 10 (Compute Capability 7.5) and CUDA-X OpenCL 1.2 and Shader Model 5.1 H.265 (HEVC) / H.264 (MPEG4/AVC) Hardware Encode & Decode NVIDIA GPUDirect® RDMA, NVENC, NVDEC |
Power Consumption | Dual RTX5000 version: 240W – depending on the application Dual RTX3000 version: 180W – depending on the application |
Operating Temperature (MIL-STD-810) | -40°C to 70°C (Rugged Air Cooled) |
Vibration / Shock (MIL-STD-810) | 0.1 g²/Hz / 40 g |
Humidity (MIL-STD-810) | 95% Without Condensation |
Software & Platform Support | Windows or Linux on x86 VPX & PCIe |
Sarsen Technology Limited is registered in England & Wales. Registration number: 03902012.
Registered office: 23-24 High Street, Marlborough, Wiltshire, SN8 1LW