Skip to content

SCOREC/core

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

SCOREC Core

The SCOREC Core is a set of C/C libraries for unstructured mesh simulations on supercomputers.

For more information, start at our wiki page

What is in this repository?

  • PUMI: parallel unstructured mesh infrastructure API User's Guide
  • PCU: Communication and file IO built on MPI
  • APF: Abstract definition of meshes, fields, and related operations
  • GMI: Common interface for geometric modeling kernels
  • MDS: Compact but flexible array-based mesh data structure
  • PARMA: Scalable partitioning and load balancing procedures
  • SPR: Superconvergent Patch Recovery error estimator
  • MA: Anisotropic mixed mesh adaptation and solution transfer
  • SAM: Sizing anisotropic meshes
  • STK: Conversion from APF meshes to Sandia's STK meshes
  • ZOLTAN: Interface to run Sandia's Zoltan code on APF meshes
  • PHASTA: Tools and file formats related to the PHASTA fluid solver
  • MTH: Math containers and routines
  • CRV: Support for curved meshes with Bezier Shapes
  • PYCORE: Python Wrappers (see python_wrappers/README.md for build instructions)
  • REE: Residual based implicit error estimator

How do I get set up?

  • Dependencies: CMake for compiling and MPI for running
  • Configuration: Typical CMake configure and build. The example_config.sh shows common options to select, use a front-end like ccmake to see a full list of options
  • Tests: the test/ subdirectory has tests and standalone tools that can be compiled by explicitly listing them as targets to make.
  • Users: make install places libraries and headers in a specified prefix, application code can use these in their own compilation process. We also install pkg-config files for all libraries.

Contribution guidelines

  • Don't break the build
  • See the STYLE file
  • If in doubt, make a branch
  • Run the ctest suite
  • Don't try to force push to master or develop; it is disabled

Who do I talk to?

Citing PUMI

If you use these tools, please cite the following paper:

Daniel A. Ibanez, E. Seegyoung Seol, Cameron W. Smith, and Mark S. Shephard. 2016. PUMI: Parallel Unstructured Mesh Infrastructure. ACM Trans. Math. Softw. 42, 3, Article 17 (May 2016), 28 pages. DOI: https://doi.org/10.1145/2814935

We would be happy to provide feedback on journal submissions using PUMI prior to publication.