search

kahypar / mt-kahypar

Mt-KaHyPar (Multi-Threaded Karlsruhe Hypergraph Partitioner) is a shared-memory multilevel graph and hypergraph partitioner equipped with parallel implementations of techniques used in the best sequential partitioning algorithms. Mt-KaHyPar can partition extremely large hypergraphs very fast and with high quality.
MIT License
122 stars 25 forks source link

performing V-cycle with a initial partition #179

Closed chenmagi closed 8 months ago

chenmagi commented 8 months ago

Hi,

I noticed the difference in functionality between KaHyPar and Mt-KaHyPar regarding the V-cycle interface in Python. Do you have a plan to enable that feature?

larsgottesbueren commented 8 months ago

Hi Magi,

You should be able to set the number of V-cycles in the context object (which passes options to the solver). If that doesn't work (please let us know), you can use the improvePartition function. It takes the number of V-cycles as an argument directly, i.e., you could first call the partitioner as you already do and then call improvePartition directly after with the desired number of cycles.

Hope this helps and if you have any further questions, please follow up.

Lars

chenmagi commented 8 months ago

Hi Lars,

Thank you for the comment. I found the example and used the API in my project. It works.