Skip to content

Use more threading in mesh algorithms #4164

@chrisrichardson

Description

@chrisrichardson

Add more threading to mesh topology computations

  • dual graph construction
  • sub-entity computation

Generally speed up mesh creation on multi-core machines, especially relevant to GPU, where often running one GPU per MPI rank, with many CPU cores.

Metadata

Metadata

Labels

roadmapPart of the FEniCS Roadmap

Type

No type

Projects

Status

No status

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions