Fast computational techniques for modeling RFX-mod fusion devices on hybrid CPU-GPU architectures

Domenico Abate, Bruno Carpentieri, Andrea G. Chiariello, Giuseppe Marchiori, Nicolo Marconato, Stefano Mastrostefano, Guglielmo Rubinacci, Salvatore Ventre, Fabio Villone

Research output: Contribution to conferencePaper

1 Citation (Scopus)

Abstract

We presents fast computational techniques applied to modelling the RFX-mod fusion device. We derive an integral equation model for estimating the current distribution in the conducting structures, and we compute the input-output transfer functions. By using hybrid CPU-GPU parallelization, speed-up factors of about 200 can be obtained against uniprocessor computation.
Original languageEnglish
DOIs
Publication statusPublished - 19 Sep 2016
Externally publishedYes
Event2016 URSI International Symposium on Electromagnetic Theory, EMTS 2016 - Espoo, Finland
Duration: 19 Sep 2016 → …

Conference

Conference2016 URSI International Symposium on Electromagnetic Theory, EMTS 2016
CountryFinland
CityEspoo
Period19/9/16 → …

    Fingerprint

All Science Journal Classification (ASJC) codes

  • Instrumentation
  • Radiation

Cite this

Abate, D., Carpentieri, B., Chiariello, A. G., Marchiori, G., Marconato, N., Mastrostefano, S., Rubinacci, G., Ventre, S., & Villone, F. (2016). Fast computational techniques for modeling RFX-mod fusion devices on hybrid CPU-GPU architectures. Paper presented at 2016 URSI International Symposium on Electromagnetic Theory, EMTS 2016, Espoo, Finland. https://doi.org/10.1109/URSI-EMTS.2016.7571350