Optimising Convolutions for Deep Learning Inference on ARM Cortex-M Processors
Por favor, use este identificador para citar o enlazar este ítem:
http://hdl.handle.net/10045/142602
Título: | Optimising Convolutions for Deep Learning Inference on ARM Cortex-M Processors |
---|---|
Autor/es: | Maciá, Antonio | Barrachina Mir, Sergio | Fabregat Llueca, Germán | Dolz, Manuel F. |
Centro, Departamento o Servicio: | Universidad de Alicante. Departamento de Tecnología Informática y Computación |
Palabras clave: | Deep learning | Convolution operator | Edge computing | Microcontrollers | High performance | ARM Cortex-M | CMSIS-NN |
Fecha de publicación: | 30-abr-2024 |
Editor: | IEEE |
Cita bibliográfica: | IEEE Internet of Things Journal. 2024. https://doi.org/10.1109/JIOT.2024.3395335 |
Resumen: | We perform a series of optimisations on the convo lution operator within the ARM CMSIS-NN library to improve the performance of deep learning tasks on Arduino development boards equipped with ARM Cortex-M4 and M7 microcontrollers. To this end, we develop custom microkernels that efficiently handle the internal computations required by the convolution operator via the lowering approach and the direct method, and we design two techniques to avoid register spilling. We also take advantage of all the RAM on the Arduino boards by reusing it as a scratchpad for the convolution filters. The integration of these techniques into CMSIS-NN, when invoked by TensorFlow Lite for microcontrollers for quantised versions of VGG, SqueezeNet, ResNet, and MobileNet-like convolutional neural networks enhances the overall inference speed by a factor ranging from 1.13× to 1.50×. |
Patrocinador/es: | This research was funded by project TED2021-129334B-I00 supported by MCIN/AEI/10.13039/501100011033 and by the “European Union NextGenerationEU/PRTR”. Manuel F. Dolz was also supported by the Plan Gen–T grant CIDEXG/2022/13 of the Generalitat Valenciana. Antonio Macia-Lillo is a PRE2021-099284 fellow supported by MCIN/AEI/10.13039/501100011033. |
URI: | http://hdl.handle.net/10045/142602 |
ISSN: | 2327-4662 |
DOI: | 10.1109/JIOT.2024.3395335 |
Idioma: | eng |
Tipo: | info:eu-repo/semantics/article |
Derechos: | This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License |
Revisión científica: | si |
Versión del editor: | https://doi.org/10.1109/JIOT.2024.3395335 |
Aparece en las colecciones: | Personal Investigador sin Adscripción a Grupo |
Archivos en este ítem:
Archivo | Descripción | Tamaño | Formato | |
---|---|---|---|---|
Macia-Lillo_etal_2024_IEEE-ITJ.pdf | 2,04 MB | Adobe PDF | Abrir Vista previa | |
Todos los documentos en RUA están protegidos por derechos de autor. Algunos derechos reservados.