[pre-commit.ci] pre-commit autoupdate #937
Open
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
![pre-commit-ci[bot]](https://avatars.githubusercontent.com/in/68672?s=60&v=4){kind=small}
pre-commit-ci
bot
force-pushed
the
pre-commit-ci-update-config
branch
from
dd1209c to
ef7f16e
Compare
pre-commit-ci
bot
force-pushed
the
pre-commit-ci-update-config
branch
from
5b8cb3d to
b83cf2d
Compare
* updating legion commit hash
* Clean up and unify the main cmake paths
* Fix legate path detection after using new project name
…1629) * Have configure tell cmake which variables it needs to dump to JSON * Don't emit the config vars if not driven by configure
All policies were being set to `NEW` and the newest policy (135) was added in CMake 3.24. The current minimum is 3.26.4 and so all policies added before that are set to new by virtue of the minimum version.
* Add tests to enhance type_info.pyx coverage * Update tests/python/unit/legate/core/test_type.py Co-authored-by: Jacob Faibussowitsch <[email protected]> * Update tests/python/unit/legate/core/test_type.py Co-authored-by: Jacob Faibussowitsch <[email protected]> * Update tests/python/unit/legate/core/test_type.py Co-authored-by: Jacob Faibussowitsch <[email protected]> * Fix rect_type alignment check * Fix primitive size check --------- Co-authored-by: Jacob Faibussowitsch <[email protected]>
* Remove variant options that we actually don't want to expose * Make the tests pass by using indefinite upper bounds everywhere * Looks like emtpy buffers cause no problem now * Memory IDs look better with std::hex * Let's set upper bounds for system memory for now * Looks like indefinite bounds make replicate write tests hang * Temporariliy disable a test that binds a buffer bigger than the extent * Update the Legion commit to fetch the onepool branch by default * Remove obsolete flags * Temporarily increase the upper bounds for Artem * Bump the legion commit * Configure memory pools for GPUs * Re-enable the test that couldn't run due to a runtime bug * Start taking pool sizes from users * Bump the Legion commit * Allow scalar extraction tasks to create buffers * Workaround for tasks that are registered through the side channel * Cherry pick a fix for the NCCL test * Missing early exit * Post merge clean-up * Legate STL tasks don't create allocations * Changes for the onepool API v4 * Bump the Legion commit * Print has_allocations * Minor fix for the store proxy * Bump the legion commit * Add some useful debugging logs * Reenable future mapping and use the right alignment for output buffers and legate metadata * Make sure we have enough space in the pool when futures are forced to be mapped to the zero-copy memory * Move instance creations out of the critical sections * Add APIs to query a single scalar and the number of scalars * Add APIs to query the launch domain and task type * Tasks are not allowed to create allocations by default from now * Filter out invalid stores in mapper proxies to reduction tasks * Bump the Legion commit * Fix C++ tests for the default option change * Update the legion commit hash * Update the legion commit hash to fetch a fix * Update the legion commit hash * Post-merge cleanup * Address some of the review comments * Ask Legion to cache all temporary allocations * Prefetch some changes from #1545 to build the code * Update the unit test * Cherry-pick 1466 * Bump the legion commit to fetch the bug fix * Fix clang-tidy issues * Take the alignment into account when picking pool sizes * Mark more tasks as allocable * Update the legion commit * Add comments about left-over items * Fix a build issue * Fix python tests * Update documentation * Typo * Address review comments * Fix clang-tidy issue * Address review comments * Bump the legion commit to fetch a fix * More review comments * fixup! More review comments
* Allow traced exception to chain
* Replace many python linters with `ruff`
* Allow top-level task to select variant
* Add unit tests for formatter * Fix the formatter test failure * Address review comments
* refactor physical_store.cc test file * fix clang-format and update Copyright * address comments * Add VariantOptions to tasks
* fixup traced exception docstring
* Clean up CODEOWNERS files to conform to new directory layout
Fixed by adding explicit template instantiations for static member `identity` of the template classes `ElementWiseMin/Max`. * Fix template instantiation warning for clang 18 * add explicit instantiation of the member variable, not the entire class * gcc doesn't like auto. Trying explicit type name
Removing tests/benchmarks from the conda package Fix the logic to only enable testing when upload is disabled. Also add arguments to switch to the new conda packaging format.
