Fitting FFT onto an Energy Efficient Massively Parallel Architecture - Inria - Institut national de recherche en sciences et technologies du numérique Accéder directement au contenu
Communication Dans Un Congrès Année : 2010

Fitting FFT onto an Energy Efficient Massively Parallel Architecture

Mihaela Malita
  • Fonction : Auteur
  • PersonId : 872189
Razvan Andonie
  • Fonction : Auteur
  • PersonId : 872190

Résumé

We present novel implementations of the Fast Fourier Transform on the massively parallel Connex ArrayTM(CA) circuit. The Estimated performance is 19 GFlops (BenchFFT metric) of parallel computing 64 FFTs of size 1024, using 5 Watts. We compare the CA and NVIDIA's GTX 285 GPU performance. The CA is not a direct NVIDIA competitor, targeting a different application area. Considering its low power dissipation, the CA is a good solution for low cost mobile computing equipment, video processing, and multi-channel high-sampling audio processing.
Fichier non déposé

Dates et versions

inria-00492915 , version 1 (17-06-2010)

Identifiants

  • HAL Id : inria-00492915 , version 1

Citer

István Lorentz, Mihaela Malita, Razvan Andonie. Fitting FFT onto an Energy Efficient Massively Parallel Architecture. IFMT'10 - Second International Forum on Next Generation Multicore/Manycore Technologies, Jun 2010, Saint Malo, France. ⟨inria-00492915⟩

Collections

IFMT10 ISCA2010
21 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More