
PNY NVIDIA Quadro GV100 Technical Specifications
CUDA Cores | 5120 |
Tensor Cores | 640 |
Peak Double Precision FP64 Performance | 7.9 TFLOPS |
Peak Single Precision FP32 Performance | 14.8 TFLOPS |
Peak Half Precision FP16 Performance | 29.6 TFLOPS |
Peak Integer Operation (INT8) Performance | 59.3 TFLOPS |
Deep Learning TFLOPS 1 | 118.5 TFLOPS |
Multi-GPU Scalability | NVLink (2-way) or SLI HB |
NVLink Bandwidth | 200 GB/s (bidirectional) |
GPU Memory | 32 GB HBM2 |
Memory Interface | 4096-bit |
Memory Bandwidth | 870 GB/s |
System Interface | PCI Express 3.0 x16 |
Maximum Power Consumption | 250 W Quadro Power Guidelines |
Energy Star Enabling | Yes |
Thermal Solution | Ultra-quiet active fansink |
Form Factor | 4.40” H x 10.50” L, Dual Slot |
Display Connectors | DP 1.4 (4) |
DisplayPort with Audio | Yes |
DVI-D Single-Link Connector | Yes, via four included adapters |
HDMI Support | Yes, via included adapter |
VGA Support | Via optional adapters |
Number of Displays Supported | 4 |
Maximum DP 1.4 Resolution | HDR 7680 x 4320 at 60Hz (30-bit color) Quadro and NVS Display Resolution Support |
5K Display Support | HDR 5120 x 2880 at 60Hz (30-bit color) |
4K Display Support | HDR 4096 x 2160 at 60Hz or 3840 x 2160 at 60Hz |
Maximum DVI-D DL Resolution | 2560 x 1600 at 60Hz via 3rd party adapter |
Maximum DVI-D SL Resolution | 1920 x 1200 at 60Hz via included adapter |
HDCP Support | Yes |
Quadro Sync II Compatible | Yes (Frame Lock and Genlock) |
NVIDIA GPU Direct Compatible | Yes |
Graphics APIs | Shader Model 5.1, OpenGL 4.5, DirectX 12.0, Vulkan 1.0 |
Compute APIs | CUDA, DirectCompute, OpenCL |
NVIEW | Yes |
NVIDIA MOSAIC | Yes (Windows 10, 8.1, 8, 7, and Linux) |
Warranty | 3 Years |
1. FP16 matrix multiply with FP16 and FP32 accumulate