Skip to content

feat: add some of the kernels using cuda.compute - #3981

Merged
ianna merged 32 commits into
scikit-hep:awkward3from
maxymnaumchyk:maxymnaumchyk/3978-awkward_indexedarray_overlay_mask
May 19, 2026
Merged

ianna merged 32 commits into
scikit-hep:awkward3from
maxymnaumchyk:maxymnaumchyk/3978-awkward_indexedarray_overlay_mask

Conversation

@maxymnaumchyk

@maxymnaumchyk maxymnaumchyk commented Apr 17, 2026 •

Copy link
Copy Markdown
Collaborator

Closes #3978, closes #3997, closes #3998, closes #3999, closes #4000, closes #4001, closes #4002, closes #4003, closes #4004, closes #4005, closes #4006, closes #4007

These kernels are only ~2 times faster-->
IndexedArray_reduce_next_nonlocal_nextshifts_64 kernel before:
Screenshot 2026-04-16 135451

IndexedArray_reduce_next_nonlocal_nextshifts_64 kernel after:
image

IndexedArray_reduce_next_64 kernel before:
image

IndexedArray_reduce_next_64 kernel after:
image

IndexedArray_overlay_mask kernel before:
image

IndexedArray_overlay_mask kernel after:
image

@github-actions

Copy link
Copy Markdown
Contributor

The documentation preview is ready to be viewed at http://preview.awkward-array.org.s3-website.us-east-1.amazonaws.com/PR3981

@maxymnaumchyk

Copy link
Copy Markdown
Collaborator Author

IndexedArray_reduce_next_64 kernel before:
Screenshot 2026-04-15 163801

IndexedArray_reduce_next_64 kernel after:
Screenshot 2026-04-15 163511

@codecov

codecov Bot commented Apr 17, 2026 •

Copy link
Copy Markdown

Codecov Report

❌ Patch coverage is 43.53448% with 131 lines in your changes missing coverage. Please review.
✅ Project coverage is 85.04%. Comparing base (6f6d816) to head (a7ac8a7).

Files with missing lines Patch % Lines
src/awkward/_connect/cuda/_compute.py 43.53% 131 Missing ⚠️

❌ Your patch check has failed because the patch coverage (43.53%) is below the target coverage (98.00%). You can increase the patch coverage or adjust the target coverage.

Additional details and impacted files
Files with missing lines Coverage Δ
src/awkward/_backends/cupy.py 100.00% <ø> (ø)
src/awkward/_connect/cuda/_compute.py 57.47% <43.53%> (-11.15%) ⬇️

@maxymnaumchyk
maxymnaumchyk marked this pull request as ready for review April 17, 2026 13:56
@maxymnaumchyk
maxymnaumchyk marked this pull request as draft April 17, 2026 13:57
@maxymnaumchyk

Copy link
Copy Markdown
Collaborator Author

ByteMaskedArray_getitem_nextcarry kernel before:
image

ByteMaskedArray_getitem_nextcarry kernel after:
Screenshot 2026-04-17 175330

@maxymnaumchyk

Copy link
Copy Markdown
Collaborator Author

awkward_ByteMaskedArray_numnull kernel before:
Screenshot 2026-04-17 175310

awkward_ByteMaskedArray_numnull kernel after:
image

@maxymnaumchyk maxymnaumchyk changed the title feat: add some of the awkward_IndexedArray kernels using cuda.compute feat: add some of the awkward_IndexedArray and ByteMaskedArray kernels using cuda.compute Apr 17, 2026
@maxymnaumchyk

Copy link
Copy Markdown
Collaborator Author

awkward_RegularArray_getitem_jagged_expand kernel before:
image

awkward_RegularArray_getitem_jagged_expand kernel after:
image

@maxymnaumchyk

Copy link
Copy Markdown
Collaborator Author

awkward_UnionArray_simplify_one kernel before:
Screenshot 2026-04-21 171211

awkward_UnionArray_simplify_one kernel after:
image

@maxymnaumchyk

Copy link
Copy Markdown
Collaborator Author

awkward_ListArray_broadcast_tooffsets kernel before:
Screenshot 2026-04-23 175153

