|
gptkbp:instanceOf
|
gptkb:architecture
|
|
gptkbp:advantage
|
energy efficiency
high throughput
|
|
gptkbp:application
|
gptkb:machine_learning
cryptography
scientific computing
audio processing
image processing
|
|
gptkbp:contrastsWith
|
gptkb:MIMD_architecture
MISD architecture
SISD architecture
|
|
gptkbp:definedIn
|
gptkb:Flynn's_taxonomy
|
|
gptkbp:enables
|
parallel processing
data-level parallelism
|
|
gptkbp:example
|
gptkb:ARM_NEON
gptkb:IBM_AltiVec
gptkb:Intel_MMX
gptkb:Intel_SSE
NVIDIA CUDA cores
|
|
gptkbp:feature
|
multiple processing elements
same instruction applied to multiple data
single control unit
|
|
gptkbp:firstDescribed
|
gptkb:Michael_J._Flynn
|
|
https://www.w3.org/2000/01/rdf-schema#label
|
SIMD architecture
|
|
gptkbp:introducedIn
|
1966
|
|
gptkbp:limitation
|
not suitable for divergent tasks
requires data parallelism
|
|
gptkbp:standsFor
|
gptkb:Single_Instruction,_Multiple_Data
|
|
gptkbp:usedIn
|
supercomputers
graphics processing units
multimedia extensions
vector processors
|
|
gptkbp:bfsParent
|
gptkb:SIMT_architecture
gptkb:GPGPU
|
|
gptkbp:bfsLayer
|
7
|