|
gptkbp:instance_of
|
gptkb:microprocessor
|
|
gptkbp:application
|
scientific computing
|
|
gptkbp:architecture
|
gptkb:ARM_architecture
RISC
|
|
gptkbp:availability
|
commercially available
|
|
gptkbp:cache_size
|
64 bytes
32 MB
L1, L2, L3 cache
16-way set associative
|
|
gptkbp:clock_speed
|
2.2 GHz
|
|
gptkbp:collaboration
|
gptkb:RIKEN_Center_for_Computational_Science
|
|
gptkbp:compatibility
|
Linux-based systems
Fujitsu chipsets
|
|
gptkbp:competes_with
|
x86 processors
|
|
gptkbp:compute_density
|
high compute density
|
|
gptkbp:design_philosophy
|
performance and efficiency
|
|
gptkbp:design_purpose
|
energy efficiency
|
|
gptkbp:designed_by
|
gptkb:Fujitsu
|
|
gptkbp:fuel_economy
|
optimized for power efficiency
|
|
gptkbp:graphics
|
No
|
|
gptkbp:has_units
|
up to 2.7 TFLOPS
|
|
https://www.w3.org/2000/01/rdf-schema#label
|
A64 FX processors
|
|
gptkbp:hyper_threading
|
No
|
|
gptkbp:innovation
|
SVE technology
|
|
gptkbp:instruction_set
|
gptkb:ARMv8.2-A
up to 4 instructions per cycle
|
|
gptkbp:is_a_platform_for
|
Fujitsu's own platform
|
|
gptkbp:is_integrated_with
|
integrated into supercomputers
|
|
gptkbp:launch_event
|
SC20
|
|
gptkbp:manufacturing_process
|
7 nm
|
|
gptkbp:market
|
gptkb:Japan
|
|
gptkbp:market_launch
|
gptkb:2020
|
|
gptkbp:market_segment
|
supercomputing
|
|
gptkbp:memory_type
|
DDR4
|
|
gptkbp:notable_feature
|
high memory bandwidth
|
|
gptkbp:number_of_cores
|
48
|
|
gptkbp:performance
|
gptkb:TOP500
low latency
High Performance Computing (HPC)
compared to Intel Xeon
up to 3.0 TOPS
|
|
gptkbp:power_consumption
|
high performance per watt
up to 240 W
|
|
gptkbp:processor
|
gptkb:microprocessor
gptkb:A64
|
|
gptkbp:production_company
|
gptkb:Fujitsu_Limited
|
|
gptkbp:ram
|
up to 512 GB/s
|
|
gptkbp:release_date
|
June 2020
|
|
gptkbp:release_year
|
gptkb:2020
|
|
gptkbp:research_focus
|
AI and machine learning applications
|
|
gptkbp:security_features
|
hardware security
|
|
gptkbp:sister_channel
|
8
|
|
gptkbp:slisupport
|
gptkb:Yes
|
|
gptkbp:socket_type
|
SP3
|
|
gptkbp:successor
|
A64 FX2
|
|
gptkbp:support
|
parallel processing
|
|
gptkbp:supports
|
64-bit computing
SVE (Scalable Vector Extension)
NUMA (Non-Uniform Memory Access)
|
|
gptkbp:target_audience
|
research institutions
|
|
gptkbp:thermal_design_power
|
TDP 240 W
|
|
gptkbp:used_in
|
gptkb:Fugaku_supercomputer
|
|
gptkbp:uses
|
HPC applications
|
|
gptkbp:virtualization_support
|
gptkb:Yes
|
|
gptkbp:bfsParent
|
gptkb:Fugaku_supercomputer
|
|
gptkbp:bfsLayer
|
8
|