Back to Search
Start Over
Efficient FeFET Crossbar Accelerator for Binary Neural Networks
- Source :
- 2020 IEEE 31st International Conference on Application-specific Systems, Architectures and Processors (ASAP), International Conference on Application-specific Systems, Architectures and Processors (ASAP), ASAP
- Publisher :
- IEEE
-
Abstract
- This paper presents a novel ferroelectric field-effect transistor (FeFET) in-memory computing architecture dedicated to accelerate Binary Neural Networks (BNNs). We present in-memory convolution, batch normalization and dense layer processing through a grid of small crossbars with reduced unit size, which enables multiple bit operation and value accumulation. Additionally, we explore the possible operations parallelization for maximized computational performance. Simulation results show that our new architecture achieves a computing performance up to 2.46 TOPS while achieving a high power efficiency reaching 111.8 TOPS/Watt and an area of 0.026 mm2 in 22nm FDSOI technology.
- Subjects :
- 010302 applied physics
Computer science
020208 electrical & electronic engineering
Transistor
Normalization (image processing)
02 engineering and technology
TOPS
Grid
01 natural sciences
Binary neural network
Computational science
law.invention
law
0103 physical sciences
0202 electrical engineering, electronic engineering, information engineering
Crossbar switch
Bitwise operation
Electrical efficiency
Subjects
Details
- Language :
- English
- ISBN :
- 978-1-72817-147-0
- ISBNs :
- 9781728171470
- Database :
- OpenAIRE
- Journal :
- 2020 IEEE 31st International Conference on Application-specific Systems, Architectures and Processors (ASAP), International Conference on Application-specific Systems, Architectures and Processors (ASAP), ASAP
- Accession number :
- edsair.doi.dedup.....621783f1248c632eac009ba224ad0af9
- Full Text :
- https://doi.org/10.1109/asap49362.2020.00027