2026-08-25T17:34:07Z https://drops.dagstuhl.de/oai

oai:drops-oai.dagstuhl.de:27141 2026-08-25T13:18:15Z ddc:004 open_access

Persistent Homology on GPU for 1d and 2d Cubical Filtrations Glisse, Marc Persistent homology cubical complex parallelism GPU This paper describes the algorithmic choices made in a pure GPU implementation to compute persistent homology for cubical complexes in dimension 1 or 2. The main ingredients are the parallel detection of apparent pairs, the simple reduction of the remaining complex when we only consider H₀, and the possibility to iterate those two steps. The result can be 200 times faster than a sequential CPU implementation. Schloss Dagstuhl – Leibniz-Zentrum für Informatik Marc Glisse 2026 Is Part Of LIPIcs, Volume 388, 34th Annual European Symposium on Algorithms (ESA 2026) InProceedings Text doc-type:ResearchArticle publishedVersion application/pdf doi:10.4230/LIPIcs.ESA.2026.5 urn:nbn:de:0030-drops-271416 https://drops.dagstuhl.de/entities/document/10.4230/LIPIcs.ESA.2026.5 eng https://creativecommons.org/licenses/by/4.0/legalcode