Fix lanczos launch grid

[achirkin]

Fix accidentally swapped block size and grid size arguments.

#3104 (comment)

[coderabbitai]

📝 Walkthrough

Summary by CodeRabbit

• Bug Fixes

  • Corrected GPU processing dimensions for sparse matrix operations.
  • Improved kernel launch configuration to ensure reliable triangular-matrix population and beta-column updates.

• Chores

  • Updated copyright attribution to include NVIDIA Corporation and affiliates.

Walkthrough

The patch updates the SPDX attribution and corrects CUDA grid-dimension calculations and kernel launch argument ordering in the Lanczos solver.

Changes

Lanczos CUDA corrections

Layer / File(s)	Summary
Correct CUDA grid and launch configuration cpp/include/raft/sparse/solver/detail/lanczos.cuh	The code uses raft::div_rounding_up_safe for grid calculations. The triangular population kernel now receives grid size before block size. The optional beta-kernel launch uses the same safe calculation. The SPDX attribution includes NVIDIA Corporation and affiliates.

Estimated code review effort: 1 (Trivial) | ~5 minutes

Suggested reviewers: lowener

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Title check	✅ Passed	The title clearly and concisely identifies the main change: fixing the Lanczos kernel launch grid.
Description check	✅ Passed	The description directly explains that the pull request corrects swapped block size and grid size arguments.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

_{Comment @coderabbitai help to get the list of available commands.}

[coderabbitai]

Actionable comments posted: 2

🧹 Nitpick comments (1)

cpp/include/raft/sparse/solver/detail/lanczos.cuh (1)

160-166: 📐 Maintainability & Code Quality | 🔵 Trivial | ⚡ Quick win

Include raft/util/integer_utils.hpp directly.

lanczos.cuh uses raft::div_rounding_up_safe but does not include its defining header. Add the direct include.

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@cpp/include/raft/sparse/solver/detail/lanczos.cuh` around lines 160 - 166,
Add a direct include of raft/util/integer_utils.hpp in lanczos.cuh so the
visible uses of raft::div_rounding_up_safe are backed by their defining header;
leave the surrounding kernel launch and block calculations unchanged.

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@cpp/include/raft/sparse/solver/detail/lanczos.cuh`:
- Around line 160-166: Before the kernel launches in the Lanczos setup path, add
RAFT_EXPECTS validation that enforces n_components + 1 < ncv and rejects k >=
ncv, ensuring invalid dimensions cannot produce zero-block launches or
out-of-bounds writes in kernel_triangular_beta_k.
- Around line 162-168: Add immediate CUDA error checks after both visible kernel
launches in the surrounding Lanczos flow: insert
RAFT_CUDA_TRY(cudaGetLastError()) directly after the triangular beta kernel
launch and the preceding kernel launch, preserving the existing conditional
structure and launch behavior.

---

Nitpick comments:
In `@cpp/include/raft/sparse/solver/detail/lanczos.cuh`:
- Around line 160-166: Add a direct include of raft/util/integer_utils.hpp in
lanczos.cuh so the visible uses of raft::div_rounding_up_safe are backed by
their defining header; leave the surrounding kernel launch and block
calculations unchanged.

🪄 Autofix

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: 7a885499-e115-4b9c-8dfd-55f2cd9dd8d9

📥 Commits

Reviewing files that changed from the base of the PR and between d254b0e and 8ad66c7.

📒 Files selected for processing (1)

• cpp/include/raft/sparse/solver/detail/lanczos.cuh

[divyegala]

Could you please add a regression test?

[achirkin]

I did want to add it, but apparently any possible configuration is valid - just slow (low residency)

[aamijar]

Good catch!

[achirkin]

/merge