Skip to content

New 1-nn primitive using cutile - #2552

Open
divyegala wants to merge 11 commits into
NVIDIA:release/26.10from
divyegala:cutile-1nn
Open

New 1-nn primitive using cutile#2552
divyegala wants to merge 11 commits into
NVIDIA:release/26.10from
divyegala:cutile-1nn

Conversation

@divyegala

@divyegala divyegala commented Sep 3, 2026 •
edited
Loading

Copy link
Copy Markdown
Contributor

Relies on #2548. This PR also adds the three ways to do a 1-nn behind a new primitive called top_1_nn.

@divyegala
divyegala requested review from a team as code owners September 3, 2026 21:24
@divyegala
divyegala requested a review from jameslamb September 3, 2026 21:24
@divyegala divyegala added feature request New feature or request non-breaking Introduces a non-breaking change labels Sep 3, 2026
@divyegala
divyegala removed the request for review from a team September 4, 2026 01:01
tarang-jain
tarang-jain reviewed Sep 4, 2026
View reviewed changes
Comment thread cpp/include/cuvs/detail/jit_lto/fused_distance_nn/fused_1nn_fragments.hpp Outdated
Comment thread cpp/src/distance/fused_distance_nn-inl.cuh Outdated
Comment thread cpp/src/distance/fused_distance_nn-inl.cuh Outdated
Comment thread cpp/src/distance/fused_distance_nn-inl.cuh Outdated
Comment thread cpp/src/distance/fused_distance_nn-inl.cuh Outdated
Comment thread cpp/src/distance/fused_distance_nn-inl.cuh Outdated
tarang-jain
tarang-jain reviewed Sep 4, 2026
View reviewed changes
Comment thread cpp/tests/neighbors/distance_nn.cu
Comment thread cpp/tests/neighbors/distance_nn_helper.cuh Outdated
tarang-jain
tarang-jain reviewed Sep 4, 2026
View reviewed changes
Comment thread cpp/src/distance/detail/fused_distance_nn/cutile/fused_1nn_tile.hpp Outdated
Comment thread cpp/src/distance/detail/fused_distance_nn/cutile/fused_1nn_tile.hpp Outdated
Comment thread cpp/CMakeLists.txt Outdated
tarang-jain
tarang-jain reviewed Sep 8, 2026
View reviewed changes
Comment thread cpp/src/distance/fused_distance_nn-inl.cuh Outdated
Comment thread cpp/src/distance/fused_distance_nn-inl.cuh Outdated
Comment thread cpp/src/distance/fused_distance_nn-inl.cuh Outdated
Comment thread cpp/src/distance/detail/fused_distance_nn/cutile/fused_1nn_tile.cu
Comment thread cpp/src/distance/fused_distance_nn-inl.cuh Outdated
tarang-jain
tarang-jain reviewed Sep 8, 2026
View reviewed changes
Comment thread cpp/tests/neighbors/distance_nn.cu
@imotov
Sync up API docs with code
952d952 
Unfortunately, when NVIDIA#2547 was merged, the docs had already managed to drift from the code. This commit catches them up.
@divyegala
divyegala changed the base branch from main to release/26.10 September 8, 2026 20:13
@divyegala
divyegala requested a review from a team as a code owner September 8, 2026 20:13
@divyegala
divyegala requested a review from a team as a code owner September 8, 2026 20:47
@divyegala
divyegala removed request for a team September 8, 2026 22:46
KyleFromNVIDIA
KyleFromNVIDIA requested changes Sep 9, 2026
View reviewed changes
Comment thread cpp/cmake/modules/generate_cutile_tile_metadata.py Outdated
tfeher
tfeher requested changes Sep 9, 2026
View reviewed changes

@tfeher tfeher left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

Thanks Divye for the PR! Looks good overall, I have only minor comments.

Comment thread cpp/src/distance/detail/fused_distance_nn/cutile/fused_1nn_kernel.py
Comment thread cpp/src/distance/detail/fused_distance_nn/cutile/fused_1nn_tile.cu Outdated
Comment thread cpp/src/distance/detail/fused_distance_nn/cutile/fused_1nn_tile.hpp Outdated
Comment thread cpp/src/distance/fused_distance_nn-inl.cuh Outdated
KyleFromNVIDIA
KyleFromNVIDIA approved these changes Sep 9, 2026
View reviewed changes

@KyleFromNVIDIA KyleFromNVIDIA 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.

Looks good, just one small request. I don't need to approve again.

Comment thread cpp/cmake/modules/cutile_tile_metadata.hpp.in Outdated
KyleFromNVIDIA
KyleFromNVIDIA reviewed Sep 9, 2026
View reviewed changes
Comment thread cpp/cmake/modules/cutile_tile_metadata.hpp.in Outdated
tfeher
tfeher approved these changes Sep 10, 2026
View reviewed changes

@tfeher tfeher left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

LGTM!

huuanhhuyn
huuanhhuyn approved these changes Sep 10, 2026
View reviewed changes

@huuanhhuyn huuanhhuyn left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

LGTM

dantegd
dantegd requested changes Sep 10, 2026
View reviewed changes

@dantegd dantegd left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

Some small comments

Comment thread cpp/src/distance/fused_distance_nn-inl.cuh
Comment thread cpp/src/distance/fused_distance_nn-inl.cuh Outdated
Comment thread cpp/src/distance/detail/fused_distance_nn/cutile/fused_1nn_kernel.py Outdated
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@KyleFromNVIDIA KyleFromNVIDIA KyleFromNVIDIA approved these changes
@dantegd dantegd dantegd requested changes
@tfeher tfeher tfeher approved these changes
@tarang-jain tarang-jain tarang-jain approved these changes
@huuanhhuyn huuanhhuyn huuanhhuyn approved these changes
@jameslamb jameslamb Awaiting requested review from jameslamb jameslamb was automatically assigned from NVIDIA/adi-packaging-codeowners

Requested changes must be addressed to merge this pull request.

Assignees

No one assigned

Labels

feature request New feature or request non-breaking Introduces a non-breaking change

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

8 participants

@divyegala @dantegd @tfeher @tarang-jain @KyleFromNVIDIA @huuanhhuyn @caryr35 @imotov