Communication-Aware Supernode Shape
- 27 June 2008
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Parallel and Distributed Systems
- Vol. 20 (4) , 498-511
Abstract
In this paper we revisit the supernode-shape selection problem, that has been widely discussed in bibliography. In general, the selection of the supernode transformation greatly affects the parallel execution time of the transformed algorithm. Since the minimization of the overall parallel execution time via an appropriate supernode transformation is very difficult to accomplish, researchers have focused on scheduling-aware supernode transformations that maximize parallelism during the execution. In this paper we argue that the communication volume of the transformed algorithm is an important criterion, and its minimization should be given high priority. For this reason we define the metric of the per process communication volume and propose a method to minimize this metric by selecting a communication-aware supernode shape. Our approach is equivalent to defining a proper Cartesian process grid with MPI_Cart_Create, which means that it can be incorporated in applications in a straightforward manner. Our experimental results illustrate that by selecting the tile shape with the proposed method, the total parallel execution time is significantly reduced due to the minimization of the communication volume, despite the fact that a few more parallel execution steps are required.Keywords
This publication has 36 references indexed in Scilit:
- Message-passing code generation for non-rectangular tiling transformationsParallel Computing, 2006
- On the parallel execution time of tiled loopsIEEE Transactions on Parallel and Distributed Systems, 2003
- On time optimal supernode shapeIEEE Transactions on Parallel and Distributed Systems, 2002
- Impact of Tile-Size Selection for Skewed Tiling2001
- Generating efficient tiled code for distributed memory machinesParallel Computing, 2000
- Tiling Optimizations for 3D Scientific Computations2000
- Reducing data communication overhead for DOACROSS loop nests1994
- On time mapping of uniform dependence algorithms into lower dimensional processor arraysIEEE Transactions on Parallel and Distributed Systems, 1992
- LINEAR SCHEDULING IS NEARLY OPTIMALParallel Processing Letters, 1991
- A loop transformation theory and an algorithm to maximize parallelismIEEE Transactions on Parallel and Distributed Systems, 1991