[Numpy-svn] [numpy/numpy] e31729: MAINT: add NPY_CPU_HAVE_UNALIGNED_ACCESS macro

[GitHub]

  Branch: refs/heads/master
  Home:   https://github.com/numpy/numpy
  Commit: e31729bb9c7abd08b1815100abab6ca63f71b8bf
      https://github.com/numpy/numpy/commit/e31729bb9c7abd08b1815100abab6ca63f71b8bf
  Author: Julian Taylor <jtaylor.debian at googlemail.com>
  Date:   2013-10-16 (Wed, 16 Oct 2013)

  Changed paths:
    M numpy/core/include/numpy/npy_cpu.h
    M numpy/core/src/multiarray/lowlevel_strided_loops.c.src

  Log Message:
  -----------
  MAINT: add NPY_CPU_HAVE_UNALIGNED_ACCESS macro

currently only i386 and amd64, must be used carefully as the compiler
may still be allowed to assume alignment and try to vectorize based on
type size alignment.


  Commit: 0f7eea9af4593f4490ff95611066e43176b5f8c8
      https://github.com/numpy/numpy/commit/0f7eea9af4593f4490ff95611066e43176b5f8c8
  Author: Julian Taylor <jtaylor.debian at googlemail.com>
  Date:   2013-10-16 (Wed, 16 Oct 2013)

  Changed paths:
    M numpy/core/src/multiarray/common.c
    M numpy/core/src/multiarray/item_selection.c

  Log Message:
  -----------
  MAINT: silence warning by casting to right type


  Commit: 935017cbc3a336f66a110b53d2b595fffc8adbd7
      https://github.com/numpy/numpy/commit/935017cbc3a336f66a110b53d2b595fffc8adbd7
  Author: Julian Taylor <jtaylor.debian at googlemail.com>
  Date:   2013-10-16 (Wed, 16 Oct 2013)

  Changed paths:
    M numpy/core/setup.py
    M numpy/core/src/multiarray/common.h
    M numpy/core/src/multiarray/dtype_transfer.c
    M numpy/core/src/multiarray/mapping.c
    M numpy/core/src/umath/ufunc_type_resolution.c
    M numpy/core/tests/test_api.py

  Log Message:
  -----------
  MAINT: add npy_memchr for memchr like operations

useful for small sizes where the function call is expensive.


  Commit: 7d4ea165817fc613c79bb92ccb3844df94d1beed
      https://github.com/numpy/numpy/commit/7d4ea165817fc613c79bb92ccb3844df94d1beed
  Author: Julian Taylor <jtaylor.debian at googlemail.com>
  Date:   2013-10-16 (Wed, 16 Oct 2013)

  Changed paths:
    M numpy/core/setup_common.py
    M numpy/core/src/multiarray/common.h

  Log Message:
  -----------
  ENH: improve npy_memchr(p, 0) by using __builtin_ctz (tzcnt on x86)

Improves sparse mask performance by about a factor of three, the worst
case of no consecutive mask elements slows down by about 10%-15%.


  Commit: 3b3fa76d5051078d27b1a30b77b586e09c8b889d
      https://github.com/numpy/numpy/commit/3b3fa76d5051078d27b1a30b77b586e09c8b889d
  Author: Charles Harris <charlesr.harris at gmail.com>
  Date:   2013-10-16 (Wed, 16 Oct 2013)

  Changed paths:
    M numpy/core/include/numpy/npy_cpu.h
    M numpy/core/setup.py
    M numpy/core/setup_common.py
    M numpy/core/src/multiarray/common.c
    M numpy/core/src/multiarray/common.h
    M numpy/core/src/multiarray/dtype_transfer.c
    M numpy/core/src/multiarray/item_selection.c
    M numpy/core/src/multiarray/lowlevel_strided_loops.c.src
    M numpy/core/src/multiarray/mapping.c
    M numpy/core/src/umath/ufunc_type_resolution.c
    M numpy/core/tests/test_api.py

  Log Message:
  -----------
  Merge pull request #3931 from juliantaylor/memchr-move

move memchr like code to a function


Compare: https://github.com/numpy/numpy/compare/cf0869ea03e6...3b3fa76d5051