8000 PERF: improve multithreaded ufunc scaling by ngoldbaum · Pull Request #27896 · numpy/numpy · GitHub
[go: up one dir, main page]
Skip to content

PERF: improve multithreaded ufunc scaling #27896

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 10 commits into from
Dec 5, 2024
Merged

PERF: improve multithreaded ufunc scaling #27896

Changes from 1 commit
Commits
Show all changes
10 commits
Select commit Hold shift + click to select a range
f019103
PERF: add a fast path to ufunc type resolution
ngoldbaum Nov 26, 2024
8c68479
MAINT: move dispatching.c to C++
ngoldbaum Dec 3, 2024
7b044a6
MAINT: move npy_hashtable to C++ and use std::shared_mutex
ngoldbaum Dec 3, 2024
cee2892
MAINT: fix windows linking
ngoldbaum Dec 3, 2024
1a1ecfc
MAINT: remove outdated comment
ngoldbaum Dec 3, 2024
c7b9b76
MAINT: only call try_promote on free-threaded build
ngoldbaum Dec 3, 2024
f2b7cc1
MAINT: try to give new function a name indicating it uses a mutex
ngoldbaum Dec 3, 2024
6146085
MAINT: only do complicated casting to get a mutex pointer once
ngoldbaum Dec 3, 2024
e150832
MAINT: use std::nothrow to avoid dealing with exceptions
ngoldbaum Dec 4, 2024
b0f744c
DOC: add changelog
ngoldbaum Dec 4, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Prev Previous commit
Next Next commit
MAINT: try to give new function a name indicating it uses a mutex
  • Loading branch information
@ngoldbaum
ngoldbaum committed Dec 3, 2024
commit f2b7cc1ff73daaa343c47c5aa6f482fbdfd2a344
5 changes: 3 additions & 2 deletions numpy/_core/src/umath/dispatching.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -904,7 +904,8 @@ promote_and_get_info_and_ufuncimpl(PyUFuncObject *ufunc,
* only acquires a write lock on a cache miss to fill the cache
*/
static inline PyObject *
try_promote_and_get_info_and_ufuncimpl(PyUFuncObject *ufunc,
promote_and_get_info_and_ufuncimpl_with_locking(
PyUFuncObject *ufunc,
PyArrayObject *const ops[],
PyArray_DTypeMeta *signature[],
PyArray_DTypeMeta *op_dtypes[],
Expand Down Expand Up @@ -1022,7 +1023,7 @@ promote_and_get_ufuncimpl(PyUFuncObject *ufunc,
}

#ifdef Py_GIL_DISABLED
PyObject *info = try_promote_and_get_info_and_ufuncimpl(ufunc,
PyObject *info = promote_and_get_info_and_ufuncimpl_with_locking(ufunc,
ops, signature, op_dtypes, legacy_promotion_is_possible);
#else
PyObject *info = promote_and_get_info_and_ufuncimpl(ufunc,
Expand Down
0