summaryrefslogtreecommitdiff
path: root/numpy
Commit message (Collapse)AuthorAgeFilesLines
...
* | | | BUG: fix uninitialized-use on invalid sort kindJulian Taylor2013-06-191-3/+2
| | | | | | | | | | | | | | | | | | | | orign is unset at that point but used in the fail goto, can probably only be triggered by using the C-api wrong.
* | | | BUG: Fix failing python long behaviour and possible heisen bugSebastian Berg2013-06-173-28/+35
| | | | | | | | | | | | | | | | | | | | | | | | | | | | The priority error fixes np.float_(64) + python_long not working anymore. The second change reformats the PyArray_GetAttrString_SuppressException to avoid a possible heisenbug decrefing an object before use.
* | | | Merge pull request #3448 from efiring/ma_compressCharles Harris2013-06-162-2/+21
|\ \ \ \ | | | | | | | | | | BUG: np.ma.compress treated inputs in wrong order; closes #2495
| * | | | BUG: np.ma.compress treated inputs in wrong order; closes #2495Eric Firing2013-06-162-2/+21
| | | | |
* | | | | BUG: add __len__ method to ma.mvoid; closes #576Eric Firing2013-06-162-0/+7
| | | | |
* | | | | Merge pull request #3436 from juliantaylor/vectorize-boolCharles Harris2013-06-163-3/+242
|\ \ \ \ \ | | | | | | | | | | | | ENH: vectorize boolean logical &&, ||, abs and not
| * | | | | ENH: vectorize boolean logical &&, ||, abs and notJulian Taylor2013-06-143-3/+242
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The code ensures the result is identical to a boolean operation even though this might be unnecessary if bools are used correctly everywhere. The overhead doesn't matter much as vectorizing single byte operations hits the memory bandwidth limit very quickly. Improves performance by about a factor of 5 to 10 depending on the cpu. These operations currently can't be autovectorized by gcc 4.8.
* | | | | | Merge pull request #3449 from efiring/mvoid_setitemCharles Harris2013-06-162-3/+32
|\ \ \ \ \ \ | |_|_|/ / / |/| | | | | BUG: field assignment in masked array did not reset mask; closes #2403
| * | | | | BUG: field assignment in masked array did not reset mask; closes #2403Eric Firing2013-06-152-3/+32
| | |/ / / | |/| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The previous behavior when setting a field after indexing to select an element was suitable for the hard mask case, but not for the default soft mask. In addition, the _hardmask value was not being set at all in the mvoid instance. With this changeset, the _hardmask is passed in and __setitem__ takes it into account.
* | | | | MAINT: fix compilation issue gh-3443.jmozmoz2013-06-161-2/+2
| | | | |
* | | | | Merge pull request #3445 from juliantaylor/revert-x86-unaligned-accessCharles Harris2013-06-151-2/+7
|\ \ \ \ \ | |/ / / / |/| | | | BUG: Revert "ENH: enable unaligned loads on x86"
| * | | | BUG: Revert "ENH: enable unaligned loads on x86"Julian Taylor2013-06-161-2/+7
| |/ / / | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This reverts commit aef286debfd11a62f1c337dea55624cee7fd4d9e. x86 platform works with unaligned access but the compiler is allowed to assume all data is aligned to its size by the C standard. This means it can vectorize instructions peeling only by the size of the type, if the data is not aligned to this size one ends up with data not correctly aligned for SSE instructions (16 byte). So this flag can only be enabled if autovectorization is disabled.
* | | | Merge pull request #3440 from WarrenWeckesser/bug-ma-averageCharles Harris2013-06-152-68/+117
|\ \ \ \ | | | | | | | | | | BUG: ma: ma.average didn't handle complex arrays correctly (issue gh-2684)
| * | | | BUG: ma: ma.average didn't handle complex arrays correctly (issue gh-2684)Warren Weckesser2013-06-152-9/+52
| | | | |
| * | | | MAINT: ma: clean up ma/test_extras.py: don't use 'import *'; PEP8 whitespaceWarren Weckesser2013-06-151-61/+67
| |/ / /
* | | | Merge pull request #3432 from fhs/lstsq-docCharles Harris2013-06-141-3/+3
|\ \ \ \ | |/ / / |/| | | DOC: fix typo in linalg.lstsq doc string
| * | | DOC: another minor tweak of linalg.lstsq doc stringFazlul Shahriar2013-06-131-1/+2
| | | |
| * | | DOC: minor tweak of linalg.lstsq doc stringFazlul Shahriar2013-06-131-2/+1
| | | |
| * | | DOC: fix typo in linalg.lstsq doc stringFazlul Shahriar2013-06-131-2/+2
| |/ /
* | | ENH: random: Allow ngood=0 or nbad=0 in mtrand.hypergeometric.Warren Weckesser2013-06-133-252/+259
|/ / | | | | | | | | Also edited the 'Parameters' section of the docstring to comply with the numpy docstring standard.
* | Merge pull request #3410 from m-d-w/enh_np_array0.1njsmith2013-06-125-65/+275
|\ \ | | | | | | ENH: Optimize array creation by avoiding errors
| * | ENH: Optimize array creation by avoiding errorsm-d-w2013-06-104-64/+126
| | |
| * | TST: Add basic tests for np.arraym-d-w2013-06-061-1/+149
| |/
* | adjust the optimal IO buffer size for npz filesBartosz Telenczuk2013-06-123-15/+14
| |
* | Chunk reads in format.read_array.Alex Ford2013-06-121-3/+14
| | | | | | | | | | | | | | | | | | Maximum data size limitations in the crc32 module cause errors when reading more than 2 ** 32 bytes from gzip streams. Work around this issue when reading large arrays from npz files by chunking reads to 256mb. This appears to resolve bug #2922.
