|
gptkbp:instanceOf
|
mainframe computer
|
|
gptkbp:announced
|
gptkb:Elon_Musk
|
|
gptkbp:application
|
gptkb:Tesla_Autopilot
gptkb:Tesla_Full_Self-Driving
|
|
gptkbp:architecture
|
custom D1 chip
|
|
gptkbp:chipset
|
gptkb:D1
|
|
gptkbp:competitor
|
Google TPU supercomputers
NVIDIA supercomputers
|
|
gptkbp:D1ChipBandwidth
|
4 TB/s
|
|
gptkbp:D1ChipCores
|
354
|
|
gptkbp:D1ChipInterconnect
|
on-chip mesh network
|
|
gptkbp:D1ChipManufacturer
|
gptkb:TSMC
|
|
gptkbp:D1ChipMemory
|
576MB SRAM
|
|
gptkbp:D1ChipPerformance
|
362 teraflops
|
|
gptkbp:D1ChipProcessNode
|
7nm
|
|
gptkbp:D1ChipTransistorCount
|
50 billion
|
|
gptkbp:developedBy
|
gptkb:Tesla,_Inc.
|
|
gptkbp:energyEfficiency
|
high
|
|
gptkbp:features
|
scalable architecture
high bandwidth interconnect
exascale computing potential
|
|
gptkbp:hasDemo
|
August 2021
|
|
https://www.w3.org/2000/01/rdf-schema#label
|
Tesla Dojo supercomputer
|
|
gptkbp:location
|
gptkb:Palo_Alto,_California
|
|
gptkbp:notableFor
|
vertical integration
custom silicon design
AI training at scale
|
|
gptkbp:purpose
|
autonomous driving
training artificial intelligence models
|
|
gptkbp:size
|
modular tile-based design
|
|
gptkbp:status
|
in development
|
|
gptkbp:unveiled
|
gptkb:Tesla_AI_Day_2021
|
|
gptkbp:uses
|
gptkb:machine_learning
neural network training
|
|
gptkbp:bfsParent
|
gptkb:Tesla_Dojo_chip
|
|
gptkbp:bfsLayer
|
8
|