Affine Vector Cache for memory bandwidth savings

Sylvain Collange 1, 2, * Alexandre Kouyoumdjian 1, 2
* Auteur correspondant
1 ARENAIRE - Computer arithmetic
Inria Grenoble - Rhône-Alpes, LIP - Laboratoire de l'Informatique du Parallélisme
Abstract : Preserving memory locality is a major issue in highly-multithreaded architectures such as GPUs. These architectures hide latency by maintaining a large number of threads in flight. As each thread needs to maintain a private working set, all threads collectively put tremendous pressure on on-chip memory arrays, at significant cost in area and power. We show that thread-private data in GPU-like implicit SIMD architectures can be compressed by a factor up to 16 by taking advantage of correlations between values held by different threads. We propose the Affine Vector Cache, a compressed cache design that complements the first level cache. Evaluation by simulation on the SDK and Rodinia benchmarks shows that a 32KB L1 cache assisted by a 16KB AVC presents a 59% larger usable capacity on average compared to a single 48KB L1 cache. It results in a global performance increase of 5.7% along with an energy reduction of 11% for a negligible hardware cost.
Type de document :
Liste complète des métadonnées

Littérature citée [28 références]  Voir  Masquer  Télécharger
Contributeur : Sylvain Collange <>
Soumis le : mercredi 7 décembre 2011 - 12:16:44
Dernière modification le : jeudi 7 février 2019 - 15:06:48
Document(s) archivé(s) le : vendredi 16 novembre 2012 - 14:40:37


Fichiers produits par l'(les) auteur(s)


  • HAL Id : ensl-00649200, version 1



Sylvain Collange, Alexandre Kouyoumdjian. Affine Vector Cache for memory bandwidth savings. 2011. 〈ensl-00649200〉



Consultations de la notice


Téléchargements de fichiers