Nvidia H100

URI: https://gptkb.org/entity/Nvidia_H100

GPTKB entity

Predicate	Object
gptkbp:instanceOf	gptkb:graphics_card
gptkbp:architecture	gptkb:Hopper
gptkbp:coreCount	16896
gptkbp:formFactor	gptkb:PCIe gptkb:SXM
https://www.w3.org/2000/01/rdf-schema#label	Nvidia H100
gptkbp:manufacturer	gptkb:Nvidia
gptkbp:memoryType	gptkb:HBM3
gptkbp:NVLink	Supported
gptkbp:PCIe_Version	gptkb:PCIe_5.0
gptkbp:peakFP64Performance	400 TFLOPS 67 TFLOPS
gptkbp:peakINT8Performance	1979 TOPS
gptkbp:peakTF32Performance	989 TFLOPS
gptkbp:processNode	gptkb:TSMC_4N
gptkbp:RAM	80 GB 94 GB
gptkbp:releaseDate	2022
gptkbp:successor	gptkb:Nvidia_A100
gptkbp:supports	gptkb:NVIDIA_AI_Enterprise gptkb:Confidential_Computing gptkb:Multi-Instance_GPU_(MIG) gptkb:NVIDIA_Triton_Inference_Server gptkb:NVIDIA_RAPIDS gptkb:NVIDIA_TensorRT gptkb:NVIDIA_cuDNN gptkb:NVSwitch gptkb:PCIe_Gen5 gptkb:NVIDIA_DeepStream gptkb:NVIDIA_Magnum_IO gptkb:NVIDIA_NCCL gptkb:NVLink_Switch_System gptkb:Secure_Boot gptkb:NVIDIA_CUDA Virtualization FP16 INT8 FP32 FP64 ECC Memory TF32 Sparsity
gptkbp:targetMarket	gptkb:cloud_service AI Inference AI Training
gptkbp:TDP	350 W 700 W
gptkbp:Tensor_Cores	528
gptkbp:transistorCount	80 billion
gptkbp:usedIn	gptkb:Nvidia_DGX_H100 gptkb:Nvidia_HGX_H100
gptkbp:bfsParent	gptkb:Nvidia_Hopper
gptkbp:bfsLayer	6