awkward_ListArray_broadcast_tooffsets kernel after:
image

@maxymnaumchyk

Copy link
Copy Markdown
Collaborator Author

awkward_ListArray_localindex kernel before:
Screenshot 2026-04-27 144334

awkward_ListArray_localindex kernel after:
image

@maxymnaumchyk

Copy link
Copy Markdown
Collaborator Author

awkward_ListArray_compact_offsets kernel before:
Screenshot 2026-04-28 122425

awkward_ListArray_compact_offsets kernel after:
image

@maxymnaumchyk

Copy link
Copy Markdown
Collaborator Author

awkward_ListArray_combinations_length kernel before:
Screenshot 2026-04-28 130635

awkward_ListArray_combinations_length kernel after:
Screenshot 2026-04-28 172829

@maxymnaumchyk

maxymnaumchyk commented May 4, 2026 •

Copy link
Copy Markdown
Collaborator Author

awkward_ListArray_combinations kernel before:
Screenshot 2026-04-28 180225

awkward_ListArray_combinations kernel after:
image

Just in case, I also manually tested the new kernel with n = 3 and replacement = True.

Also, there is a possible optimization that can be done -- to use the calculated offsets directly from awkward_ListArray_combinations_length (now, they are being discarded).

@maxymnaumchyk

Copy link
Copy Markdown
Collaborator Author

The awkward_UnionArray_nestedfill_tags_index kernel turned out to be a little slower: 0.006263 vs 0.006439 seconds. Add just for archive.

@maxymnaumchyk maxymnaumchyk changed the title feat: add some of the awkward_IndexedArray and ByteMaskedArray kernels using cuda.compute feat: add some of the kernels using cuda.compute May 7, 2026
@maxymnaumchyk
maxymnaumchyk marked this pull request as ready for review May 7, 2026 09:46
@ianna
ianna changed the base branch from main to awkward3 May 16, 2026 19:45

@ianna ianna left a comment

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@maxymnaumchyk - Thanks! 12 more kernels migrated to cuda.compute! I'll enable auto-merge. The benchmarks will be updated after it is merged later today. Thanks.

@ianna
ianna merged commit 85e945d into scikit-hep:awkward3 May 19, 2026
35 of 38 checks passed
ianna added a commit that referenced this pull request Jun 30, 2026
* feat: complete CPU kernel migration from `parents` to `offsets` (#3988)

* next step in kernel migration from parents to offsets

* linter fix

* add jax bincount

* add cpp kernel to convert parents to offsets

* fix typo

* fix typo in an auto-generated test

* almost there

* nearly there

* cleanup

* fix remaining kernels

* final cleanup

* format json data

* format

* initialize maxnextparents  = -1 (a sentinel meaning "no bin was touched")

* update the kernels to work on offsets!

* format

* migrate cupy rawkernels from parents to offsets

* migrate jax reducers

* fix for platfroms where the int64 counts can't be safely cast

* add bincount for cupy backend

* compact loc

* fix windows build and add optional OpenMP support

* update cuda kernels

* feat: add `missing_repeat` kernel implementation using cuda.compute (#3922)

* feat: add missing_repeat implementation using cuda.compute

* keep the same name as in cpu kernel

* add tests for repetitions>1 and regularsize>1

* style: pre-commit fixes

* add keywords

* style fixes

---------

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

* feat: add `index_rpad_and_clip` kernels implementations using cuda.compute (#3923)

* feat: add `index_rpad_and_clip` kernels implementation using cuda.compute

* style fix

* add a test for `index_rpad_and_clip_axis0` that would have target>length

* add keyword names

* style

* Apply suggestions from Ianna

Co-authored-by: Ianna Osborne <ianna.osborne@cern.ch>

* style: pre-commit fixes

---------

Co-authored-by: Ianna Osborne <ianna.osborne@cern.ch>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

* fix: array attrs not being validated at creation and being of inconsistent type (#3996)

* convert dict attrs to Attrs type

* add test

* fix: ensure jax backend uses arrays on cpu only (#3990)

