README 2.2 KB

12345678910111213141516171819202122232425262728293031323334353637383940414243
  1. NumPy is a general-purpose array-processing package designed to
  2. efficiently manipulate large multi-dimensional arrays of arbitrary
  3. records without sacrificing too much speed for small multi-dimensional
  4. arrays. NumPy is built on the Numeric code base and adds features
  5. introduced by numarray as well as an extended C-API and the ability to
  6. create arrays of arbitrary type which also makes NumPy suitable for
  7. interfacing with general-purpose data-base applications.
  8. There are also basic facilities for discrete fourier transform, basic
  9. linear algebra and random number generation.
  10. If you need to build numpy for debugging, set DEBUG=y. If you use software
  11. which is having problems with numpy's new relaxed strides checking, set
  12. NPY_RSC=0.
  13. It is highly recommended to install libraries implementing BLAS and LAPACK
  14. before installing numpy. You may choose between:
  15. a) BLAS and LAPACK (reference but unoptimized and thus slow)
  16. b) OpenBLAS (optimized, provides LAPACK too)
  17. c) ATLAS and LAPACK (optimized), good to read README.ATLAS
  18. All these are available on SlackBuilds.org.
  19. If you want to use the UMFPACK library instead of SuperLU to solve unsymmetric
  20. sparse linear systems, then run this Slackbuild with NO_UMFPACK set to "no"
  21. and then install scikit-umfpack on top of scipy. In this context, UMFPACK is an
  22. optional dependency for numpy. Nevertheless, note that presently scikit-umfpack
  23. is not available on SlackBuilds.org while its dependencies are.
  24. NOTE: If you use this SlackBuild, numpy will run with the python version
  25. provided by Slackware Linux, which is presently 2.7.xx. If you'd like to
  26. use python 3.x then you have to install numpy with the numpy3 SlackBuild.
  27. IMPORTANT: The version installed by this SlackBuild does NOT include the
  28. oldnumeric and numarray compatibility modules since starting with
  29. version 1.9.0 these modules got removed by the numpy developers.
  30. If you need these compatibility modules please consider the
  31. numpy-legacy SlackBuild.
  32. THUS: This SlackBuild conflicts with the numpy-legacy SlackBuild
  33. which installs versions < 1.9.0!
  34. This numpy3 SlackBuild creates bindings for python3 and can be installed
  35. without conflict alongside the standard numpy SlackBuild.