Skip to content

Latest commit

 

History

History
299 lines (208 loc) · 8.39 KB

CHANGES.rst

File metadata and controls

299 lines (208 loc) · 8.39 KB
Raw

3.0.0 (future)

Breaking Changes

  • Package will be renamed to corrfunc from Corrfunc

New features

  • conda installable package
  • GPU version

2.5.2 (04/10/2023)

Enhancements

  • Simplified and faster FALLBACK kernels [#303]

Infrastructure

  • Fix python_requires in setup.py [#302]
  • Automatic uploads of new releases to PyPI [#305]
  • Added @dependabot for automatic dependency updates in GitHub Actions [#306, #307, #308]

2.5.1 (28/07/2023)

Enhancements

  • Corrfunc now compiles and runs on Apple M1/M2 cpus (using the FALLBACK kernels) [#295]

Changes

  • Python >= 3.7 and numpy >= 1.16 are required for python extensions [#291]

Enhancements

  • Warn about loss of precision for float32 calculations involving small theta in DDtheta_mocks and large mu in DDsmu_mocks [#299]

2.5.0 (2022-12-23)

Enhancements

  • Allow user to specify periodicity and box size per dimension [#276]
  • Allow larger Rmax (up to half the boxsize) [#277]

Changes

  • Add Corrfunc/tests.py to CI testing [#260]
  • Migrate doctests to Python 3.8 [#261]
  • Migrate Python tests to pytest [#265]

Fixes

  • Add additional check to tell if it's safe to redirect stdout/err [#270]
  • Check and fix z vs cz in DDrppi_mocks and DDsmu_mocks only if comoving distance flag is not set [#275]
  • Update GNU assembler bug detection [#278]
  • Fix installation instructions and update README.rst [#285]
  • Remove mistaken references to "projected" correlation function in docs [#289]

2.4.0 (2021-09-30)

This release adds the boxsize parameter to the command line interfaces and requires the user to specify the box size in the Python interfaces to the periodic theory functions. It also contains a number of performance, code-quality, and user-experience improvements.

Breaking Changes

  • Require user to specify boxsize rather than automatically detect particle extent in periodic theory boxes. Applies to Python, command line, and C API. The order of some Python keyword args has also changed. [#199]

Enhancements

  • In the theoretical VPF calculation (theory.vpf), the total volume of the random spheres can now exceed the volume of the sample [#238]
  • Gridlink (the binning of particles into cells) now uses a parallel algorithm for the theory module [#239]
  • Add detection of known-bad Cray hugepages library at NERSC [#246]
  • Replace np.float with np.float64 to fix numpy 1.20 deprecation [#250]
  • Test Numpy versions as old as 1.14 and recent as 1.20 [#251]
  • Add lscpu and preprocessor defs to CI output [#259]

Bug fixes

  • Fix Python reference leak to results struct [#229]
  • Fix parsing error when periodic=False and boxsize not given in the theory module [#257]

2.3.4 (2019-07-21)

This is a bug-fix release and contains general code quality improvements.

Enhancements

  • A new helper routine to find the combination of (RA, DEC) refinements that produces fastest runtime in DDtheta_mocks [#216]
  • Further testing via GitHub Actions [#220]
  • Added Ubuntu-Xenial on Travis [#222]

Bug fixes

  • Fixing docs build failure on Travis [#215]
  • Fixing compile failure on missing 'CC' in environment [#226]

2.3.3 (2019-02-03)

This is a bug-fix release and contains general code quality improvements.

Enhancements

Bug fixes

  • Installation does not require python(3)-config anymore [#209, #211]
  • Better handling of terminal colours for unknown terminals [#209]
  • Prevent incorrect calculations with periodic boundaries for large ratios of (zmax, Rmax) to Lbox [#210]

2.3.2 (2019-12-24)

This is a release for bug-fixes and general code quality improvements. Travis now also tests for python3.7.

Enhancements

  • Improved code quality and adherence to PEP8 [#189]
  • Documentation no longer shows duplicate entries [#205]

Bug fixes

  • Incorrect calculations for non-native endian data [#191]
  • Large Rmax to Lbox ratio now supported for periodic boundaries [#192]
  • Workaround for GNU Assembler bug causing incorrect calculations [#196]
  • Only report runtime isa support if we also have compiler support [#200]
  • Example code to illustrate how to code custom weights with AVX512F [#205]

2.3.1 (2019-06-21)

Enhancements

  • Reduce memory footprint of the cell pairs [#186]

2.3.0 (2019-05-20)

Breaking Changes

New features

  • AVX512F kernels for all pair-counters, faster code from new optimizations using the minimum separation between pairs of cells, option to use the input particle arrays directly and not make a copy of the particle positions, internal code changes to (hopefully) achieve better OpenMP scaling [#167, #170, #173]

Bug fixes

  • Fix segmentation fault in vpf_mocks [#168]
  • Fix automatic uniform weights array when only one set of weights (or a scalar) is passed [#180]
  • Fix memory leak due to Python reference leak when using weights [#181]

2.2.0 (2018-08-18)

Breaking Changes

  • Drop Python 2.6 support

New features

  • Progress bar is displayed in Jupyter notebooks [#158]

Bug fixes

  • Fix virtualenv install issue [#159]
  • Error messages are displayed in Jupyter notebooks instead of the unhelpful "TypeError: 'NoneType' object is not iterable". [#158]

2.1.0 (2018-08-17)

New features

  • New pair counter DD(s, mu) for theory and mocks (contributed by @nickhand, in #130 and #132) [#166]

Enhancements

  • GSL version now specified and tested by Travis [#164]
  • Now possible to specify the number of Newton-Raphson steps to improve accuracy of approximate reciprocals. Available in DD(rp, pi) for mocks, and DD(s, mu) for both theory and mocks

2.0.0 (2017-04-06)

New features

  • Library behaviour can now be controlled at runtime
  • Calculates with doubles and floats transparently (passing arrays of doubles ensures calculation in double precision)
  • Both the API and ABI should be future proof
  • Extensive docs (first version with docs)
  • Arbitrary cosmology can be accounted for in the mocks routines #71

Breaking Changes

  • API has changed from previous version. Two additional inputs are now required for every statistic (#73)

Enhancements

  • Ctrl-C now aborts even within python extensions (cleans up memory too!, see issue #12)
  • Significantly improved installation for python
    • compiler can now be specified within python setup.py install CC=yourcompiler (see issue #31)
    • python via an alias is now solved #52

Bug fixes

  • Fixed bug in DDrppi_mocks where the minimum number of grid cells had to be 1 #70

Outstanding issues

  • Conda package still is pending (#49)
  • Recursive Makefile needs to be replaced with a more monolithic Makefile (#14)
  • Parameter parsing in python extensions can be flaky (#79)

1.1.0 (2016-06-08)

  • SSE kernels for all statistics
  • Incorrect normalization in xi. ALL previous xi calculations were wrong.

1.0.0 (2016-04-14)

  • Improved installation process
  • Detecting AVX capable CPU at compile time
  • Double-counting bug fixes in wp and xi

0.2.3 (2016-03-30)

  • Streamlined compilation on MACs
  • PyPI version is not verbose by default

0.2.2 (2016-02-09)

  • First version on PyPI

0.2.1 (2016-02-06)

  • AVX enabled by default

0.2.0 (2016-02-05)

  • Python 2/3 compatible

0.0.1 (2015-11-11)

  • Initial release