GPU Architecture

NVIDIA Blackwell

NVIDIA® CUDA® Cores

24,064

Tensor Cores

Fifth generation

Ray Tracing Cores

Fourth generation

AI TOPS

3511 AI TOPS (1,2)

Single-Precision Performance

110 TFLOPS (1)

RT Core Performance

333 TFLOPS (1)

GPU Memory

96 GB GDDR7 with ECC

Memory Interface

512-bit

Memory Bandwidth

1792 GB/s

System Interface

PCIe 5.0 x16

Display Connectors

4x DisplayPort 2.1b

Max Simultaneous Displays

>4x 4096 x 2160 @ 120 Hz
 >4x 5120 x 2880 @ 60 Hz
 >2x 7680 x 4320 @ 60 Hz

Video Engines

>4x NVENC (ninth generation)
>4x NVDEC (sixth generation)

MIG Instance Types

>Up to 4x 24 GB
 >Up to 2x 48 GB
 >Up to 1x 96 GB

Power Consumption

Total board power: 300 W

Power Connector

1x PCIe CEM5 16-pin

Thermal Solution

Active

Form Factor

4.4” x 10.5” L, dual slot, full height

Graphics APIs

DirectX 12, Shader Model 6.6, OpenGL 4.63, Vulkan 1.4 (3)

Compute APIs

CUDA 12.8, OpenCL 3.0, DirectCompute

(1) Peak rates based on GPU Boost Clock
(2) Effective FP4 TOPS with sparsity
(3) Product is based on a published Khronos specification and is expected to pass the Khronos conformation testing process when available. Current conformance status can be found at www.khronos.org/conformance