Skip to content

Fake cupyx.scipy: copy and deepcopy of proxy objects - #98

Merged
max-models merged 1 commit into
develfrom
fake-cupyx-proxy-deepcopy
Oct 10, 2026
Merged

max-models merged 1 commit into
develfrom
fake-cupyx-proxy-deepcopy

Conversation

@max-models

Copy link
Copy Markdown
Owner

No description provided.

@max-models
max-models marked this pull request as ready for review October 10, 2026 15:24
@max-models
max-models merged commit c64ef2d into devel Oct 10, 2026
17 checks passed
max-models added a commit that referenced this pull request Oct 13, 2026
* Numpy docstrings (#91)

* Docstring tooling: ruff D rules, numpydoc validation, doctests, autodoc reference

* Rewrite docstrings in NumPy format with runnable examples

* Trim api.md to an overview; move its details into guides and a CUDA headers page

* Fix CudaStruct.from_signature with postponed annotations

* Load the postponed-annotations test helper by path

* Add mpi exchange (#92)

* Added xp.mpi.exchange

* Added testing/smoke_wheel.py

* Fix docs

* Add fake cupyx.scipy and batched xp.optimize.newton (#93)

* Add fake cupyx.scipy and batched xp.optimize.newton

* Docs: resolve scipy cross-references via intersphinx

* fix: segment_sum on the fake CuPy sums on the host buffers (#94)

* More xp scipy (#95)

* fix: segment_sum on the fake CuPy sums on the host buffers

* add xp.scipy.linalg.solve_circulant on the device

* Add xp.optimize.fsolve/root/minimize and xp.integrate.quad/odeint with device callbacks

* Added MIT license

* More xp scipy (#96)

* fix: segment_sum on the fake CuPy sums on the host buffers

* add xp.scipy.linalg.solve_circulant on the device

* Add xp.optimize.fsolve/root/minimize and xp.integrate.quad/odeint with device callbacks

* Added MIT license

* Support CuPy >= 14.0: fake cupyx.scipy names of 14.0.0, checked against CuPy in the GPU CI

* Hide the real cupyx in the fake_cupyx test fixture

* Fake cupyx.scipy: copy and deepcopy of proxy objects (#98)

* Faster type checking on numpy (#99)

* fix: segment_sum on the fake CuPy sums on the host buffers

* add xp.scipy.linalg.solve_circulant on the device

* Add xp.optimize.fsolve/root/minimize and xp.integrate.quad/odeint with device callbacks

* Added MIT license

* Support CuPy >= 14.0: fake cupyx.scipy names of 14.0.0, checked against CuPy in the GPU CI

* Hide the real cupyx in the fake_cupyx test fixture

* more xp scipy

* formatting

* numpydoc fix

* Amd support (#100)

* AMD support via hipify

* Updated job names

* Remove _trap() on HIP

* Fix tests

* formatting
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant