• Aucun résultat trouvé

Message Scheduling for Data Redistribution through High Performance Networks

N/A
N/A
Protected

Academic year: 2021

Partager "Message Scheduling for Data Redistribution through High Performance Networks"

Copied!
35
0
0

Texte intégral

Loading

Figure

Figure 1. two distant clusters
Figure 2. Communication matrix
Figure 4. initial graph, k = 3
Figure 7. heuristics
+7

Références

Documents relatifs

This chapter shows that a fine proportional conflict redistribution rule must be preferred for the combination in the belief function theory.. Keywords: Experts fusion, DST,

Data Redistribution using One-sided Transfers to In-memory HDF5 Files 5 congestion if two simulation processes were sending data to the same DSM pro- cess using the block

Unit´e de recherche INRIA Lorraine, Technopˆole de Nancy-Brabois, Campus scientifique, ` NANCY 615 rue du Jardin Botanique, BP 101, 54600 VILLERS LES Unit´e de recherche INRIA

Algorithm 4 returns in polynomial time an optimal s hedule σ for the following s heduling problem: minimizing the makespan on a star-platform with homogeneous ommuni a- tion links

However, if we assume that each processor initially holds more data than it needs to send during the whole execution of the algo- rithm (what we call a light redistribution), then

There maybe several reasons why Pj+1 would also send a data item to Pj : a Pj+1 is the extremity of a positive maximal slice Cj+1,k and it sends a data item due to Steps 4 through

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des

Such structures raise two problems for in situ middleware: (1) It can be difficult to express these structures with current data interfaces inherited from traditional I/O