PR573_FILE_SUMMARY.md

PR #573 File Changes Summary

Legend: A = Added | M = Modified | D = Deleted | R = Renamed

---

📁 .claude/

[M] CLAUDE.md

Refresh .claude/CLAUDE.md documentation docs: add ILKernelGenerator documentation and refactor plan docs: update CLAUDE.md - mark medium severity bugs as fixed docs: update CLAUDE.md bug list - mark fixed bugs feat(api): complete kernel API audit with NumPy 2.x alignment feat(kernel): add SIMD axis reduction kernels and fix NaN edge cases feat: complete IL kernel migration batch 2 - Dot.NDMD, CumSum axis, Shift, Var/Std SIMD fix(unique): sort unique values to match NumPy behavior

📁 Root

[A] .gitattributes

chore: add .gitattributes for consistent line endings

[M] .gitignore

chore(.gitignore): exclude local design documents chore: remove worktree tracking from git, add to .gitignore

[A] CHANGES.md

chore: remove internal planning docs and changelog from PR docs: fix broken documentation URLs (scipy → numpy.org) refactor(random): align parameter names and docs with NumPy API refactor: move broadcast utilities from DefaultEngine to Shape struct

[M] README.md

docs: fix broken documentation URLs (scipy → numpy.org) docs: fix versioned numpy URLs in README to use /stable/

[A] RELEASE_0.41.0-prerelease.md

docs: fix broken documentation URLs (scipy → numpy.org)

📁 benchmark/NumSharp.Benchmark.GraphEngine/Benchmarks/Allocation/

[A] AllocationMicroBenchmarks.cs

feat(benchmark): add NativeMemory allocation benchmarks for issue #528 feat(simd): add SIMD helpers for reductions, fix np.any bug, NumPy NaN handling

[A] AllocationSizeBenchmarks.cs

feat(benchmark): add NativeMemory allocation benchmarks for issue #528 feat(simd): add SIMD helpers for reductions, fix np.any bug, NumPy NaN handling

[A] NumSharpAllocationBenchmarks.cs

feat(benchmark): add NativeMemory allocation benchmarks for issue #528 feat(simd): add SIMD helpers for reductions, fix np.any bug, NumPy NaN handling

[A] ZeroInitBenchmarks.cs

feat(benchmark): add NativeMemory allocation benchmarks for issue #528 feat(simd): add SIMD helpers for reductions, fix np.any bug, NumPy NaN handling

📁 benchmark/NumSharp.Benchmark.GraphEngine/Benchmarks/Arithmetic/

[M] ModuloBenchmarks.cs

feat(benchmark): comprehensive benchmark infrastructure with full NumPy comparison

📁 benchmark/NumSharp.Benchmark.GraphEngine/Benchmarks/

[A] SimdVsScalarBenchmarks.cs

bench: add SIMD vs scalar benchmark suite

📁 benchmark/NumSharp.Benchmark.GraphEngine/Infrastructure/

[M] BenchmarkBase.cs

feat(benchmark): comprehensive benchmark infrastructure with full NumPy comparison

📁 benchmark/NumSharp.Benchmark.GraphEngine/

[M] Program.cs

bench: add SIMD vs scalar benchmark suite

📁 benchmark/NumSharp.Benchmark.Python/

[M] numpy_benchmark.py

feat(benchmark): comprehensive benchmark infrastructure with full NumPy comparison

📁 benchmark/

[M] README.md

feat(benchmark): comprehensive benchmark infrastructure with full NumPy comparison

[M] benchmark-report.md

feat(benchmark): comprehensive benchmark infrastructure with full NumPy comparison

[M] run-benchmarks.ps1

feat(benchmark): comprehensive benchmark infrastructure with full NumPy comparison

📁 benchmark/scripts/

[M] merge-results.py

feat(benchmark): comprehensive benchmark infrastructure with full NumPy comparison

📁 docs/

[A] exceptions.md

Added plans to unfinished work

📁 docs/issues/

[M] issue-0070-let-more-people-know-about-numsharp.md

chore: normalize line endings in issue docs

[M] issue-0075-implement-numpy.asarray.md

chore: normalize line endings in issue docs docs: fix broken documentation URLs (scipy → numpy.org)

[M] issue-0078-implement-numpy.where.md

chore: normalize line endings in issue docs docs: fix broken documentation URLs (scipy → numpy.org)

[M] issue-0095-extend-the-guidelines.md

chore: normalize line endings in issue docs

[M] issue-0105-implement-numpy.vdot.md

chore: normalize line endings in issue docs docs: fix broken documentation URLs (scipy → numpy.org)

[M] issue-0106-implement-numpy.inner.md

chore: normalize line endings in issue docs docs: fix broken documentation URLs (scipy → numpy.org)

[M] issue-0108-implement-numpy.tensordot.md

chore: normalize line endings in issue docs docs: fix broken documentation URLs (scipy → numpy.org)

[M] issue-0114-implement-numpy.fft.fft.md

chore: normalize line endings in issue docs docs: fix broken documentation URLs (scipy → numpy.org)

[M] issue-0116-intel-math-kernel-library-mkl.md

chore: normalize line endings in issue docs

[M] issue-0129-doc-better-specification-for-all-classes-what-class-has-what-task.md

chore: normalize line endings in issue docs

[M] issue-0190-compressed-sparse-format.md

chore: normalize line endings in issue docs

[M] issue-0202-implement-np.pad.md

chore: normalize line endings in issue docs docs: fix broken documentation URLs (scipy → numpy.org)

[M] issue-0210-add-numpy.all.md

chore: normalize line endings in issue docs docs: fix broken documentation URLs (scipy → numpy.org)

[M] issue-0211-implement-scipy-interpolate.md

chore: normalize line endings in issue docs docs: fix broken documentation URLs (scipy → numpy.org)

[M] issue-0220-numpy.flip.md

chore: normalize line endings in issue docs docs: fix broken documentation URLs (scipy → numpy.org)

[M] issue-0221-numpy.rot90.md

docs: fix broken documentation URLs (scipy → numpy.org)

[M] issue-0238-how-to-mimic-pythons-nice-column-and-row-access-i.e-matrix-2.md

chore: normalize line endings in issue docs

[M] issue-0239-np.linalg.norm.md

chore: normalize line endings in issue docs

[M] issue-0284-discussion-ground-rules-and-library-structure-architecture.md

chore: normalize line endings in issue docs

[M] issue-0298-implement-numpy.random.choice.md

chore: normalize line endings in issue docs docs: fix broken documentation URLs (scipy → numpy.org)

[M] issue-0315-tostring-should-truncate-its-output.md

chore: normalize line endings in issue docs

[M] issue-0326-lazy-loading.md

chore: normalize line endings in issue docs

[M] issue-0340-memory-limitations.md

chore: normalize line endings in issue docs

[M] issue-0341-ndarray-string-problem.md

chore: normalize line endings in issue docs

[M] issue-0343-built-in-system.drawing.image-and-bitmap-methods.md

chore: normalize line endings in issue docs

[M] issue-0349-scipy.signal.md

chore: normalize line endings in issue docs

[M] issue-0351-proper-way-to-iterate-using-ienumerable-t.md

chore: normalize line endings in issue docs

[M] issue-0360-np.any.md

chore: normalize line endings in issue docs docs: fix broken documentation URLs (scipy → numpy.org)

[M] issue-0361-mixing-indices-and-slices-in-ndarray.md

chore: normalize line endings in issue docs

[M] issue-0362-implicit-operators-for.md

chore: normalize line endings in issue docs docs: fix broken documentation URLs (scipy → numpy.org)

[M] issue-0363-add-nditerator-t-overload-with-support-for-specific-axis.md

chore: normalize line endings in issue docs

[M] issue-0365-np.nonzero.md

chore: normalize line endings in issue docs docs: fix broken documentation URLs (scipy → numpy.org)

[M] issue-0366-masking-ndarray-nd.md

chore: normalize line endings in issue docs

[M] issue-0368-masking-a-slice-...-returns-null.md

chore: normalize line endings in issue docs

[M] issue-0369-slicing-notsupportedexception.md

chore: normalize line endings in issue docs

[M] issue-0372-clustering-example.md

chore: normalize line endings in issue docs

[M] issue-0374-np.append.md

chore: normalize line endings in issue docs

[M] issue-0375-slice-assignment.md

chore: normalize line endings in issue docs

[M] issue-0378-add-np.frombuffer.md

chore: normalize line endings in issue docs docs: fix broken documentation URLs (scipy → numpy.org)

[M] issue-0383-is-there-any-way-to-convert-numsharp.ndarray-to-numpy.ndarray.md

chore: normalize line endings in issue docs

[M] issue-0384-save-ndarray-as-png-image.md

chore: normalize line endings in issue docs

[M] issue-0386-how-to-read-.csv-file-with-numsharp.md

chore: normalize line endings in issue docs

[M] issue-0390-how-to-create-an-ndarray-from-pointer-and-nptypecode.md

chore: normalize line endings in issue docs

[M] issue-0396-bitmap.tondarray-problem-with-odd-bitmap-width.md

chore: normalize line endings in issue docs

[M] issue-0397-missing-np.tile.md

chore: normalize line endings in issue docs

[M] issue-0398-typo-in-library-np.random.stardard.md

chore: normalize line endings in issue docs

[M] issue-0401-how-to-convert-ndarray-to-list.md

chore: normalize line endings in issue docs

[M] issue-0405-np.argsort-not-sorting-properly.md

chore: normalize line endings in issue docs

[M] issue-0406-c-convert-image-to-ndarray.md

chore: normalize line endings in issue docs

[M] issue-0407-np.negative-is-not-working.md

chore: normalize line endings in issue docs

[M] issue-0408-np.meshgrid-has-a-hidden-error-returning-wrong-results.md

chore: normalize line endings in issue docs

[M] issue-0410-np.save-fails-with-indexoutofrangeexception-for-jagged-arrays.md

chore: normalize line endings in issue docs

[M] issue-0411-pyobject-to-ndarray.md

chore: normalize line endings in issue docs

[M] issue-0412-the-type-ndarray-exists-in-both-numsharp.core-version-0.20.5.0-and-numsh.md

chore: normalize line endings in issue docs

[M] issue-0413-ndarray-split.md

chore: normalize line endings in issue docs

[M] issue-0414-implementation-of-np.delete-working.md

chore: normalize line endings in issue docs

[M] issue-0416-how-to-make-numsharp.ndarray-from-numpy.ndarray.md

chore: normalize line endings in issue docs

[M] issue-0418-help-me.md

chore: normalize line endings in issue docs

[M] issue-0421-performance.md

chore: normalize line endings in issue docs

[M] issue-0422-index-of-element-with-a-condiction.md

chore: normalize line endings in issue docs

[M] issue-0423-system.notimplementedexception-somearray-np.frombuffer-bytebuffer.toa.md

chore: normalize line endings in issue docs

[M] issue-0424-the-type-or-namespace-name-numsharp-could-not-be-found-are-you-missing-a-usin.md

chore: normalize line endings in issue docs

[M] issue-0426-arctan2-returning-incorrect-value.md

chore: normalize line endings in issue docs

[M] issue-0427-performance-on-np.matmul.md

chore: normalize line endings in issue docs

[M] issue-0428-typo-in-ndarray.tomuliarray-method-name.md

chore: normalize line endings in issue docs

[M] issue-0430-numsharp.backends.unmanaged.unmanagedmemoryblock1-fails-on-mono-on-linux.md

chore: normalize line endings in issue docs

[M] issue-0433-ndarray-exists-in-both-numsharp.core-version-0.20.5.0-and-numsharp.lite-versio.md

chore: normalize line endings in issue docs

[M] issue-0434-accessviolationexception-when-selecting-indexes-using-ndarray-ndarray-and-setti.md

chore: normalize line endings in issue docs

[M] issue-0435-complex-number-support.md

chore: normalize line endings in issue docs

[M] issue-0436-np.searchsorted-error.md

chore: normalize line endings in issue docs

[M] issue-0437-argmin-is-not-the-same-with-numpy.md

chore: normalize line endings in issue docs

[M] issue-0438-how-to-get-the-inverse-of-a-2d-matrix.md

chore: normalize line endings in issue docs

[M] issue-0439-where-is-np.where-function.md

chore: normalize line endings in issue docs

[M] issue-0440-ndarray.tobitmap-has-critical-issue-with-24bpp-vertical-images.md

chore: normalize line endings in issue docs

[M] issue-0443-0.3.0-from-nuget-throwing-notsupportedexception-on-negate-function-call.md

chore: normalize line endings in issue docs

[M] issue-0445-how-can-provide-output-for-np.dot.md

chore: normalize line endings in issue docs

[M] issue-0446-unable-to-use-np.dot-due-to-specified-method-unsupported-error.md

chore: normalize line endings in issue docs

[M] issue-0447-np.sum-is-supported-on-numsharp0.20.5-but-not-on-numsharp0.30.0.md

chore: normalize line endings in issue docs

[M] issue-0448-debug.assert-...-causes-tests-to-stop-the-entire-process.md

chore: normalize line endings in issue docs

[M] issue-0449-isclose-is-not-implemented-and-allclose-test-is-ignored.md

chore: normalize line endings in issue docs

[M] issue-0451-np.argmax-is-slow.md

chore: normalize line endings in issue docs

[M] issue-0452-missing-feature-s-numsharps-np.around-method-is-missing-decimals-parameter.md

chore: normalize line endings in issue docs

[M] issue-0454-ndarray.lstqr-doesnt-work.md

chore: normalize line endings in issue docs

[M] issue-0455-numsharp-does-not-allow-building-with-il2cpp-via-unity.md

chore: normalize line endings in issue docs

[M] issue-0456-silent-catastrophe-in-implicit-casting-singleton-array-to-value-type.md

chore: normalize line endings in issue docs

[M] issue-0461-np.save-incorrectly-saves-system.byte-arrays-as-signed.md

chore: normalize line endings in issue docs

[M] issue-0462-how-to-use-the-repo-to-convert-some-python-code.md

chore: normalize line endings in issue docs

[M] issue-0464-new-api-request-to-port-np.random.triangular.md

chore: normalize line endings in issue docs

[M] issue-0465-how-could-i-transform-between-numsharp.ndarray-with-tensorflow.numpy.ndarray.md

chore: normalize line endings in issue docs

[M] issue-0466-bug-np.random.choice-raise-exception.md

chore: normalize line endings in issue docs

[M] issue-0467-numsharp-and-tensorflow.net-works-on-desktop-but-fails-on-cloud-web-service-.ne.md

chore: normalize line endings in issue docs

[M] issue-0468-np-array.convolve-returning-null.md

chore: normalize line endings in issue docs

[M] issue-0470-numsharp0.30.0-np.random.choice-method-missing-cause-exception.md

chore: normalize line endings in issue docs

[M] issue-0471-unhandled-exception-system.notsupportedexception-specified-method-is-not-suppo.md

chore: normalize line endings in issue docs

[M] issue-0472-how-to-calculate-the-rank-of-a-matrix-with-numsharp.md

chore: normalize line endings in issue docs

[M] issue-0473-bit-shift-and-bit-or.md

chore: normalize line endings in issue docs

[M] issue-0475-tobitmap-fails-if-not-contiguous-because-of-broadcast-mismatch.md

chore: normalize line endings in issue docs

[M] issue-0476-numsharp.core-contains-many-debug.assert-lines.md

chore: normalize line endings in issue docs

[M] issue-0477-different-result-between-numpy-and-numsharp-with-np.matmul-function.md

chore: normalize line endings in issue docs

[M] issue-0479-lacking-outdated-documentation.md

chore: normalize line endings in issue docs

[M] issue-0480-numsharp-equivalent-for-unravel-index.md

chore: normalize line endings in issue docs

[M] issue-0481-normal-disttribution-in-numsharp.md

chore: normalize line endings in issue docs

[M] issue-0483-how-to-convert-list-ndarray-to-ndarray.md

chore: normalize line endings in issue docs

[M] issue-0484-np.load-system.exception.md

chore: normalize line endings in issue docs

[M] issue-0486-slice-assign.md

chore: normalize line endings in issue docs

[M] issue-0487-linspace-to-array-as-type-float-while-other-functions-as-type-double.md

chore: normalize line endings in issue docs

[M] issue-0488-np.random.choice-raised-system.notsupportedexception.md

chore: normalize line endings in issue docs

[M] issue-0490-np.random.choice-with-replace-false-produces-duplicates.md

chore: normalize line endings in issue docs

[M] issue-0491-tobitmap-datatype-mistmatch.md

chore: normalize line endings in issue docs

[M] issue-0492-critical-vulnerability-in-version-5.0.2-of-system.drawing.common.md

chore: normalize line endings in issue docs

[M] issue-0493-numsharp-array-output-in-.net-interactive-notebooks-is-misleading.md

chore: normalize line endings in issue docs

[M] issue-0494-hello-has-scisharp-numsharp-stopped-development-and-maintenance.md

chore: normalize line endings in issue docs

[M] issue-0497-np.linalg.pinv-not-supported.md

chore: normalize line endings in issue docs

[M] issue-0498-is-there-an-example-on-how-to-use-it-with-ironpython.md

chore: normalize line endings in issue docs

[M] issue-0499-possible-typo-tomulidimarray.md

chore: normalize line endings in issue docs

[M] issue-0501-memory-leak.md

chore: normalize line endings in issue docs

[M] issue-0505-np.convolve-return-null-exception.md

chore: normalize line endings in issue docs

[M] issue-0506-cannot-create-an-ndarray-of-shorts.md

chore: normalize line endings in issue docs

[M] issue-0507-np.maximum-error.md

chore: normalize line endings in issue docs

[M] issue-0508-np.hstack-has-diffrent-effect-from-python.md

chore: normalize line endings in issue docs

[M] issue-0509-extremely-poor-performance-on-sum-reduce.md

chore: normalize line endings in issue docs

[M] issue-0510-how-to-save-a-nested-dictionary-with-save-npz.md

chore: normalize line endings in issue docs

[M] issue-0511-does-numsharp-work-in-unity-with-the-il2cpp-backend.md

chore: normalize line endings in issue docs

[M] issue-0512-how-to-change-to-microsoft.ml.onnxruntime.tensors.densetensor-float.md

chore: normalize line endings in issue docs

[M] issue-0514-setitem-for-multiple-ids-not-working.md

chore: normalize line endings in issue docs

[M] issue-0517-error-when-loading-a-.npy-file-containing-a-scalar-value.md

chore: normalize line endings in issue docs

[M] issue-0519-bug-ndarray-filted-array-ori-array-max-prob-conf-threshold.md

chore: normalize line endings in issue docs

📁 docs/neps/

[M] NEP42.md

docs: fix broken documentation URLs (scipy → numpy.org)

[M] NEP55.md

docs: fix broken documentation URLs (scipy → numpy.org)

📁 docs/numpy-alignment/

[A] kernel-ops-test-matrix.md

fix: comprehensive bug fixes from parallel agent battle-testing

📁 docs/numpy/

[A] NUMPY_STRING_TYPES.md

feat: added our own numpy docs

[A] PERFORMANCE_EXECUTION_PATHS.md

feat: added our own numpy docs

[A] PERFORMANCE_NUMSHARP_RECOMMENDATIONS.md

feat: added our own numpy docs

[A] PERFORMANCE_OPTIMIZATION_STRATEGIES.md

feat: added our own numpy docs

[A] PERFORMANCE_SHARED_INFRASTRUCTURE.md

feat: added our own numpy docs

📁 docs/plans/

[A] EXCEPTION_REDESIGN.md

Added plans to unfinished work

[A] UNIFIED_ITERATOR_DESIGN.md

Added plans to unfinished work

[M] numpy-1x-deprecation-findings.md

docs: fix broken documentation URLs (scipy → numpy.org)

📁 src/NumSharp.Core/APIs/

[M] np.array_manipulation.cs

refactor: remove IKernelProvider interface, make ILKernelGenerator static

[A] np.count_nonzero.cs

fix: test assertion bugs and API mismatches in PowerEdgeCaseTests and ClipEdgeCaseTests

[M] np.cs

fix: np.intp now uses nint (native-sized integer) instead of int

[A] np.cumprod.cs

docs: fix broken documentation URLs (scipy → numpy.org) perf: CumProd, MethodInfo cache, Integer Abs/Sign bitwise, Vector512

📁 src/NumSharp.Core/Backends/Default/ArrayManipulation/

[M] Default.Broadcasting.cs

refactor: move broadcast utilities from DefaultEngine to Shape struct

[M] Default.Transpose.cs

fix: multiple NumPy alignment bug fixes

📁 src/NumSharp.Core/Backends/Default/

[M] DefaultEngine.cs

refactor(kernel): integrate IKernelProvider into DefaultEngine refactor(kernel): use DefaultKernelProvider for Enabled/VectorBits checks refactor: remove IKernelProvider interface, make ILKernelGenerator static refactor: remove all Parallel.For usage and switch to single-threaded execution refactor: route np.* and NDArray operations through TensorEngine

📁 src/NumSharp.Core/Backends/Default/Indexing/

[A] Default.BooleanMask.cs

refactor: route np.* and NDArray operations through TensorEngine

[M] Default.NonZero.cs

feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations feat(simd): add SIMD helpers for reductions, fix np.any bug, NumPy NaN handling feat: IL kernel migration for reductions, scans, and math ops fix: kernel day bug fixes and SIMD enhancements fix: test assertion bugs and API mismatches in PowerEdgeCaseTests and ClipEdgeCaseTests refactor(kernel): use DefaultKernelProvider for Enabled/VectorBits checks refactor: remove IKernelProvider interface, make ILKernelGenerator static

📁 src/NumSharp.Core/Backends/Default/Logic/

[M] Default.All.cs

refactor: route np.* and NDArray operations through TensorEngine

[A] Default.Any.cs

refactor: route np.* and NDArray operations through TensorEngine

[M] Default.IsClose.cs

chore: cleanup dead code and fix IsClose/All/Any fix: comprehensive OpenBugs fixes (45 tests fixed, 108→63 failures)

[M] Default.IsFinite.cs

fix: comprehensive OpenBugs fixes (45 tests fixed, 108→63 failures) refactor: proper NumPy-aligned implementations replacing hacks

[A] Default.IsInf.cs

fix: test assertion bugs and API mismatches in PowerEdgeCaseTests and ClipEdgeCaseTests

[M] Default.IsNan.cs

fix: comprehensive OpenBugs fixes (45 tests fixed, 108→63 failures) refactor: proper NumPy-aligned implementations replacing hacks

📁 src/NumSharp.Core/Backends/Default/Math/Add/

[D] Default.Add.Boolean.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Add.Byte.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Add.Char.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Add.Decimal.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Add.Double.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Add.Int16.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Add.Int32.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Add.Int64.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Add.Single.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Add.UInt16.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Add.UInt32.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Add.UInt64.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

📁 src/NumSharp.Core/Backends/Default/Math/BLAS/

[M] Default.Dot.NDMD.cs

feat: complete IL kernel migration batch 2 - Dot.NDMD, CumSum axis, Shift, Var/Std SIMD fix: IL kernel battle-tested fixes for shift overflow and Dot non-contiguous arrays fix: multiple NumPy alignment bug fixes

[M] Default.Dot.cs

fix: implement np.dot(1D, 2D) - treats 1D as row vector

[M] Default.MatMul.2D2D.cs

feat: SIMD-optimized MatMul with 35-100x speedup over scalar path feat: SIMD-optimized MatMul with cache blocking (no parallelization) feat: cache-blocked SIMD MatMul achieving 14-17 GFLOPS fix: IL kernel battle-tested fixes for shift overflow and Dot non-contiguous arrays

[M] Default.MatMul.cs

fix: np.matmul broadcasting crash with >2D arrays

📁 src/NumSharp.Core/Backends/Default/Math/

[M] Default.ACos.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[M] Default.ASin.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[M] Default.ATan.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[M] Default.ATan2.cs

feat(api): complete kernel API audit with NumPy 2.x alignment feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations feat: complete IL kernel migration - ATan2, ClipNDArray, NaN reductions refactor: replace DecimalMath dependency with internal implementation refactor: route np.* and NDArray operations through TensorEngine

[M] Default.Abs.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code fix: multiple NumPy alignment bug fixes

[M] Default.Add.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[A] Default.Cbrt.cs

feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations

[M] Default.Ceil.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[M] Default.Clip.cs

feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations feat: IL kernel migration for reductions, scans, and math ops fix: test assertion bugs and API mismatches in PowerEdgeCaseTests and ClipEdgeCaseTests

[M] Default.ClipNDArray.cs

feat: complete IL kernel migration - ATan2, ClipNDArray, NaN reductions fix: test assertion bugs and API mismatches in PowerEdgeCaseTests and ClipEdgeCaseTests

[M] Default.Cos.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[M] Default.Cosh.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[A] Default.Deg2Rad.cs

feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations

[M] Default.Divide.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[M] Default.Exp.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[M] Default.Exp2.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[M] Default.Expm1.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[M] Default.Floor.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[A] Default.FloorDivide.cs

feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations

[A] Default.Invert.cs

feat(api): complete kernel API audit with NumPy 2.x alignment feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations

[M] Default.Log.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[M] Default.Log10.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[M] Default.Log1p.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[M] Default.Log2.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[M] Default.Mod.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[M] Default.Modf.cs

feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations feat: IL kernel migration for reductions, scans, and math ops

[M] Default.Multiply.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[M] Default.Negate.cs

feat(api): complete kernel API audit with NumPy 2.x alignment feat: IL Kernel Generator replaces 500K+ lines of generated code

[M] Default.Power.cs

feat(api): complete kernel API audit with NumPy 2.x alignment feat: IL kernel migration for reductions, scans, and math ops fix: multiple NumPy alignment bug fixes

[A] Default.Rad2Deg.cs

feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations

[A] Default.Reciprocal.cs

feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations

[M] Default.Round.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[A] Default.Shift.cs

feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations feat: complete IL kernel migration batch 2 - Dot.NDMD, CumSum axis, Shift, Var/Std SIMD fix: correct shift operations and ATan2 tests

[M] Default.Sign.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code fix: multiple NumPy alignment bug fixes

[M] Default.Sin.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[M] Default.Sinh.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[M] Default.Sqrt.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[A] Default.Square.cs

feat(api): complete kernel API audit with NumPy 2.x alignment feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations

[M] Default.Subtract.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[M] Default.Tan.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[M] Default.Tanh.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[A] Default.Truncate.cs

feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations

[A] DefaultEngine.BinaryOp.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code feat: IL kernel migration for reductions, scans, and math ops refactor(kernel): complete scalar delegate integration via IKernelProvider refactor(kernel): integrate IKernelProvider into DefaultEngine refactor: route np.* and NDArray operations through TensorEngine

[A] DefaultEngine.BitwiseOp.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[A] DefaultEngine.CompareOp.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code refactor(kernel): complete scalar delegate integration via IKernelProvider refactor(kernel): integrate IKernelProvider into DefaultEngine refactor: route np.* and NDArray operations through TensorEngine

[A] DefaultEngine.ReductionOp.cs

Unify ArgMin/ArgMax to IL kernels (NaN/bool) feat(kernel): wire axis reduction SIMD to production + port NumPy tests feat(simd): add SIMD helpers for reductions, fix np.any bug, NumPy NaN handling feat: IL Kernel Generator replaces 500K+ lines of generated code feat: NumPy 2.4.2 alignment investigation with bug fixes refactor(kernel): integrate IKernelProvider into DefaultEngine refactor: route np.* and NDArray operations through TensorEngine

[A] DefaultEngine.UnaryOp.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code fix(unary): preserve Boolean type for LogicalNot operation refactor(kernel): complete scalar delegate integration via IKernelProvider refactor(kernel): integrate IKernelProvider into DefaultEngine refactor: proper NumPy-aligned implementations replacing hacks refactor: route np.* and NDArray operations through TensorEngine

📁 src/NumSharp.Core/Backends/Default/Math/Divide/

[D] Default.Divide.Boolean.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Divide.Byte.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Divide.Char.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Divide.Decimal.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Divide.Double.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Divide.Int16.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Divide.Int32.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Divide.Int64.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Divide.Single.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Divide.UInt16.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Divide.UInt32.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Divide.UInt64.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

📁 src/NumSharp.Core/Backends/Default/Math/Mod/

[D] Default.Mod.Boolean.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Mod.Byte.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Mod.Char.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Mod.Decimal.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Mod.Double.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Mod.Int16.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Mod.Int32.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Mod.Int64.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Mod.Single.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Mod.UInt16.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Mod.UInt32.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Mod.UInt64.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

📁 src/NumSharp.Core/Backends/Default/Math/Multiply/

[D] Default.Multiply.Boolean.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Multiply.Byte.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Multiply.Char.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Multiply.Decimal.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Multiply.Double.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Multiply.Int16.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Multiply.Int32.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Multiply.Int64.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Multiply.Single.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Multiply.UInt16.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Multiply.UInt32.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Multiply.UInt64.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

📁 src/NumSharp.Core/Backends/Default/Math/Reduction/

[M] Default.Reduction.AMax.cs

feat(kernel): wire axis reduction SIMD to production + port NumPy tests feat: IL Kernel Generator replaces 500K+ lines of generated code feat: NumPy 2.4.2 alignment investigation with bug fixes fix: comprehensive bug fixes from parallel agent battle-testing fix: extend keepdims fix to all reduction operations fix: test assertion bugs and API mismatches in PowerEdgeCaseTests and ClipEdgeCaseTests refactor: replace Regen axis reduction templates with IL kernel dispatch

[M] Default.Reduction.AMin.cs

feat(kernel): wire axis reduction SIMD to production + port NumPy tests feat: IL Kernel Generator replaces 500K+ lines of generated code feat: NumPy 2.4.2 alignment investigation with bug fixes fix: comprehensive bug fixes from parallel agent battle-testing fix: extend keepdims fix to all reduction operations fix: test assertion bugs and API mismatches in PowerEdgeCaseTests and ClipEdgeCaseTests refactor: replace Regen axis reduction templates with IL kernel dispatch

[M] Default.Reduction.Add.cs

feat(kernel): wire axis reduction SIMD to production + port NumPy tests feat: IL Kernel Generator replaces 500K+ lines of generated code fix: comprehensive OpenBugs fixes (45 tests fixed, 108→63 failures) fix: comprehensive bug fixes from parallel agent battle-testing fix: keepdims returns correct shape for element-wise reductions fix: sum axis reduction for broadcast arrays + NEP50 test fixes (6 more OpenBugs) fix: test assertion bugs and API mismatches in PowerEdgeCaseTests and ClipEdgeCaseTests refactor: proper NumPy-aligned implementations replacing hacks refactor: replace Regen axis reduction templates with IL kernel dispatch

[M] Default.Reduction.ArgMax.cs

Unify ArgMin/ArgMax to IL kernels (NaN/bool) chore: cleanup dead code and fix IsClose/All/Any feat(api): complete kernel API audit with NumPy 2.x alignment feat(kernel): add SIMD axis reduction kernels and fix NaN edge cases feat(simd): add SIMD helpers for reductions, fix np.any bug, NumPy NaN handling feat: IL Kernel Generator replaces 500K+ lines of generated code feat: IL kernel migration for reductions, scans, and math ops feat: NumPy 2.4.2 alignment investigation with bug fixes fix: comprehensive bug fixes from parallel agent battle-testing

[M] Default.Reduction.ArgMin.cs

Unify ArgMin/ArgMax to IL kernels (NaN/bool) chore: cleanup dead code and fix IsClose/All/Any feat(api): complete kernel API audit with NumPy 2.x alignment feat(kernel): add SIMD axis reduction kernels and fix NaN edge cases feat(simd): add SIMD helpers for reductions, fix np.any bug, NumPy NaN handling feat: IL Kernel Generator replaces 500K+ lines of generated code feat: IL kernel migration for reductions, scans, and math ops feat: NumPy 2.4.2 alignment investigation with bug fixes fix: comprehensive bug fixes from parallel agent battle-testing

[M] Default.Reduction.CumAdd.cs

feat: IL kernel migration for reductions, scans, and math ops feat: complete IL kernel migration batch 2 - Dot.NDMD, CumSum axis, Shift, Var/Std SIMD fix: comprehensive OpenBugs fixes (45 tests fixed, 108→63 failures) fix: comprehensive bug fixes from parallel agent battle-testing fix: empty array handling for std/var, cumsum refactor removing 4K lines Regen fix: resolve 6 OpenBugs (3 fixed, 3 verified already working) refactor: proper NumPy-aligned implementations replacing hacks

[A] Default.Reduction.CumMul.cs

perf: CumProd, MethodInfo cache, Integer Abs/Sign bitwise, Vector512

[M] Default.Reduction.Mean.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code fix: comprehensive bug fixes from parallel agent battle-testing fix: keepdims returns correct shape for element-wise reductions refactor: replace Regen axis reduction templates with IL kernel dispatch

[A] Default.Reduction.Nan.cs

refactor: route np.* and NDArray operations through TensorEngine

[M] Default.Reduction.Product.cs

feat(kernel): wire axis reduction SIMD to production + port NumPy tests feat: IL Kernel Generator replaces 500K+ lines of generated code fix: comprehensive bug fixes from parallel agent battle-testing fix: extend keepdims fix to all reduction operations refactor: replace Regen axis reduction templates with IL kernel dispatch

[M] Default.Reduction.Std.cs

feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations feat: IL kernel migration for reductions, scans, and math ops feat: complete IL kernel migration batch 2 - Dot.NDMD, CumSum axis, Shift, Var/Std SIMD fix: comprehensive bug fixes from parallel agent battle-testing fix: empty array handling for std/var, cumsum refactor removing 4K lines Regen fix: extend keepdims fix to all reduction operations refactor: replace DecimalMath dependency with internal implementation

[M] Default.Reduction.Var.cs

feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations feat: IL kernel migration for reductions, scans, and math ops feat: complete IL kernel migration batch 2 - Dot.NDMD, CumSum axis, Shift, Var/Std SIMD fix: comprehensive bug fixes from parallel agent battle-testing fix: empty array handling for std/var, cumsum refactor removing 4K lines Regen fix: extend keepdims fix to all reduction operations

📁 src/NumSharp.Core/Backends/Default/Math/Subtract/

[D] Default.Subtract.Boolean.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Subtract.Byte.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Subtract.Char.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Subtract.Decimal.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Subtract.Double.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Subtract.Int16.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Subtract.Int32.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Subtract.Int64.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Subtract.Single.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Subtract.UInt16.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Subtract.UInt32.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Subtract.UInt64.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

📁 src/NumSharp.Core/Backends/Default/Math/Templates/

[D] Default.Op.Dot.Boolean.template.cs

chore: cleanup dead code and fix IsClose/All/Any

[D] Default.Op.Dot.template.cs

chore: cleanup dead code and fix IsClose/All/Any

[D] Default.Op.General.template.cs

chore: cleanup dead code and fix IsClose/All/Any

📁 src/NumSharp.Core/Backends/Default/Statistics/

[M] Default.ArgMax.cs

feat(api): complete kernel API audit with NumPy 2.x alignment

[M] Default.ArgMin.cs

feat(api): complete kernel API audit with NumPy 2.x alignment

📁 src/NumSharp.Core/Backends/Iterators/

[M] MultiIterator.cs

refactor: move broadcast utilities from DefaultEngine to Shape struct

[D] NDIterator.template.cs

Delete NDIterator.template.cs

📁 src/NumSharp.Core/Backends/Kernels/

[A] BinaryKernel.cs

feat(kernel): add IKernelProvider abstraction layer feat: IL Kernel Generator replaces 500K+ lines of generated code

[A] ILKernelGenerator.Binary.cs

feat(kernel): add IKernelProvider abstraction layer feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations feat(simd): add SIMD helpers for reductions, fix np.any bug, NumPy NaN handling perf: 4x loop unrolling for SIMD kernels refactor: remove IKernelProvider interface, make ILKernelGenerator static refactor: remove dead code and cleanup IL kernel infrastructure refactor: replace null-forgiving operators with fail-fast exceptions in CachedMethods refactor: split ILKernelGenerator.cs into partial classes

[A] ILKernelGenerator.Clip.cs

feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations feat: IL kernel migration for reductions, scans, and math ops feat: complete IL kernel migration - ATan2, ClipNDArray, NaN reductions fix: test assertion bugs and API mismatches in PowerEdgeCaseTests and ClipEdgeCaseTests perf: CumProd, MethodInfo cache, Integer Abs/Sign bitwise, Vector512 refactor: remove IKernelProvider interface, make ILKernelGenerator static

[A] ILKernelGenerator.Comparison.cs

feat(kernel): add IKernelProvider abstraction layer feat(simd): add SIMD helpers for reductions, fix np.any bug, NumPy NaN handling perf: 4x loop unrolling for SIMD kernels refactor(kernel): complete scalar delegate integration via IKernelProvider refactor: remove IKernelProvider interface, make ILKernelGenerator static refactor: remove dead code and cleanup IL kernel infrastructure refactor: replace null-forgiving operators with fail-fast exceptions in CachedMethods refactor: split ILKernelGenerator.cs into partial classes

[A] ILKernelGenerator.Masking.Boolean.cs

refactor: remove IKernelProvider interface, make ILKernelGenerator static refactor: split ILKernelGenerator.Reduction.Axis.cs and Masking.cs into focused partial classes

[A] ILKernelGenerator.Masking.NaN.cs

refactor: remove IKernelProvider interface, make ILKernelGenerator static refactor: split ILKernelGenerator.Reduction.Axis.cs and Masking.cs into focused partial classes

[A] ILKernelGenerator.Masking.VarStd.cs

perf: CumProd, MethodInfo cache, Integer Abs/Sign bitwise, Vector512 refactor: remove IKernelProvider interface, make ILKernelGenerator static refactor: split ILKernelGenerator.Reduction.Axis.cs and Masking.cs into focused partial classes

[A] ILKernelGenerator.Masking.cs

refactor: remove IKernelProvider interface, make ILKernelGenerator static refactor: split ILKernelGenerator.Reduction.Axis.cs and Masking.cs into focused partial classes refactor: split ILKernelGenerator.Reduction.cs and Unary.cs into focused partial classes

[A] ILKernelGenerator.MatMul.cs

feat: SIMD-optimized MatMul with 35-100x speedup over scalar path feat: SIMD-optimized MatMul with cache blocking (no parallelization) fix: IL MatMul - declare locals before executable code fix: SIMD MatMul IL generation - method lookup and Store argument order refactor: remove IKernelProvider interface, make ILKernelGenerator static refactor: remove dead code and cleanup IL kernel infrastructure refactor: replace null-forgiving operators with fail-fast exceptions in CachedMethods

[A] ILKernelGenerator.MixedType.cs

feat(kernel): add IKernelProvider abstraction layer feat(simd): add SIMD helpers for reductions, fix np.any bug, NumPy NaN handling feat: IL kernel migration for reductions, scans, and math ops feat: complete IL kernel migration - ATan2, ClipNDArray, NaN reductions feat: complete IL kernel migration batch 2 - Dot.NDMD, CumSum axis, Shift, Var/Std SIMD perf: 4x loop unrolling for SIMD kernels refactor: remove IKernelProvider interface, make ILKernelGenerator static refactor: remove dead code and cleanup IL kernel infrastructure refactor: replace null-forgiving operators with fail-fast exceptions in CachedMethods refactor: split ILKernelGenerator.cs into partial classes

[A] ILKernelGenerator.Modf.cs

feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations feat: IL kernel migration for reductions, scans, and math ops perf: CumProd, MethodInfo cache, Integer Abs/Sign bitwise, Vector512 refactor: remove IKernelProvider interface, make ILKernelGenerator static

[A] ILKernelGenerator.Reduction.Arg.cs

Unify ArgMin/ArgMax to IL kernels (NaN/bool) refactor: remove IKernelProvider interface, make ILKernelGenerator static refactor: split ILKernelGenerator.Reduction.cs and Unary.cs into focused partial classes

[A] ILKernelGenerator.Reduction.Axis.Arg.cs

refactor: remove IKernelProvider interface, make ILKernelGenerator static refactor: split ILKernelGenerator.Reduction.Axis.cs and Masking.cs into focused partial classes

[A] ILKernelGenerator.Reduction.Axis.NaN.cs

feat: complete IL kernel migration - ATan2, ClipNDArray, NaN reductions refactor: remove IKernelProvider interface, make ILKernelGenerator static refactor: remove dead code and cleanup IL kernel infrastructure

[A] ILKernelGenerator.Reduction.Axis.Simd.cs

perf: SIMD axis reductions with AVX2 gather and parallel outer loop (#576) refactor: remove IKernelProvider interface, make ILKernelGenerator static refactor: remove all Parallel.For usage and switch to single-threaded execution refactor: split ILKernelGenerator.Reduction.Axis.cs and Masking.cs into focused partial classes

[A] ILKernelGenerator.Reduction.Axis.VarStd.cs

perf: SIMD axis reductions with AVX2 gather and parallel outer loop (#576) refactor: remove IKernelProvider interface, make ILKernelGenerator static refactor: remove all Parallel.For usage and switch to single-threaded execution refactor: split ILKernelGenerator.Reduction.Axis.cs and Masking.cs into focused partial classes

[A] ILKernelGenerator.Reduction.Axis.cs

refactor: remove IKernelProvider interface, make ILKernelGenerator static refactor: remove dead code and cleanup IL kernel infrastructure refactor: split ILKernelGenerator.Reduction.Axis.cs and Masking.cs into focused partial classes refactor: split ILKernelGenerator.Reduction.cs and Unary.cs into focused partial classes

[A] ILKernelGenerator.Reduction.Boolean.cs

refactor: remove IKernelProvider interface, make ILKernelGenerator static refactor: split ILKernelGenerator.Reduction.cs and Unary.cs into focused partial classes

[A] ILKernelGenerator.Reduction.cs

Unify ArgMin/ArgMax to IL kernels (NaN/bool) feat(kernel): add IKernelProvider abstraction layer feat(kernel): add SIMD axis reduction kernels and fix NaN edge cases feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations feat(simd): add SIMD helpers for reductions, fix np.any bug, NumPy NaN handling feat: IL kernel migration for reductions, scans, and math ops feat: complete IL kernel migration batch 2 - Dot.NDMD, CumSum axis, Shift, Var/Std SIMD fix: kernel day bug fixes and SIMD enhancements perf: 4x loop unrolling for SIMD kernels refactor: remove IKernelProvider interface, make ILKernelGenerator static refactor: remove dead code and cleanup IL kernel infrastructure refactor: replace Regen axis reduction templates with IL kernel dispatch refactor: replace null-forgiving operators with fail-fast exceptions in CachedMethods refactor: split ILKernelGenerator.Reduction.cs and Unary.cs into focused partial classes refactor: split ILKernelGenerator.cs into partial classes

[A] ILKernelGenerator.Scalar.cs

refactor: remove IKernelProvider interface, make ILKernelGenerator static refactor: remove dead code and cleanup IL kernel infrastructure refactor: split ILKernelGenerator.Reduction.cs and Unary.cs into focused partial classes

[A] ILKernelGenerator.Scan.cs

feat: IL kernel migration for reductions, scans, and math ops feat: complete IL kernel migration batch 2 - Dot.NDMD, CumSum axis, Shift, Var/Std SIMD perf: CumProd, MethodInfo cache, Integer Abs/Sign bitwise, Vector512 refactor: remove IKernelProvider interface, make ILKernelGenerator static refactor: remove dead code and cleanup IL kernel infrastructure refactor: replace null-forgiving operators with fail-fast exceptions in CachedMethods

[A] ILKernelGenerator.Shift.cs

feat: complete IL kernel migration batch 2 - Dot.NDMD, CumSum axis, Shift, Var/Std SIMD fix: IL kernel battle-tested fixes for shift overflow and Dot non-contiguous arrays refactor: remove IKernelProvider interface, make ILKernelGenerator static refactor: remove redundant BUG 81 references from Shift kernel comments refactor: replace null-forgiving operators with fail-fast exceptions in CachedMethods

[A] ILKernelGenerator.Unary.Decimal.cs

refactor: remove IKernelProvider interface, make ILKernelGenerator static refactor: replace null-forgiving operators with fail-fast exceptions in CachedMethods refactor: split ILKernelGenerator.Reduction.cs and Unary.cs into focused partial classes

[A] ILKernelGenerator.Unary.Math.cs

perf: CumProd, MethodInfo cache, Integer Abs/Sign bitwise, Vector512 refactor: remove IKernelProvider interface, make ILKernelGenerator static refactor: replace null-forgiving operators with fail-fast exceptions in CachedMethods refactor: split ILKernelGenerator.Reduction.cs and Unary.cs into focused partial classes

[A] ILKernelGenerator.Unary.Predicate.cs

refactor: remove IKernelProvider interface, make ILKernelGenerator static refactor: split ILKernelGenerator.Reduction.cs and Unary.cs into focused partial classes

[A] ILKernelGenerator.Unary.Vector.cs

refactor: remove IKernelProvider interface, make ILKernelGenerator static refactor: replace null-forgiving operators with fail-fast exceptions in CachedMethods refactor: split ILKernelGenerator.Reduction.cs and Unary.cs into focused partial classes

[A] ILKernelGenerator.Unary.cs

feat(api): complete kernel API audit with NumPy 2.x alignment feat(kernel): add IKernelProvider abstraction layer feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations feat(simd): add SIMD helpers for reductions, fix np.any bug, NumPy NaN handling fix: kernel day bug fixes and SIMD enhancements perf: 4x loop unrolling for SIMD kernels refactor: proper NumPy-aligned implementations replacing hacks refactor: remove IKernelProvider interface, make ILKernelGenerator static refactor: remove dead code and cleanup IL kernel infrastructure refactor: replace null-forgiving operators with fail-fast exceptions in CachedMethods refactor: split ILKernelGenerator.Reduction.cs and Unary.cs into focused partial classes refactor: split ILKernelGenerator.cs into partial classes

[A] ILKernelGenerator.cs

feat(SIMD): add SIMD scalar paths to IL kernel generator feat(SIMD): dynamic vector width detection for IL kernels feat(kernel): add IKernelProvider abstraction layer feat(kernel): add SIMD axis reduction kernels and fix NaN edge cases feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations feat(simd): add SIMD helpers for reductions, fix np.any bug, NumPy NaN handling feat: IL Kernel Generator replaces 500K+ lines of generated code feat: complete IL kernel migration - ATan2, ClipNDArray, NaN reductions fix: comprehensive OpenBugs fixes (45 tests fixed, 108→63 failures) perf: CumProd, MethodInfo cache, Integer Abs/Sign bitwise, Vector512 refactor(kernel): complete scalar delegate integration via IKernelProvider refactor: remove IKernelProvider interface, make ILKernelGenerator static refactor: remove dead code and cleanup IL kernel infrastructure refactor: replace DecimalMath dependency with internal implementation refactor: replace null-forgiving operators with fail-fast exceptions in CachedMethods refactor: split ILKernelGenerator.cs into partial classes

[A] KernelKey.cs

feat(kernel): add IKernelProvider abstraction layer

[A] KernelOp.cs

feat(api): complete kernel API audit with NumPy 2.x alignment feat(kernel): add IKernelProvider abstraction layer feat: complete IL kernel migration - ATan2, ClipNDArray, NaN reductions perf: CumProd, MethodInfo cache, Integer Abs/Sign bitwise, Vector512 refactor: proper NumPy-aligned implementations replacing hacks

[A] KernelSignatures.cs

feat(kernel): add IKernelProvider abstraction layer refactor: remove IKernelProvider interface, make ILKernelGenerator static

[A] ReductionKernel.cs

feat(kernel): add IKernelProvider abstraction layer feat(simd): add SIMD helpers for reductions, fix np.any bug, NumPy NaN handling feat: IL Kernel Generator replaces 500K+ lines of generated code feat: complete IL kernel migration batch 2 - Dot.NDMD, CumSum axis, Shift, Var/Std SIMD

[A] ScalarKernel.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code refactor(kernel): complete scalar delegate integration via IKernelProvider

[A] SimdMatMul.cs

feat: cache-blocked SIMD MatMul achieving 14-17 GFLOPS perf: full panel packing for MatMul achieving 20+ GFLOPS

[A] SimdThresholds.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[A] StrideDetector.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[A] TypeRules.cs

feat(kernel): add IKernelProvider abstraction layer fix: resolve 6 OpenBugs (3 fixed, 3 verified already working)

📁 src/NumSharp.Core/Backends/

[M] NDArray.cs

refactor: remove IKernelProvider interface, make ILKernelGenerator static

[M] NPTypeCode.cs

fix: multiple NumPy alignment bug fixes

[M] TensorEngine.cs

feat(api): complete kernel API audit with NumPy 2.x alignment feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations feat: IL Kernel Generator replaces 500K+ lines of generated code fix: test assertion bugs and API mismatches in PowerEdgeCaseTests and ClipEdgeCaseTests perf: CumProd, MethodInfo cache, Integer Abs/Sign bitwise, Vector512 refactor: route np.* and NDArray operations through TensorEngine

📁 src/NumSharp.Core/Backends/Unmanaged/Pooling/

[M] StackedMemoryPool.cs

refactor: modernize allocation with NativeMemory API (#528)

📁 src/NumSharp.Core/Backends/Unmanaged/

[M] UnmanagedMemoryBlock1.cs`

refactor: modernize allocation with NativeMemory API (#528) refactor: rename AllocationType.AllocHGlobal to Native

[M] UnmanagedStorage.Getters.cs

fix: np.matmul broadcasting crash with >2D arrays

[M] UnmanagedStorage.Slicing.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code fix: NumPy 2.x alignment for array creation and indexing refactor: proper NumPy-aligned implementations replacing hacks

📁 src/NumSharp.Core/Casting/Implicit/

[M] NdArray.Implicit.ValueTypes.cs

fix: resolve 6 OpenBugs (3 fixed, 3 verified already working)

📁 src/NumSharp.Core/Casting/

[M] NdArray.ToString.cs

fix: NumPy 2.x alignment for array creation and indexing

📁 src/NumSharp.Core/Creation/

[M] np.arange.cs

fix: NumPy 2.x alignment for array creation and indexing

[M] np.are_broadcastable.cs

refactor: move broadcast utilities from DefaultEngine to Shape struct

[M] np.array.cs

refactor: remove all Parallel.For usage and switch to single-threaded execution

[M] np.broadcast.cs

refactor: move broadcast utilities from DefaultEngine to Shape struct

[M] np.broadcast_arrays.cs

refactor: move broadcast utilities from DefaultEngine to Shape struct

[M] np.broadcast_to.cs

refactor: move broadcast utilities from DefaultEngine to Shape struct

[M] np.full.cs

fix: NumPy 2.x alignment for array creation and indexing

[M] np.linspace.cs

fix: NumPy 2.x alignment for array creation and indexing fix: multiple NumPy alignment bug fixes

📁 src/NumSharp.Core/Generics/

[A] NDArray1.Operators.cs`

feat: IL Kernel Generator replaces 500K+ lines of generated code

📁 src/NumSharp.Core/Logic/

[M] np.all.cs

feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations feat(simd): add SIMD helpers for reductions, fix np.any bug, NumPy NaN handling fix: kernel day bug fixes and SIMD enhancements refactor(kernel): use DefaultKernelProvider for Enabled/VectorBits checks refactor: route np.* and NDArray operations through TensorEngine

[M] np.any.cs

feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations feat(simd): add SIMD helpers for reductions, fix np.any bug, NumPy NaN handling fix: kernel day bug fixes and SIMD enhancements refactor(kernel): use DefaultKernelProvider for Enabled/VectorBits checks refactor: route np.* and NDArray operations through TensorEngine

[A] np.comparison.cs

feat(api): complete kernel API audit with NumPy 2.x alignment

[M] np.is.cs

fix: test assertion bugs and API mismatches in PowerEdgeCaseTests and ClipEdgeCaseTests

[A] np.logical.cs

feat(api): complete kernel API audit with NumPy 2.x alignment

📁 src/NumSharp.Core/Manipulation/

[M] NDArray.unique.cs

fix(unique): sort unique values to match NumPy behavior fix: comprehensive bug fixes from parallel agent battle-testing

[M] np.repeat.cs

fix: comprehensive OpenBugs fixes (45 tests fixed, 108→63 failures) refactor: proper NumPy-aligned implementations replacing hacks

📁 src/NumSharp.Core/Math/

[M] NDArray.negative.cs

fix: multiple NumPy alignment bug fixes

[M] NDArray.positive.cs

fix: multiple NumPy alignment bug fixes

[M] NdArray.Convolve.cs

fix: comprehensive OpenBugs fixes (45 tests fixed, 108→63 failures) fix: multiple NumPy alignment bug fixes

[M] np.absolute.cs

feat(api): complete kernel API audit with NumPy 2.x alignment

[A] np.cbrt.cs

feat(api): complete kernel API audit with NumPy 2.x alignment feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations

[M] np.ceil.cs

feat(api): complete kernel API audit with NumPy 2.x alignment

[M] np.clip.cs

feat(api): complete kernel API audit with NumPy 2.x alignment

[M] np.cos.cs

feat(api): complete kernel API audit with NumPy 2.x alignment

[A] np.deg2rad.cs

feat(api): complete kernel API audit with NumPy 2.x alignment feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations

[M] np.floor.cs

feat(api): complete kernel API audit with NumPy 2.x alignment

[A] np.floor_divide.cs

feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations

[A] np.invert.cs

docs: fix broken documentation URLs (scipy → numpy.org) feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations

[A] np.left_shift.cs

feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations

[M] np.log.cs

feat(api): complete kernel API audit with NumPy 2.x alignment

[M] np.maximum.cs

feat(api): complete kernel API audit with NumPy 2.x alignment

[M] np.minimum.cs

feat(api): complete kernel API audit with NumPy 2.x alignment

[M] np.modf.cs

feat(api): complete kernel API audit with NumPy 2.x alignment

[A] np.nanprod.cs

feat(kernel): add SIMD axis reduction kernels and fix NaN edge cases feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations feat: complete IL kernel migration - ATan2, ClipNDArray, NaN reductions refactor: route np.* and NDArray operations through TensorEngine

[A] np.nansum.cs

feat(kernel): add SIMD axis reduction kernels and fix NaN edge cases feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations feat: complete IL kernel migration - ATan2, ClipNDArray, NaN reductions refactor: route np.* and NDArray operations through TensorEngine

[M] np.power.cs

docs: fix broken documentation URLs (scipy → numpy.org) feat(api): complete kernel API audit with NumPy 2.x alignment feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations

[A] np.rad2deg.cs

feat(api): complete kernel API audit with NumPy 2.x alignment feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations

[A] np.reciprocal.cs

feat(api): complete kernel API audit with NumPy 2.x alignment feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations

[A] np.right_shift.cs

feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations

[M] np.round.cs

feat(api): complete kernel API audit with NumPy 2.x alignment

[M] np.sign.cs

feat(api): complete kernel API audit with NumPy 2.x alignment

[M] np.sin.cs

feat(api): complete kernel API audit with NumPy 2.x alignment

[M] np.sqrt.cs

feat(api): complete kernel API audit with NumPy 2.x alignment

[M] np.tan.cs

feat(api): complete kernel API audit with NumPy 2.x alignment

[A] np.trunc.cs

feat(api): complete kernel API audit with NumPy 2.x alignment feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations

📁 src/NumSharp.Core/

[M] NumSharp.Core.csproj

chore: remove obsolete template file references from csproj feat(SIMD): dynamic vector width detection for IL kernels feat(simd): add SIMD helpers for reductions, fix np.any bug, NumPy NaN handling

📁 src/NumSharp.Core/Operations/Elementwise/Equals/

[D] Default.Equals.Boolean.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Equals.Byte.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Equals.Char.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Equals.Decimal.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Equals.Double.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Equals.Int16.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Equals.Int32.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Equals.Int64.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Equals.Single.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Equals.UInt16.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Equals.UInt32.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Equals.UInt64.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[D] Default.Equals.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

📁 src/NumSharp.Core/Operations/Elementwise/

[M] NDArray.AND.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[M] NDArray.Equals.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code fix: comprehensive bug fixes from parallel agent battle-testing

[M] NDArray.Greater.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code fix: comprehensive bug fixes from parallel agent battle-testing

[M] NDArray.Lower.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code fix: comprehensive bug fixes from parallel agent battle-testing

[M] NDArray.NotEquals.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code fix: comprehensive bug fixes from parallel agent battle-testing

