summaryrefslogtreecommitdiff
path: root/numpy
Commit message (Collapse)AuthorAgeFilesLines
* Merge pull request #18074 from tylerjereddy/treddy_exprtype_regex_simplifyCharles Harris2020-12-261-3/+3
|\ | | | | MAINT: exprtype regex simplify
| * MAINT: exprtype regex simplifyTyler Reddy2020-12-261-3/+3
| | | | | | | | | | | | | | | | | | | | | | | | | | * `determineexprtype_re_1` was modified to remove extraneous character class markers around a single `,` * a similar change was applied for the word metacharacter in `determineexprtype_re_2` and `determineexprtype_re_3` * the third character class in `determineexprtype_re_3` was simplified to remove an escape sequence--by placing the `-` at the start of the character class its metacharacter status can be avoided
* | Merge pull request #18072 from tylerjereddy/treddy_crackfortran_regex_simplify_1Charles Harris2020-12-261-2/+2
|\ \ | |/ |/| MAINT: crackfortran regex simplify
| * MAINT: PR 18072 revisionsTyler Reddy2020-12-261-1/+1
| | | | | | | | | | | | * `name_match` regular expression now starts by matching a letter only, based on reviewer feedback
| * MAINT: crackfortran regex simplifyTyler Reddy2020-12-251-2/+2
| | | | | | | | | | | | | | | | | | | | | | * remove extraneous character class markers used in `crackline_re_1`: `\w` and `=` on their own have no benefit to character class `[]` inclusion * `name_match` has a character class that can be simplified because `\w` metacharacter already encompasses the digit metacharacter and the underscore
* | Merge pull request #17587 from seiko2plus/to_npyv_sincos_f32Matti Picus2020-12-2610-290/+268
|\ \ | | | | | | SIMD: Replace raw SIMD of sin/cos with NPYV(universal intrinsics)
| * | MAINT: Suppress maybe-uninitialized warning in gcc on VSXSayed Adel2020-12-261-0/+10
| | |
| * | SIMD: Replace raw SIMD of sin/cos with NPYVSayed Adel2020-12-269-290/+258
| |/ | | | | | | | | | | The new code improves the performance of non-contiguous memory access for the output array without any reduction in performance. For PPC64LE the performance increased by 2-3.0, and 1.5-2.0 on aarch64.
* | Merge pull request #18066 from BvB93/typing-testsMatti Picus2020-12-269-437/+511
|\ \ | |/ |/| TST: Allow mypy output types to be specified via aliases
| * TST: Use type-aliases for describing `np.number` subclassesBas van Beek2020-12-238-434/+432
| |
| * TST: Allow mypy output types to be specified via aliasesBas van Beek2020-12-231-3/+79
| |
* | MAINT: Remove obsolete workaround to set ndarray.__hash__ = NoneRonan Lamy2020-12-242-11/+0
| | | | | | | | | | | | In Python 3, a type that sets tp_richcompare but leaves tp_hash null is always unhashable. In Python 2, such types used to inherit tp_hash from their bases.
* | Merge pull request #18001 from seiko2plus/fix_detect_aarch64_osxCharles Harris2020-12-233-87/+85
|\ \ | | | | | | BLD, BUG: Fix detecting aarch64 on macOS
| * | BLD, MAINT: add platform info to the final optimization reportSayed Adel2020-12-201-10/+12
| | |
| * | BLD, BUG: Fix detecting aarch64 on macOSSayed Adel2020-12-203-77/+73
| | | | | | | | | | | | Co-authored-by: h-vetinari <h.vetinari@gmx.com>
* | | Merge pull request #18052 from seberg/concat-with-string-dtypeCharles Harris2020-12-234-40/+116
|\ \ \ | | | | | | | | BUG: Fix concatenation when the output is "S" or "U"
| * | | BUG: Fix concatenation when the output is "S" or "U"Sebastian Berg2020-12-214-40/+116
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Previously, the dtype was used, this now assumes that we want to cast to a string of (unknown) length. This is a simplified version of what happens in `np.array()` or `arr.astype()` (it does never inspect the values, e.g. for object casts). This is more complex as I would like, and with the refactor of ResultType and similar can be cleaned up a bit more hopefully. Note that currently, object to "S" or "U" casts simply return length 64 strings, but with the new version, this will be an error (although the error message probably needs improvement). This is a behaviour inherited from other places however.
* | | | Merge pull request #18012 from Qiyu8/einsum-sumMatti Picus2020-12-231-429/+81
|\ \ \ \ | | | | | | | | | | SIMD: Optimize the performance of einsum's submodule sum.
| * | | | simplify the index related code.Qiyu82020-12-231-3/+3
| | | | |
| * | | | passing pointer not the address.Qiyu82020-12-231-14/+14
| | | | |
| * | | | add missing opt 3 flag.Qiyu82020-12-181-1/+1
| | | | |
| * | | | add NPY_GCC_OPT_3 option.Qiyu82020-12-171-3/+3
| | | | |
| * | | | Optimize the performance of einsum's submodule sum.Qiyu82020-12-171-426/+78
| | | | |
* | | | | Add docstrings to indicate which intrinsics are tested.Qiyu82020-12-231-0/+11
| | | | |
* | | | | Merge branch 'master' of github.com:numpy/numpy into countnzQiyu82020-12-2340-401/+1390
|\ \ \ \ \
| * \ \ \ \ Merge pull request #17790 from seiko2plus/npyv_new_intrinsic_pk1Matti Picus2020-12-2221-5/+598
| |\ \ \ \ \ | | |_|_|_|/ | |/| | | | ENH, SIMD: Add new NPYV intrinsics pack(1)
| | * | | | ENH, SIMD: Add new NPYV intrinsics pack(1)Sayed Adel2020-12-2221-5/+598
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | - add bitwise logical operations for boolean vectors - add round conversion for float vectors - add NAN test for float vectors - add conditional addition and subtraction - add #definition NPY_SIMD_FMA3 to check FUSED native support - add testing cases for all of the above
| * | | | | Merge pull request #17843 from BvB93/precision-pluginRalf Gommers2020-12-2217-368/+674
| |\ \ \ \ \ | | | | | | | | | | | | | | ENH: Add a mypy plugin for inferring platform-specific `np.number` precisions
| | * | | | | Update mypy_plugin.pyBas van Beek2020-12-221-1/+1
| | | | | | |
| | * | | | | DOC: Add a comment about truncating and sub-string matchingBas van Beek2020-12-221-0/+3
| | | | | | | | | | | | | | | | | | | | | | | | | | | | Co-Authored-By: Eric Wieser <425260+eric-wieser@users.noreply.github.com>
| | * | | | | ENH: Add unit support to the `datetime64`/`timedelta64` character codesBas van Beek2020-12-221-6/+66
| | | | | | |
| | * | | | | MAINT: Import `mypy` via a try/except approachBas van Beek2020-12-221-15/+23
| | | | | | | | | | | | | | | | | | | | | | | | | | | | Fixes an issue where the `pypy` tests would fail
| | * | | | | MAINT,TST: Fixed a failure due platform-dependent precisionsBas van Beek2020-12-221-4/+4
| | | | | | |
| | * | | | | TST: Add `numpy.typing.mypy_plugin` to the list of public modulesBas van Beek2020-12-222-0/+3
| | | | | | |
| | * | | | | DOC: Add a docstring to the mypy plugin entry-pointBas van Beek2020-12-221-0/+1
| | | | | | |
| | * | | | | DOC,STY: Fixed a few typos and the likesBas van Beek2020-12-221-5/+5
| | | | | | |
| | * | | | | ENH: Take the precision of `np.int_` into account arithmetic operationsBas van Beek2020-12-224-48/+50
| | | | | | |
| | * | | | | ENH: Use `intp` as return type for all appropiate functionsBas van Beek2020-12-224-28/+29
| | | | | | | | | | | | | | | | | | | | | | | | | | | | Since `intp` is now annotated it can be used in place of `signedinteger[Any]`
| | * | | | | ENH: Centralize all `DTypeLike` string aliases in `numpy.typing`Bas van Beek2020-12-223-246/+282
| | | | | | |
| | * | | | | ENH: Added a mypy plugin for handling platform-specific `np.number` precisionsBas van Beek2020-12-227-37/+229
| | | | | | |
| * | | | | | Merge pull request #18050 from BvB93/array-like2Charles Harris2020-12-222-7/+96
| |\ \ \ \ \ \ | | |/ / / / / | |/| | | | | MAINT: Add aliases for commonly used `ArrayLike` objects
| | * | | | | MAINT: Add aliases for commonly used `ArrayLike` objectsBas van Beek2020-12-212-7/+96
| | | | | | |
| * | | | | | DOC: Update examples in stack docstrings (#18054)daaawx2020-12-221-21/+21
| |/ / / / / | | | | | | | | | | | | The numbers chosen in the stacking examples can make them difficult to interpret. I suggest using different numbers to make it unambiguous.
* | | | | | remove one inner loops in order to save more cycles.Qiyu82020-12-221-26/+24
| | | | | |
* | | | | | Merge branch 'master' of github.com:numpy/numpy into countnzQiyu82020-12-2131-1113/+1750
|\ \ \ \ \ \ | |/ / / / /
| * | | | | Merge pull request #17985 from seiko2plus/ditch_simd_arithmeticCharles Harris2020-12-199-986/+1224
| |\ \ \ \ \ | | | | | | | | | | | | | | ENH, SIMD: Ditching the old CPU dispatcher(Arithmetic)
| | * | | | | MAINT: Small style fixes.Charles Harris2020-12-191-1/+1
| | | | | | |
| | * | | | | ENH, SIMD: Ditching the old CPU dispatcher(Arithmetic)Sayed Adel2020-12-149-967/+1190
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The first patch in a series of pull-requests aims to facilitate the migration process to our new SIMD interface(NPYV). It is basically a process that focuses on getting rid of the main umath SIMD source `simd.inc`, which contains almost all SIMD kernels, by splitting it into several dispatch-able sources without changing the base code, which facilitates the review process in order to speed up access to the nominal target. In this patch, we have moved the arithmetic operations of real and complex for single/double precision to the new CPU dispatcher. NOTE: previously, the SIMD code of AVX2 and AVX512F for single/double precision wasn't dispatched in runtime before.
| | * | | | | ENH: add new option 'cfunc_alias' to umath generatorSayed Adel2020-12-111-16/+23
| | | | | | |
| | * | | | | MAINT, SIMD: reduce the number of preprocessor CPU runtime dispatcher callsSayed Adel2020-12-101-15/+23
| | | | | | |
View Lorry Controller Status