|
gptkbp:instance_of
|
gptkb:microprocessor
|
|
gptkbp:application
|
AI training
various research applications
|
|
gptkbp:architecture
|
gptkb:Wafer_Scale_Engine_2
|
|
gptkbp:availability
|
limited availability
|
|
gptkbp:awards
|
received industry awards
|
|
gptkbp:case_analysis
|
published case studies
|
|
gptkbp:collaboration
|
academic institutions
|
|
gptkbp:collaborative_projects
|
collaborative research projects
|
|
gptkbp:community
|
active user community
|
|
gptkbp:competitors
|
gptkb:NVIDIA
|
|
gptkbp:connects
|
high-speed fabric
|
|
gptkbp:cooling_system
|
liquid cooling
|
|
gptkbp:designed_for
|
large-scale AI models
|
|
gptkbp:development
|
ongoing improvements
|
|
gptkbp:dimensions
|
4 U rack size
|
|
gptkbp:economic_impact
|
significant industry impact
|
|
gptkbp:education
|
user training programs
|
|
gptkbp:feedback
|
user feedback incorporated
|
|
gptkbp:form_factor
|
rack-mounted
|
|
gptkbp:fuel_economy
|
high energy efficiency
|
|
gptkbp:funding
|
funded by grants
|
|
gptkbp:has_programs
|
Cerebras Software Stack
|
|
gptkbp:has_publications
|
numerous research papers published
|
|
gptkbp:has_research_center
|
designed for data centers
|
|
https://www.w3.org/2000/01/rdf-schema#label
|
Cerebras CS-2 system
|
|
gptkbp:input_output
|
data from various sources
trained AI models
|
|
gptkbp:is_scalable
|
high scalability
|
|
gptkbp:manufacturer
|
gptkb:Cerebras_Systems
|
|
gptkbp:market_segment
|
enterprise AI
|
|
gptkbp:network
|
gptkb:Ethernet
|
|
gptkbp:notable_feature
|
largest chip ever made
|
|
gptkbp:partnerships
|
technology companies
|
|
gptkbp:performance
|
gptkb:MLPerf
various performance metrics
up to 220 petaflops
|
|
gptkbp:power_consumption
|
20 kilowatts
|
|
gptkbp:price
|
not publicly disclosed
|
|
gptkbp:processor
|
gptkb:Cerebras_WSE-2
|
|
gptkbp:project
|
future developments planned
|
|
gptkbp:provides_information_on
|
large datasets
|
|
gptkbp:ram
|
2.4 terabytes
|
|
gptkbp:reach
|
international customers
|
|
gptkbp:released
|
gptkb:2021
|
|
gptkbp:research_focus
|
gptkb:neural_networks
|
|
gptkbp:successor
|
gptkb:Cerebras_CS-1
|
|
gptkbp:support
|
AI researchers
dedicated support team
|
|
gptkbp:supports
|
multiple frameworks
|
|
gptkbp:target_market
|
research institutions
|
|
gptkbp:training_speed
|
accelerated training speed
|
|
gptkbp:upgrades
|
regular upgrades
|
|
gptkbp:user_base
|
AI developers
|
|
gptkbp:user_interface
|
web-based interface
|
|
gptkbp:uses
|
deep learning
|
|
gptkbp:weight
|
over 1,000 pounds
|
|
gptkbp:bfsParent
|
gptkb:Cerebras_Cloud
gptkb:Cerebras_Software_Platform
|
|
gptkbp:bfsLayer
|
7
|