CUDA Cores       4608

NVIDIA RT Cores              72

NVIDIA Tensor Cores     576

GPU Memory    48 GB GDDR6 with ECC

RTX-OPS              84T

Rays Cast             10 Giga Rays/Sec

Peak FP32 Performance                16.3 TFLOPS

Peak FP16 Performance                32.6 TFLOPS

Peak INT8 Performance                65.2 TOPS

Deep Learning TFLOPS   130.5 Tensor TFLOPS

Multi-GPU Scalability      NVLink

NVLink Bandwidth           100 GB/Sec

Memory Bandwidth       672 GB/s

System Interface             PCI Express 3.0 x16

Maximum Power Consumption 295 W

Energy Star Enabling       Yes

Thermal Solution              Ultra-quiet active fansink

Form Factor        111.8 mm H * 266.7 mm L, Dual Slot

Display Connectors         DisplayPort 1.4 (4) + VirtualLink (1)

DisplayPort with Audio  Yes

DVI-D Single-Link Connector       Yes, via included adapter

HDMI Support   Yes, via included adapter

Number of Displays Supported  4

Maximum DP 1.4 Resolution       HDR 7680 * 4320 at 60 Hz

5K Display Support          HDR 5120 * 2880 at 60 Hz

4K Display Support          HDR 4096 * 2160 or 3840 * 2160 at 120 Hz

Maximum DVI-D DL Resolution  2560 * 1600 at 60 Hz via 3rd party adapter

Maximum DVI-D SL Resolution  1920 * 1200 at 60 Hz via included adapter

HDCP Support   Yes

Professional 3D Support               Via optional Stereo Connector Bracket

Quadro Sync II Compatible          Yes (Frame Lock and Genlock)

NVIDIA GPU Direct Compatible Yes

Graph APIs         Shader Model 5.1, OpenGL 4.5, DirectX 12.0, Vulkan 1.0

Compute APIs   CUDA, DirectCompute, OpenCL

NVIEW  Yes

NVIDIA MOSAIC               Yes