|
gptkbp:instanceOf
|
gptkb:personal_computer
|
|
gptkbp:announced
|
2014
|
|
gptkbp:competitor
|
gptkb:AMD_Infinity_Fabric
gptkb:PCI_Express
|
|
gptkbp:developedBy
|
gptkb:NVIDIA
|
|
gptkbp:enables
|
multi-GPU scaling
CPU-to-GPU communication
GPU-to-GPU communication
shared memory access
|
|
https://www.w3.org/2000/01/rdf-schema#label
|
NVLink interconnect
|
|
gptkbp:introducedIn
|
gptkb:NVIDIA_Pascal_architecture
|
|
gptkbp:memoryBusWidth
|
up to 25 GB/s per lane (NVLink 2.0)
up to 50 GB/s per lane (NVLink 3.0)
|
|
gptkbp:successor
|
PCI Express (in some use cases)
|
|
gptkbp:supports
|
deep learning
high-performance computing
data analytics
AI workloads
coherent memory access
scalable GPU clusters
high-bandwidth communication
|
|
gptkbp:supportsProtocol
|
gptkb:NVLink_2.0
gptkb:NVLink_3.0
gptkb:NVLink_4.0
gptkb:NVLink_1.0
point-to-point interconnect
|
|
gptkbp:usedIn
|
gptkb:NVIDIA_A100
gptkb:NVIDIA_Tesla_V100
gptkb:NVIDIA_DGX-1
gptkb:NVIDIA_HGX_systems
gptkb:NVIDIA_Tesla_P100
gptkb:NVIDIA_GPUs
gptkb:NVIDIA_DGX_systems
gptkb:NVIDIA_Ampere_architecture
gptkb:NVIDIA_DGX-2
gptkb:NVIDIA_DGX_A100
gptkb:NVIDIA_DGX_H100
gptkb:NVIDIA_Grace_CPU
gptkb:NVIDIA_Grace_Hopper_Superchip
gptkb:NVIDIA_H100
gptkb:NVIDIA_Hopper_architecture
gptkb:NVIDIA_NVSwitch
gptkb:NVIDIA_Quadro_GV100
gptkb:NVIDIA_Turing_architecture
gptkb:NVIDIA_Volta_architecture
gptkb:NVIDIA_Titan_V
NVIDIA Mellanox NVSwitch
NVIDIA data center products
|
|
gptkbp:website
|
https://www.nvidia.com/en-us/data-center/nvlink/
|
|
gptkbp:bfsParent
|
gptkb:Nvidia_DGX_systems
|
|
gptkbp:bfsLayer
|
7
|