5739

Stack-less SIMT reconvergence at low cost

Sylvain Collange
ENS de Lyon, Universite de Lyon, LIP (UMR 5668 CNRS – ENS de Lyon – INRIA – UCBL), Ecole Normale Superieure de Lyon, 46 allee d’Italie, 69364 Lyon Cedex 07, France
hal-00622654, version 1, 2011

@article{collange2011stack,

   title={Stack-less SIMT reconvergence at low cost},

   author={Collange, S.},

   year={2011}

}

Download Download (PDF)   View View   Source Source   

761

views

Parallel architectures following the SIMT model such as GPUs benefit from application regularity by issuing concurrent threads running in lockstep on SIMD units. As threads take different paths across the control-flow graph, lockstep execution is partially lost, and must be regained whenever possible in order to maximize the occupancy of SIMD units. In this paper, we propose a technique to handle SIMT control divergence that operates in constant space and handles indirect jumps and recursion. We describe a possible implementation which leverage the existing memory divergence management unit, ensuring a low hardware cost. In terms of performance, this solution is at least as efficient as existing techniques.
No votes yet.
Please wait...

* * *

* * *

HGPU group © 2010-2017 hgpu.org

All rights belong to the respective authors

Contact us: