|
gptkbp:instanceOf
|
gptkb:graphics_card
|
|
gptkbp:announced
|
2019
|
|
gptkbp:application
|
gptkb:artificial_intelligence
gptkb:HPC
gptkb:Scientific_Computing
|
|
gptkbp:architecture
|
gptkb:Xe_HPC
|
|
gptkbp:codename
|
gptkb:Ponte_Vecchio
|
|
gptkbp:countryOfOrigin
|
gptkb:United_States
|
|
gptkbp:energyEfficiency
|
High
|
|
gptkbp:firstDelivery
|
2022
|
|
gptkbp:formFactor
|
gptkb:OAM
gptkb:PCIe
|
|
gptkbp:manufacturer
|
gptkb:Intel
|
|
gptkbp:market
|
gptkb:cloud_service
|
|
gptkbp:memoryType
|
gptkb:HBM2e
|
|
gptkbp:notableFeature
|
gptkb:EMIB
gptkb:Foveros_3D_stacking
Advanced packaging
Designed for exascale computing
High bandwidth memory
Multi-process tile design
|
|
gptkbp:processNode
|
gptkb:TSMC_N5
gptkb:Intel_7
|
|
gptkbp:productType
|
gptkb:Intel_Data_Center_GPU_Max_Series
|
|
gptkbp:RAM
|
128 GB
|
|
gptkbp:releaseYear
|
2022
|
|
gptkbp:successor
|
gptkb:Intel_Rialto_Bridge
|
|
gptkbp:supports
|
gptkb:oneAPI
gptkb:PCIe_5.0
gptkb:CUDA_(via_compatibility_layer)
gptkb:Xe_Link
gptkb:Xe_Matrix_Extensions_(XMX)
gptkb:OpenCL
gptkb:AVX-512
gptkb:DirectX_12
BF16
FP16
INT8
FP32
FP64
Tensor operations
Ray tracing
|
|
gptkbp:tileCount
|
47
|
|
gptkbp:type
|
Multi-tile
|
|
gptkbp:usedIn
|
gptkb:Aurora_supercomputer
gptkb:Intel_Data_Center_GPU_Max_1100
gptkb:Intel_Data_Center_GPU_Max_1350
gptkb:Intel_Data_Center_GPU_Max_1450
gptkb:Intel_Data_Center_GPU_Max_1550
|
|
gptkbp:bfsParent
|
gptkb:Intel_Xe
|
|
gptkbp:bfsLayer
|
6
|
|
https://www.w3.org/2000/01/rdf-schema#label
|
Intel Ponte Vecchio
|