CUDA: add basic complex number operation tests\n\nAdds kernel-based t…

[CodersAcademy006]

This PR adds kernel-based tests for device-side complex64 operations: addition, multiplication, and absolute value.

• Skips under cudasim due to divergent semantics

• Uses NumPy as the reference oracle

• Scope is intentionally limited to safe, well-defined CUDA operations (no division, power, or mixed-type ops)

This PR is intentionally narrow and reviewer-friendly, following the same discipline as previous CUDA test PRs.
#515

[copy-pr-bot]

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

[greptile-apps]

Greptile Summary

Adds three kernel-based tests for device-side complex64 operations: addition, multiplication, and absolute value. The tests use NumPy as the reference oracle and skip execution under cudasim due to divergent semantics.

Key changes:

• Added skip_on_cudasim import to support decorator usage
• Implemented three CUDA kernels: complex_add_kernel, complex_mul_kernel, and complex_abs_kernel
• Added TestCudaComplexBasicOps test class with three test methods that validate kernel behavior against NumPy
• All code properly structured with SPDX header, imports, and kernel definitions in correct order

The implementation is clean, follows existing test patterns in the codebase, and addresses issue #515 by expanding CUDA test coverage for complex number operations.

Confidence Score: 5/5

• Safe to merge - well-structured test addition with no runtime risks
• All previous syntax and structural issues have been resolved. The code has proper imports, correct file structure with SPDX header at top, kernels and test class at the end, no duplicates, and follows existing test patterns in the codebase.
• No files require special attention

Important Files Changed

Filename	Overview
numba_cuda/numba/cuda/tests/cudapy/test_complex.py	Added three new tests for basic complex64 operations (addition, multiplication, absolute value) with proper CUDA kernel implementations - no issues found

[greptile-apps]

_{1 file reviewed, 2 comments}

_{Edit Code Review Agent Settings | Greptile}

[greptile-apps]

Additional Comments (1)

1. numba_cuda/numba/cuda/tests/cudapy/test_complex.py, line 62-66 (link)

   syntax: duplicate numpy import - already imported on line 2 as np

_{1 file reviewed, 2 comments}

_{Edit Code Review Agent Settings | Greptile}

[greptile-apps]

_{1 file reviewed, 1 comment}

_{Edit Code Review Agent Settings | Greptile}

[greptile-apps]

_{1 file reviewed, 1 comment}

_{Edit Code Review Agent Settings | Greptile}

[greptile-apps]

_{1 file reviewed, 1 comment}

_{Edit Code Review Agent Settings | Greptile}

[greptile-apps]

_{1 file reviewed, 2 comments}

_{Edit Code Review Agent Settings | Greptile}

[greptile-apps]

_{1 file reviewed, 2 comments}

_{Edit Code Review Agent Settings | Greptile}

[greptile-apps]

_{1 file reviewed, 3 comments}

_{Edit Code Review Agent Settings | Greptile}

[greptile-apps]

Additional Comments (1)

1. numba_cuda/numba/cuda/tests/cudapy/test_complex.py, line 102-141 (link)

   style: duplicate imports - these imports are redundant since they're already present in lines 5-44 (after fixing the decorator on line 4)

_{1 file reviewed, 4 comments}

_{Edit Code Review Agent Settings | Greptile}