Statements (29)
| Predicate | Object |
|---|---|
| gptkbp:instanceOf |
Neural Engine
|
| gptkbp:announced |
2019
|
| gptkbp:architecture |
custom AI accelerator
|
| gptkbp:competitor |
gptkb:NVIDIA_A100
gptkb:Google_TPU |
| gptkbp:countryOfOrigin |
gptkb:Israel
|
| gptkbp:designedFor |
deep learning training
|
| gptkbp:formFactor |
PCIe card
|
| https://www.w3.org/2000/01/rdf-schema#label |
Habana Gaudi
|
| gptkbp:manufacturer |
gptkb:Habana_Labs
|
| gptkbp:memoryType |
gptkb:HBM2
|
| gptkbp:numberOfChipsPerCard |
1
|
| gptkbp:openSource |
SynapseAI
|
| gptkbp:parentCompany |
gptkb:Intel
|
| gptkbp:PCIeGeneration |
Gen4
|
| gptkbp:platform |
gptkb:TensorFlow
gptkb:PyTorch |
| gptkbp:powerSource |
200W
|
| gptkbp:RAM |
32 GB
|
| gptkbp:successor |
Habana Gaudi2
|
| gptkbp:supports |
BF16
INT8 FP32 Ethernet-based interconnect |
| gptkbp:targetMarket |
cloud service providers
enterprise AI |
| gptkbp:usedIn |
data centers
|
| gptkbp:bfsParent |
gptkb:Intel_Nervana_NNP
|
| gptkbp:bfsLayer |
7
|