Static Compilation Analysis for Host-Accelerator Communication Optimization - Mines Paris Accéder directement au contenu
Communication Dans Un Congrès Lecture Notes in Computer Science Année : 2012

Static Compilation Analysis for Host-Accelerator Communication Optimization


We present an automatic, static program transformation that schedules and generates e cient memory transfers between a computer host and its hardware accelerator, addressing a well-known performance bottleneck. Our automatic approach uses two simple heuristics: to perform transfers to the accelerator as early as possible and to delay transfers back from the accelerator as late as possible. We implemented this transformation as a middle-end compilation pass in the pips/Par4All compiler. In the generated code, redundant communications due to data reuse between kernel executions are avoided. Instructions that initiate transfers are scheduled e ectively at compile-time. We present experimental results obtained with the Polybench 2.0, some Rodinia benchmarks, and with a real numerical simulation. We obtain an average speedup of 4 to 5 when compared to a naïve parallelization using a modern gpu with Par4All, hmpp, and pgi, and 3.5 when compared to an OpenMP version using a 12-core multiprocessor.
Fichier principal
Vignette du fichier
A-476.pdf (319.1 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-00743496 , version 1 (19-10-2012)


  • HAL Id : hal-00743496 , version 1


Mehdi Amini, Fabien Coelho, François Irigoin, Ronan Keryell. Static Compilation Analysis for Host-Accelerator Communication Optimization. LCPC'2011 : The 24th International Workshop on Languages and Compilers for Parallel Computing, Sep 2011, Fort Collins, Colorado, United States. pp. 237-251. ⟨hal-00743496⟩
178 Consultations
616 Téléchargements


Gmail Facebook X LinkedIn More