* Use TaskConfig in Python tasks
… (#2033) * Add error checking for file(DOWNLOAD) when downloading rapids-cmake
* return a better error message * fixup * fixup * Update src/cpp/legate/runtime/detail/mpi_detection.cc Co-authored-by: Jacob Faibussowitsch <[email protected]> * Update src/cpp/legate/runtime/detail/mpi_detection.cc Co-authored-by: Jacob Faibussowitsch <[email protected]> * fixup --------- Co-authored-by: Jacob Faibussowitsch <[email protected]>
* Append upload action with 'tested' if corresponding tests pass. * Upload takes place in the context of "with_tests", but passes upload-enabled=true, which actually will upload the package w/o with_tests. * The "build-has-tests" needs to be false for no "with_tests" * Remove trailing space * Combine upload jobs into 1. * Run upload irrespective of test status. * Remove always() * Use always() to run the job irrespective of the dependency failure. * Use legate-gh-ci v1.30 for upload. --------- Co-authored-by: Marcin Zalewski <[email protected]>
* Add Span::rbegin/rend and Span::front/back
Push source code to nv-legate/legate as part of the nightly CI run. Tracking issue: LLRDO-335.
* Support MPICH mpirun in legate launcher
The UCX wheel library loader has logic to attempt to prefer the system UCX over the bundled wheels in pip. Add some logic to make the init aware of being in a wheel build and perform some custom logic for loading UCX before any other libraries are loaded. Thanks to Vyas Ramasubramani and Jacob Faibussowitsch for advice on putting this together.
…. (#1676) Closes #1200 Currently, store.transpose() gets mapped to physical memory in the dimension order of the original store and not store.transpose(). This PR adds adjust_dimension_ordering() to the transform stack of a store so that the dimension ordering for the layout can be adjusted based on any transpose() operations in the transform stack. * WIP: Create the inital test. Needs fixes * WIP: Fix for the transpose order bug is in * more tests for transpose order * some cleanup and fixes * linter fixes * fix build after merge from main * implement the inverse transform for dimension ordering * WIP: cleanup * WIP: cleanup and renaming * WIP: rework the test code * tests for promote, project and delinarize when combined with transpose * tidy fixes * more tidy and build fixes * fixed lint error * tidy fixes * fix format Co-authored-by: Jacob Faibussowitsch <[email protected]> * remove comment Co-authored-by: Jacob Faibussowitsch <[email protected]> * fix spacing Co-authored-by: Jacob Faibussowitsch <[email protected]> * use move Co-authored-by: Jacob Faibussowitsch <[email protected]> * fix spacing. Co-authored-by: Jacob Faibussowitsch <[email protected]> * use move Co-authored-by: Jacob Faibussowitsch <[email protected]> * simplify test code structure * WIP: address review comments * Address review comments * tidy fixes * Address review comments * add some missing stuff to span * address review comments * tidy fix --------- Co-authored-by: Jacob Faibussowitsch <[email protected]>
…(#2044) * Maintenance on StoreMapping, add proper constructors and accessors
Change the default back to wheels and let the user override to opt in.
The git SHA and build number should be at the end of the build string, the build number is especially important in assisting the solver in resolving the latest version that should be installed.
* Expose many more Python classes' documentation
…ke 4.0 (#2063) * Set macOS deployment target and sdkroot environment variables for cmake 4.0
…ne mapping (#2051) * Keep a reference to the Logical store in the Physical store when inline mapping
* Minor cleanup of the legate launcher environment variables * Remove most of the Legion paths entries * Remove most of the LegatePaths entries as well
…020) This keeps allowing the `Registrar` including fixing all of the changes that were needed to cupynumeric/legate-boost It emoves any direct `Registarar.record_task()` but these are effectively considered private API. I don't think the options are useful and I think using the same name is not great, but this is the minimal (maximum compatible) start. Signed-off-by: Sebastian Berg <[email protected]> * Address Jacob's review comments Co-authored-by: Jacob Faibussowitsch <[email protected]> Signed-off-by: Sebastian Berg <[email protected]> * Tweak docs, and don't use doxygen as we don't want it publically documented Signed-off-by: Sebastian Berg <[email protected]> --------- Signed-off-by: Sebastian Berg <[email protected]> Co-authored-by: Jacob Faibussowitsch <[email protected]>
* Link to example code in io docs
* refactor TaskStoreTests * improve scalar type conversion and task registration * add inline comments for literal bool parameters * Apply angle-includes formatting * Fixed clang-tidy warning
updates: - [github.com/adrienverge/yamllint: v1.36.0 → v1.37.0](adrienverge/yamllint@v1.36.0...v1.37.0) - [github.com/astral-sh/ruff-pre-commit: v0.11.0 → v0.11.4](astral-sh/ruff-pre-commit@v0.11.0...v0.11.4)
pre-commit-ci
bot
force-pushed
the
pre-commit-ci-update-config
branch
from
434f8c0 to
2454b8d
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
updates: