Changes between Version 14 and Version 15 of ScramblerPaper


Ignore:
Timestamp:
05/17/11 09:50:38 (14 years ago)
Author:
Brandon Shroyer
Comment:

—

Legend:

Unmodified
Added
Removed
Modified

  • ScramblerPaper

    v14 v15  
     1[[PageOutline]]
     2
    13 = Efficient Parallelization for AMR MHD Multiphysics Calculations; Implementaion in AstroBEAR =
    2 * Abstract
     4*Abstract
    35  * Current and future AMR simulations will require algorithms that are highly parallelized and manage memory efficiently.  In the redesign of AstroBEAR we have attempted to employ new techniques to achieve both of these goals.  Patch based AMR often employs ghost cells to decouple the hyperbolic advances of each grid.  This decoupling allows each patch to be advanced independently.  In AstroBEAR we utilize this independence to allow patches across multiple refinement levels to be advanced independently with preference going to the finer level patches.  This allows for global load balancing instead of level by level load balancing and allows for greater parallelization across both space and AMR level which can improve performance in deep simulations with many levels of refinement.  With regard to memory management we have employed a distributed tree algorithm so processors no longer need to store the entire tree structure.  We have also implemented a sweep method that pipelines the computations required for updating the fluid variables using unsplit algorithms.  This can dramatically reduce the memory overhead required for intermediate variables.
    46