|
gptkbp:instanceOf
|
gptkb:personal_computer
|
|
gptkbp:announced
|
gptkb:NVIDIA
|
|
gptkbp:architecture
|
gptkb:Volta
gptkb:Hopper
gptkb:Ampere
|
|
gptkbp:designedFor
|
gptkb:artificial_intelligence
high-performance computing
|
|
gptkbp:enables
|
gptkb:simulation
data science
deep learning
scientific computing
data analytics
AI inference
large language model training
scalable AI training
|
|
gptkbp:firstReleased
|
2017
|
|
gptkbp:formFactor
|
server board
|
|
https://www.w3.org/2000/01/rdf-schema#label
|
NVIDIA HGX
|
|
gptkbp:manufacturer
|
gptkb:NVIDIA
|
|
gptkbp:supports
|
gptkb:NVIDIA_A100
gptkb:NVLink
gptkb:PCIe
gptkb:NVIDIA_GPUs
gptkb:NVIDIA_Grace_CPU
gptkb:NVIDIA_H100
gptkb:NVIDIA_NVSwitch
gptkb:InfiniBand_networking
gptkb:NVIDIA_V100
Ethernet networking
liquid cooling
high-speed interconnects
high-bandwidth memory
air cooling
multi-GPU configurations
up to 8 GPUs per board
|
|
gptkbp:targetMarket
|
gptkb:business
research institutions
cloud service providers
|
|
gptkbp:usedBy
|
gptkb:Google_Cloud_Platform
gptkb:Oracle_Cloud
gptkb:Amazon_Web_Services
gptkb:Microsoft_Azure
|
|
gptkbp:usedIn
|
gptkb:NVIDIA_DGX_systems
cloud computing
data centers
supercomputers
OEM server platforms
|
|
gptkbp:website
|
https://www.nvidia.com/en-us/data-center/hgx/
|
|
gptkbp:bfsParent
|
gptkb:NVIDIA_DGX
gptkb:DGX
|
|
gptkbp:bfsLayer
|
6
|