* | BUG: fix loading large npz files (fixes #2922)Bartosz Telenczuk2013-06-122-6/+25
| |
* | Merge pull request #3421 from juliantaylor/complex-copyCharles Harris2013-06-113-24/+9
|\ \ | | | | | | Poor performance complex strided copies since 1.7
| * | BUG: fix non-swapping strided copies in GetStridedCopySwapJulian Taylor2013-06-111-3/+3
| | | | | | | | | | | | | | | | | | | | | | | | | | | Fixed PyArray_GetStridedCopySwap{Pair,}Fn returns paths that do not swap the data. It never triggered because the alignment requirement for complex data was too low (4-8 bytes) and always fullfiled by the memory allocator. As this is now fixed it caused failures on sparc with 16 byte long doubles.
| * | ENH: enable unaligned loads on x86Julian Taylor2013-06-101-4/+2
| | | | | | | | | | | | | | | | | | | | | | | | x86 can handle unaligned load and there is no hand vectorized code in this file. It would be a serious compiler bug if it adds vectorization without checking for alignment. Enables fast complex128 copies which are unaligned on 32 bit gcc unless compiled with -malign-double.
| * | ENH: properly fix unaligned load of complex typesJulian Taylor2013-06-102-17/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | the workaround of setting the aligned to false unconditionally makes copying strided complex data extremely slow as it will always do unaligned elementwise memmoves. Instead set the alignment requirement for complex types correct in the dtype to begin with. Note that on 32 bit gcc complex double will still be aligned to 8 bytes unless compiled with -malign-double. It is possible this will introduce new segfaults on architectures without unaligned loads, but this now indicates a missing alignment check in the affected code as the array description is correct.
* | | Merge pull request #3419 from juliantaylor/vectorize-restCharles Harris2013-06-114-21/+369
|\ \ \ | | | | | | | | min/max and base math vectorization
| * | | MAINT: fix an uninitialized use in the no fenv fallbackJulian Taylor2013-06-111-1/+1
| | | |
| * | | MAINT: guard prototypes with HAVE_EMMINTRIN_HJulian Taylor2013-06-111-2/+2
| | | | | | | | | | | | | | | | avoids declared but not defined warnings
| * | | ENH: vectorize base math with SSE2Julian Taylor2013-06-093-0/+235
| | | | | | | | | | | | | | | | | | | | Improves performance by ~1.5/3.0 for float/double for inplace or cpu cached operations
| * | | ENH: Vectorize float min/max operation with sse2Julian Taylor2013-06-093-21/+134
| | | | | | | | | | | | | | | | Improves performance by ~1.5/3.0 for float/double.
* | | | Merge pull request #3422 from charris/deprecate-modulesCharles Harris2013-06-105-1/+30
|\ \ \ \ | | | | | | | | | | DEP: Deprecate the oldnumeric and numarray modules.
| * | | | DEP: Deprecate the oldnumeric and numarray modules.Charles Harris2013-06-105-1/+30
| | |/ / | |/| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The numarray and oldnumeric modules are deprecated. This is a bit tricky as raising a DeprecationWarning on import causes an error when tests are run. To deal with that, a ModuleDeprecationWarning class is added to numpy and NoseTester is modified to ignore that warning during testing. Closes #2905
* | | | Merge pull request #3417 from seberg/subspace-decrefCharles Harris2013-06-102-20/+93
|\ \ \ \ | |/ / / |/| | | BUG: Decref subspace when subspace is 0-d in fancy indexing
| * | | TST: Refactor some test_indexing doc stringsSebastian Berg2013-06-101-11/+40
| | | |
| * | | BUG: Decref subspace when subspace is 0-d in fancy indexingSebastian Berg2013-06-092-12/+56
| |/ / | | | | | | | | | | | | | | | | | | | | | | | | | | | Not decrefing it causes a major memory leak. This adapts the automated tests to cover this case as well and adds refcounting to it. Replaces some 'skip' with 0s, but I believe the skip tests are not that important. Added 1-d special case tests using the same indices as the multi-dim test though. Thanks to Julian Taylor for noting the bug.
* | | MAINT: use PyOS_snprintf instead of snprintfChristoph Gohlke2013-06-091-1/+1
|/ / | | | | PyOS_snprintf is portable and more secure than snprintf.
* | Merge pull request #3243 from seberg/deprecate-non-integer-arguments-newCharles Harris2013-06-0912-414/+387
|\ \ | | | | | | Deprecate non integer arguments
| * | MAINT: Disallow np.bool_ being used as integerSebastian Berg2013-06-092-1/+14
| | | | | | | | | | | | | | | | | | This makes the deprecation of bools used as integers complete from the numpy perspective. It would not be unsensible to move the bool checks into indexing specific code though.
| * | MAINT: Minor style fix and integer conversion optimizationsSebastian Berg2013-06-091-12/+12
| | |
| * | STY: Style fixes for integer deprecation changesSebastian Berg2013-06-094-14/+15
| | | | | | | | | | | | Also minor changes in the documentation.
| * | MAINT: Use warnings context manager instead of doing it by handSebastian Berg2013-06-091-12/+11
| | |
| * | DOC: Add integer deprecations to release notesSebastian Berg2013-06-091-2/+2
| | | | | | | | | | | | (also minor spelling fix in comments)
| * | MAINT: Create new function PyArray_IntpFromIndexSequenceSebastian Berg2013-05-312-7/+26
| | | | | | | | | | | | | | | This function uses npy_intp to in principle support larger then int sequence lengths.
| * | STY: style and small doc fixes for integer deprecationsSebastian Berg2013-05-312-4/+7
| | |
| * | TST: Rework test_deprecations.py and add new tests.Sebastian Berg2013-05-312-108/+192
| | |
View Lorry Controller Status