|
gptkbp:instanceOf
|
GPU server platform
|
|
gptkbp:announced
|
2022
|
|
gptkbp:architecture
|
gptkb:Nvidia_Hopper
|
|
gptkbp:formFactor
|
server board
|
|
gptkbp:hasEthernetPort
|
yes
|
|
https://www.w3.org/2000/01/rdf-schema#label
|
Nvidia HGX H100
|
|
gptkbp:manufacturer
|
gptkb:Nvidia
|
|
gptkbp:maximum_GPUs
|
8
|
|
gptkbp:memoryType
|
gptkb:HBM2e
|
|
gptkbp:successor
|
Nvidia HGX A100
|
|
gptkbp:supports_air_cooling
|
yes
|
|
gptkbp:supports_BF16_precision
|
yes
|
|
gptkbp:supports_confidential_computing
|
yes
|
|
gptkbp:supports_DPUs
|
yes
|
|
gptkbp:supports_FP16_precision
|
yes
|
|
gptkbp:supports_FP8_precision
|
yes
|
|
gptkbp:supports_InfiniBand
|
yes
|
|
gptkbp:supports_INT4_precision
|
yes
|
|
gptkbp:supports_INT8_precision
|
yes
|
|
gptkbp:supports_liquid_cooling
|
yes
|
|
gptkbp:supports_MIG
|
yes
|
|
gptkbp:supports_multi-GPU_interconnect
|
yes
|
|
gptkbp:supports_NCCL
|
yes
|
|
gptkbp:supports_NVIDIA_AI_Enterprise
|
yes
|
|
gptkbp:supports_NVIDIA_BlueField
|
yes
|
|
gptkbp:supports_NVIDIA_Clara
|
yes
|
|
gptkbp:supports_NVIDIA_ConnectX-7
|
yes
|
|
gptkbp:supports_NVIDIA_cuBLAS
|
yes
|
|
gptkbp:supports_NVIDIA_cuDNN
|
yes
|
|
gptkbp:supports_NVIDIA_cuGraph
|
yes
|
|
gptkbp:supports_NVIDIA_cuML
|
yes
|
|
gptkbp:supports_NVIDIA_Merlin
|
yes
|
|
gptkbp:supports_NVIDIA_Omniverse
|
yes
|
|
gptkbp:supports_NVIDIA_Quantum-2
|
yes
|
|
gptkbp:supports_NVIDIA_RAPIDS
|
yes
|
|
gptkbp:supports_NVIDIA_Spectrum-4
|
yes
|
|
gptkbp:supports_NVIDIA_Triton_Inference_Server
|
yes
|
|
gptkbp:supports_NVSwitch
|
yes
|
|
gptkbp:supports_PCIe_Gen4
|
yes
|
|
gptkbp:supports_PCIe_Gen5
|
yes
|
|
gptkbp:supports_secure_boot
|
yes
|
|
gptkbp:supports_TensorRT
|
yes
|
|
gptkbp:supportsCUDA
|
yes
|
|
gptkbp:supportsNVLink
|
yes
|
|
gptkbp:targetMarket
|
gptkb:artificial_intelligence
data centers
high-performance computing
|
|
gptkbp:used_in
|
supercomputers
cloud infrastructure
AI training clusters
|
|
gptkbp:uses_GPU
|
gptkb:Nvidia_H100
|
|
gptkbp:bfsParent
|
gptkb:Nvidia_H100
|
|
gptkbp:bfsLayer
|
7
|