Deep learning models have reached state of the art performance in many machine learning tasks. Benefits in terms of energy, bandwidth, latency, etc., can be obtained by evaluating these models directly within Internet of Things end nodes, rather than in the cloud. This calls for implementations of deep learning tasks that can run in resource limited environments with low energy footprints. Research and industry have recently investigated these aspects, coming up with specialized hardware accelerators for low power deep learning. One effective technique adopted in these devices consists in reducing the bit-width of calculations, exploiting the error resilience of deep learning. However, bit-widths are tipically set statically for a given model, regardless of input data. Unless models are retrained, this solution invariably sacrifices accuracy for energy efficiency. In this paper, we propose a new approach for implementing input-dependant dynamic bit-width reconfiguration in deep learning accelerators. Our method is based on a fully automatic characterization phase, and can be applied to popular models without retraining. Using the energy data from a real deep learning accelerator chip, we show that 50% energy reduction can be achieved with respect to a static bit-width selection, with less than 1% accuracy loss.

Dynamic bit-width reconfiguration for energy-efficient deep learning hardware / Jahier Pagliari, D.; Macii, E.; Poncino, M.. - ELETTRONICO. - (2018), pp. 1-6. (Intervento presentato al convegno 23rd IEEE/ACM International Symposium on Low Power Electronics and Design, ISLPED 2018 tenutosi a Hyatt Regency Bellevue, usa nel 2018) [10.1145/3218603.3218611].

Dynamic bit-width reconfiguration for energy-efficient deep learning hardware

Jahier Pagliari D.;Macii E.;Poncino M.
2018

Abstract

Deep learning models have reached state of the art performance in many machine learning tasks. Benefits in terms of energy, bandwidth, latency, etc., can be obtained by evaluating these models directly within Internet of Things end nodes, rather than in the cloud. This calls for implementations of deep learning tasks that can run in resource limited environments with low energy footprints. Research and industry have recently investigated these aspects, coming up with specialized hardware accelerators for low power deep learning. One effective technique adopted in these devices consists in reducing the bit-width of calculations, exploiting the error resilience of deep learning. However, bit-widths are tipically set statically for a given model, regardless of input data. Unless models are retrained, this solution invariably sacrifices accuracy for energy efficiency. In this paper, we propose a new approach for implementing input-dependant dynamic bit-width reconfiguration in deep learning accelerators. Our method is based on a fully automatic characterization phase, and can be applied to popular models without retraining. Using the energy data from a real deep learning accelerator chip, we show that 50% energy reduction can be achieved with respect to a static bit-width selection, with less than 1% accuracy loss.
2018
9781450357043
File in questo prodotto:
File Dimensione Formato  
main.pdf

non disponibili

Descrizione: Articolo principale
Tipologia: 2a Post-print versione editoriale / Version of Record
Licenza: Non Pubblico - Accesso privato/ristretto
Dimensione 784.95 kB
Formato Adobe PDF
784.95 kB Adobe PDF   Visualizza/Apri   Richiedi una copia
postprint.pdf

accesso aperto

Descrizione: Post Print
Tipologia: 2. Post-print / Author's Accepted Manuscript
Licenza: PUBBLICO - Tutti i diritti riservati
Dimensione 713.44 kB
Formato Adobe PDF
713.44 kB Adobe PDF Visualizza/Apri
Pubblicazioni consigliate

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11583/2785755