Message Clustering Technique towards Efficient Irregular Data Redistribution in Clusters and Grids

  title={Message Clustering Technique towards Efficient Irregular Data Redistribution in Clusters and Grids},
  author={Shih-Chang Chen and Tai-Lung Chen and Ching-Hsien Hsu},
Efficient scheduling algorithms are essential to irregular data redistribution in cluster grid Cluster grid is an environment with heterogeneous computing nodes and complex network It is important for schedulers to keep an eye on load balance and low communication cost while distributing different size of data segment on various processors High Performance Fortran Version 2 (HPF2) provides GEN_BLOCK distribution format which facilitates generalized block distributions In this paper, we present… 


Job scheduling and data replication on data grids
Message Scheduling for Irregular Data Redistribution in Parallelizing Compilers
An efficient algorithm is developed based on the list scheduling and its experimental results are compared with previous algorithms to obtain suboptimal solutions while satisfying the minimal communication costs condition and the minimal step condition.
An Efficient Communication Scheduling Method for the Processor Mapping Technique Applied Data Redistribution
A unified approach to constructing optimal communication schedules for the processor mapping technique applied Block-Cyclic redistribution is presented and can more efficiently construct the required communication schedules than other optimal scheduling methods.
Scheduling Block-Cyclic Array Redistribution
This article builds upon results of Walker and Otto, who solved a particular instance of the problem, and derives an optimal scheduling for the most general case, namely, moving from a CYCLIC(r) distribution on a P-processor grid to a CYclIC(s) Distribution on a Q-processor Grid, for arbitrary values of the redistribution parameters P, Q, r, and s.
Efficient Multidimensional Data Redistribution for Resizable Parallel Computations
This paper derives an algorithm for redistributing two-dimensional block-cyclic arrays from P to Q processors, organized as 2-D processor grids, and ensures a contention-free communication schedule for data redistribution if Pr ≤ Qr and Pc ≤ Qc.
ISO: Comprehensive Techniques Toward Efficient GEN_BLOCK Redistribution with Multidimensional Arrays
This paper presents comprehensive techniques to implement GEN_BLOCK redistribution in parallelizing compilers, including Indexing schemes for communication sets generation, a contention-free communication Scheduling algorithm and an Optimization technique for improving communication efficiency.
Essential Cycle Calculation Method for Irregular Array Redistribution
An Essential Cycle Calculation method for unequal block sizes array redistribution is presented and the experimental results show that the ECC method greatly outperforms the Sequence method for all test samples.
Efficient Algorithms for Block-Cyclic Redistribution of Arrays
This framework is based on a generalized circulant matrix formalism of the redistribution problem and a general purpose distributed memory model of the parallel machine, which allows for efficient and distributed algorithms for index set computation that do not require any interprocessor communication.
Scheduling contention-free irregular redistributions in parallelizing compilers
A two-phase degree-reduction (TPDR) method for scheduling HPF2 irregular array redistribution that not only avoids node contention but also shortens the overall communication cost.
Fast Runtime Block Cyclic Data Redistribution on Multiprocessors
The data redistribution algorithms implemented in the ScaLAPACK library in order to go from one block cyclic distribution on one grid to that on another grid are presented.