A merge tree is a topological summary of a real-valued function on a graph. Merge trees can be used to find stable features in the data, report the number of connected components above any threshold, or compute other topological descriptors. A local-global merge tree provides a way of distributing a merge tree among multiple processors so that queries can be performed with minimal communication. While this makes them efficient in massively parallel setting, the only known algorithm for computing a local-global merge tree involves global reduction. Motivated by applications in cosmological simulations, we consider a restricted version of the problem: we compute a local-global tree down to a threshold fixed by the user. We describe two algorithms for computing such a tree via only local exchanges between processors. We present a number of experiments that show the advantage of our method on different simulations.
机构:Univ Fed Rio de Janeiro, COPPE, PEE, Engn Grad Program, BR-21941914 Rio De Janeiro, Brazil
Peres, R. T.
Pedreira, C. E.
论文数: 0引用数: 0
h-index: 0
机构:
Univ Fed Rio de Janeiro, COPPE, PEE, Engn Grad Program, BR-21941914 Rio De Janeiro, BrazilUniv Fed Rio de Janeiro, COPPE, PEE, Engn Grad Program, BR-21941914 Rio De Janeiro, Brazil
机构:
Univ Paris Diderot, UMR 7219, Lab SPHERE, Equipe REHSEIS, F-75205 Paris 13, FranceUniv Paris Diderot, UMR 7219, Lab SPHERE, Equipe REHSEIS, F-75205 Paris 13, France