* ensure jax uses cpu by default

* remove now redundant jax_platform_name setting

* change error messages

* better errors

* remove DeviceArray mentions as that does not exist while we're at it too

* almost there

* cleanup

* cleanup

* remove old data file

* update cupy kernels

* remove exlicit test for depricated kernel

* add complex kernels and port them to use offsets

* fix complex reducers

* add remaining kernels

* add kernels for complex and bool sum

* move to segmented_reduce

* fix typo

* promote type

* fix complex bool reducer

* try another algo

* missed one

* use type inference

* make numba happy

* remove reducer overloads

* use sum op func for complex types

* remove test of depricated code

* avoid using == to compare floating-point products

* fix boundary tests

* cleanup

* cleanup

* remove dead code

* revert lexsort to argmin/max

* handrolled lexsort

* remove dead code

* remove bincount

* import numba cuda for jit

* fix tests

---------

Co-authored-by: maxymnaumchyk <maxymnaumchyk@gmail.com>
Co-authored-by: maxymnaumchyk <70752300+maxymnaumchyk@users.noreply.github.com>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Co-authored-by: Iason Krommydas <iason.krom@gmail.com>

* chore: profile: add benchmark script (#4073)

add benchmark script

* feat: add some of the kernels using cuda.compute (#3981)

* feat: add `awkward_IndexedArray_overlay_mask` kernel using cuda.compute

* feat: add `awkward_IndexedArray_reduce_next_64` kernel using cuda.compute

* feat: add `IndexedArray_reduce_next_nonlocal_nextshifts_64` kernel using cuda.compute

* feat: add `ByteMaskedArray_getitem_nextcarry` kernel using cuda.compute

* feat: add `awkward_ByteMaskedArray_numnull` kernel using cuda.compute

* feat: add `awkward_RegularArray_getitem_jagged_expand` kernel using cuda.compute

* add an upper bound

* style: pre-commit fixes

* feat: add `awkward_RegularArray_getitem_jagged_expand` kernel using cuda.compute

* feat: add `awkward_UnionArray_simplify_one` kernel using cuda.compute

* feat: add `awkward_ListArray_broadcast_tooffsets` kernel using cuda.compute

* feat: add `awkward_ListArray_localindex` kernel using cuda.compute

* fix the impl

* feat: add `awkward_ListArray_compact_offsets` kernel using cuda.compute

* feat: add `awkward_ListArray_combinations_length` kernel using cuda.compute

* feat: add `awkward_ListArray_combinations` kernel using cuda.compute

* feat: add `awkward_UnionArray_nestedfill_tags_index` kernel using cuda.compute

* fix the tests for kernels that are deliberately raising errors

* compare `starts` and `stops` separately

* ignore `memptr` argument for pylint

* style: pre-commit fixes

* Add functions for indexing and repeating arrays

* style: pre-commit fixes

* return unary_transform call for segment_ids

* style: pre-commit fixes

* update the `awkward_IndexedArray_reduce_next_64` to work with offsets

* style: pre-commit fixes

---------

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Co-authored-by: Ianna Osborne <ianna.osborne@cern.ch>

* feat: migrate some more kernels to cuda.compute (#4019)

* feat: add reimplementation of the `awkward_RegularArray` kernels using cuda.compute

* reimplement more kernels using cuda.compute

* reimplement more kernels using cuda.compute

* style: pre-commit fixes

* style fix

* use the named functions instead of lambdas

* add new kernels

* add new kernels

* fix a type mismatch

* style: pre-commit fixes

* add new kernels

* fix the overflow error

* new batch of kernels (draft)

* add comments and cleanup

* fix the tests messages

* add new `BitMaskedArray` kernels

* modify the `test_1381_check_errors.py` test

* resolve the conflicts

---------

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

* Fable fixes based on the report

* sort kernel issues addressed

* add tests

* add benchmarks

* add benchmarks

* fix memory leak

* add benchmarks/cuda_compute_cache_leak_reproducer.py

* fix ruff checks

* synchronization

* fix it

* fix memory leak

* more fixes

* more tests

* reduce temporaries

* more tests

* relax conditions

* fix bench

* bigfix

* fix bug

* relax condition

* add kernel fusion with parrot

* the same for more reducers

* cleanup

* apply __restrict__ uniformly

* rename macro-generated ABI symbols to match the kernel name

* add tests

* add O2 build

* fix

* add max_segment_size

* add benchmarks and cccl profiler

* pin cccl and clean up

* cleanup

* add test

* fix overflow int64

---------

Co-authored-by: maxymnaumchyk <maxymnaumchyk@gmail.com>
Co-authored-by: maxymnaumchyk <70752300+maxymnaumchyk@users.noreply.github.com>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Co-authored-by: Iason Krommydas <iason.krom@gmail.com>
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.

awkward_IndexedArray_overlay_mask.cu

2 participants