Releases: hguo/ftk
Releases · hguo/ftk
Release 0.0.9
- Integration with DeformingSpacetimeMesh
- Critical line tracking in 3D unstructured grid
- Particle tracer for 2D/3D regular grid vector fields
- Unit tests for SYCL-accelerated 2D critical point tracking
- Application-specific updates
- XGC: Handling of magnetic fields from equilibrium data
- MPAS-Ocean: CPU- and GPU-based Lagrangian particle tracer with vertical velocity support
Release 0.0.8
- Support for 3D periodical spacetime meshes (the same 2D spatial meshes with designated way of spacetime triangulation)
- Adding an option to enable/disable cell reordering in unstructured 3D grids
- Fixing unit tests when building with metis; updating catch2 to 2.13.10 to fix linux build
- Various performance improvement and bug fixes
Release 0.0.7.1
- Allowing partitioning of 2D unstructured mesh for critical point tracking in parallel.
- Adding options to resample unstructured grids in input streams
- Various performance improvements and bug fixes.
Release 0.0.6
- New ParaView filter
CriticalPointTracker
- The new filter is the replacement of the previous
CriticalPoint2DTracker
,CriticalPointTracker2DUnstructured
, andCriticalPoint3DTracker
- The new integrated filter will determine mesh type (e.g.,
vtkUnstructuredGrid
andvtkImageData
) and dimensionality (2 or 3) for tracking critical points - The new filter uses correct coordinates and extents for
vtkStructuredGrid
,vtkRectilinearGrid
, andvtkImageData
- The new filter is the replacement of the previous
- New ParaView filter
CriticalLineTracker
that tracks intersections of isosurfaces in two scalar functions; GPU accelerated critical line tracker - SYCL support (experimental)
- 2D critical point tracking in regular grids (experimental)
- New features for XGC
- XGC 3D blob core line tracking with CUDA acceleration
- New executable
ftk.xgc.poincare
that computes Poincare plot for XGC simulations; CUDA accelerated Poincare plot computation
- New features for MPAS-O
- I/O for MPAS-O netCDF outputs
- Eddy center tracking in MPAS-O outputs
- Improved I/O for netCDF, HDF5, ADIOS2, and VTK
- Documentation
- New example of in situ critical point tracking in a heat transfer simulation
- Various bug fixes and performance improvements
Release 0.0.5.1
- ParaView synthetic data sources: added new sources including
MovingExtremum3DSource
,MovingRamp3DSource
,MovingDualRamp3DSource
; added new options to inject noise toSpiralWoven2DSource
; fixed implementation ofDoubleGyre2DSource
- ParaView filters: added
Levelset3DTracker
; added output options ("traced" and "sliced") for all existing filters (CriticalPoint2DTracker
,CriticalPoint3DTracker
,Levelset2DTracker
, andLevelset3DTracker
) - Documentation
- A new page of different installation and compilation options
- A new page to document a number of options to produce synthetic data for tutorial and testing purposes; all examples are provided with command lines and ParaView state files (
.pvsm
) for reproducibility - A new example of tracking critical points in a 2D flow-past-cylinder dataset
- Various bug fixes and performance improvements
Release 0.0.5
- Easing compilation and installation with
spack install ftk
- Bug fixing and performance improvements
- Supporting data reading from ADIOS2
.bp
files - Unit testing with CUDA accelerated critical point tracking
Release 0.0.4
- Integrated
track_critical_points
,track_tdgl_vortex
,track_contours
into one single executableftk
- GPU acceleration for tracking TDGL vortices and contours
- bug fixing and performance improvements
Release 0.0.2.2 - Introduction of PyFTK
- Python bindings of FTK
- Numerous improvements and bug fix
Release 0.0.2
This release features several new capabilities:
- A ParaView plugin
vtkCriticalPoint2DTracker
that tracks minima, maxima, and saddles in 2D scalar fields - Distributed and parallel 2D/3D critical point tracking with MPI
- CUDA-accelerated 2D/3D critical point tracking
- Automated testing with Travis CI / Google Test