[M] NDArray.OR.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

📁 src/NumSharp.Core/Operations/Elementwise/Templates/

[D] Default.Op.Boolean.template.cs

Unify ArgMin/ArgMax to IL kernels (NaN/bool) refactor: move broadcast utilities from DefaultEngine to Shape struct

[D] Default.Op.Equals.template.cs

Unify ArgMin/ArgMax to IL kernels (NaN/bool) refactor: move broadcast utilities from DefaultEngine to Shape struct

📁 src/NumSharp.Core/RandomSampling/

[M] np.random.bernoulli.cs

refactor(random): align parameter names and docs with NumPy API

[M] np.random.beta.cs

feat(random): add Shape overloads for randn, normal, standard_normal refactor(random): align parameter names and docs with NumPy API

[M] np.random.binomial.cs

refactor(random): align parameter names and docs with NumPy API

[M] np.random.chisquare.cs

refactor(random): align parameter names and docs with NumPy API

[M] np.random.choice.cs

refactor(random): align parameter names and docs with NumPy API

[M] np.random.exponential.cs

refactor(random): align parameter names and docs with NumPy API

[M] np.random.gamma.cs

refactor(random): align parameter names and docs with NumPy API

[M] np.random.geometric.cs

refactor(random): align parameter names and docs with NumPy API

[M] np.random.lognormal.cs

refactor(random): align parameter names and docs with NumPy API

[M] np.random.permutation.cs

refactor(random): align parameter names and docs with NumPy API

[M] np.random.poisson.cs

refactor(random): align parameter names and docs with NumPy API

[M] np.random.rand.cs

feat(random): add Shape overloads for randn, normal, standard_normal fix(random): fix standard_normal typo and add random() alias refactor(random): align parameter names and docs with NumPy API

[M] np.random.randint.cs

refactor(random): align parameter names and docs with NumPy API

[M] np.random.randn.cs

feat(random): add Shape overloads for randn, normal, standard_normal fix(random): fix standard_normal typo and add random() alias refactor(random): align parameter names and docs with NumPy API refactor(random): remove stardard_normal backwards compat alias

[M] np.random.shuffle.cs

feat(shuffle): add axis parameter and fix NumPy alignment (closes #582) fix(shuffle): align with NumPy legacy API (no axis parameter)

[M] np.random.uniform.cs

refactor(random): align parameter names and docs with NumPy API

📁 src/NumSharp.Core/Selection/

[M] NDArray.Indexing.Masking.cs

feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations feat(simd): add SIMD helpers for reductions, fix np.any bug, NumPy NaN handling fix: NumPy 2.x alignment for array creation and indexing fix: comprehensive OpenBugs fixes (45 tests fixed, 108→63 failures) fix: test assertion bugs and API mismatches in PowerEdgeCaseTests and ClipEdgeCaseTests refactor(kernel): use DefaultKernelProvider for Enabled/VectorBits checks refactor: remove IKernelProvider interface, make ILKernelGenerator static refactor: route np.* and NDArray operations through TensorEngine

[M] NDArray.Indexing.Selection.Getter.cs

fix: kernel day bug fixes and SIMD enhancements

[M] NDArray.Indexing.Selection.Setter.cs

fix: comprehensive OpenBugs fixes (45 tests fixed, 108→63 failures)

📁 src/NumSharp.Core/Sorting_Searching_Counting/

[M] ndarray.argsort.cs

fix: comprehensive OpenBugs fixes (45 tests fixed, 108→63 failures) fix: test assertion bugs and API mismatches in PowerEdgeCaseTests and ClipEdgeCaseTests

[M] np.amax.cs

fix: extend keepdims fix to all reduction operations

[M] np.argmax.cs

feat(api): complete kernel API audit with NumPy 2.x alignment feat: NumPy 2.4.2 alignment investigation with bug fixes

[M] np.argsort.cs

fix: comprehensive OpenBugs fixes (45 tests fixed, 108→63 failures)

[M] np.min.cs

fix: extend keepdims fix to all reduction operations

[A] np.nanmax.cs

feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations feat: complete IL kernel migration - ATan2, ClipNDArray, NaN reductions refactor: route np.* and NDArray operations through TensorEngine

[A] np.nanmin.cs

feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations feat: complete IL kernel migration - ATan2, ClipNDArray, NaN reductions refactor: route np.* and NDArray operations through TensorEngine

[M] np.searchsorted.cs

fix(searchsorted): use type-agnostic value extraction for all dtypes fix: medium severity bug fixes (BUG-12, BUG-16, BUG-17)

📁 src/NumSharp.Core/Statistics/

[M] NDArray.argmax.cs

feat(api): complete kernel API audit with NumPy 2.x alignment feat(simd): add SIMD helpers for reductions, fix np.any bug, NumPy NaN handling feat: NumPy 2.4.2 alignment investigation with bug fixes

[M] NDArray.argmin.cs

feat(api): complete kernel API audit with NumPy 2.x alignment feat(simd): add SIMD helpers for reductions, fix np.any bug, NumPy NaN handling feat: NumPy 2.4.2 alignment investigation with bug fixes

[A] np.nanmean.cs

feat: NumPy 2.4.2 alignment investigation with bug fixes

[A] np.nanstd.cs

feat: NumPy 2.4.2 alignment investigation with bug fixes

[A] np.nanvar.cs

feat: NumPy 2.4.2 alignment investigation with bug fixes

[M] np.std.cs

feat: IL kernel migration for reductions, scans, and math ops

[M] np.var.cs

feat: IL kernel migration for reductions, scans, and math ops

📁 src/NumSharp.Core/Utilities/

[M] ArrayConvert.cs

refactor: remove all Parallel.For usage and switch to single-threaded execution

[M] Converts.Native.cs

fix: medium severity bug fixes (BUG-12, BUG-16, BUG-17) test: comprehensive edge case battle-testing for recent fixes

[D] DecimalEx.cs

refactor: replace DecimalMath dependency with internal implementation

[A] DecimalMath.cs

refactor: replace DecimalMath dependency with internal implementation

📁 src/NumSharp.Core/View/

[A] Shape.Broadcasting.cs

docs: fix broken documentation URLs (scipy → numpy.org) refactor: move broadcast utilities from DefaultEngine to Shape struct

[M] Shape.cs

docs: fix misleading ALIGNED flag comment

📁 src/

[A] numpy

feat(simd): add SIMD helpers for reductions, fix np.any bug, NumPy NaN handling

📁 test/NumSharp.Benchmark/

[A] GlobalUsings.cs

fix: test assertion bugs and API mismatches in PowerEdgeCaseTests and ClipEdgeCaseTests

📁 test/NumSharp.UnitTest/APIs/

[A] CountNonzeroTests.cs

fix: test assertion bugs and API mismatches in PowerEdgeCaseTests and ClipEdgeCaseTests

[A] np_searchsorted_edge_cases.cs

test: comprehensive edge case battle-testing for recent fixes

📁 test/NumSharp.UnitTest/Backends/Kernels/

[A] ArgMaxArgMinComprehensiveTests.cs

feat: IL kernel migration for reductions, scans, and math ops

[A] ArgMaxNaNTests.cs

feat: NumPy 2.4.2 alignment investigation with bug fixes fix: kernel day bug fixes and SIMD enhancements test: add comprehensive tests for SIMD optimizations and NumPy compatibility

[A] AxisReductionBenchmarkTests.cs

perf: SIMD axis reductions with AVX2 gather and parallel outer loop (#576)

[A] AxisReductionEdgeCaseTests.cs

test: comprehensive edge case battle-testing for recent fixes

[A] AxisReductionMemoryTests.cs

fix: comprehensive bug fixes from parallel agent battle-testing

[A] AxisReductionSimdTests.cs

feat(kernel): add SIMD axis reduction kernels and fix NaN edge cases

[A] BattleProofTests.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[A] BinaryOpTests.cs

feat(api): complete kernel API audit with NumPy 2.x alignment feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations feat: IL Kernel Generator replaces 500K+ lines of generated code feat: complete IL kernel migration - ATan2, ClipNDArray, NaN reductions fix: correct shift operations and ATan2 tests

[A] BitwiseOpTests.cs

Unify ArgMin/ArgMax to IL kernels (NaN/bool) feat: IL Kernel Generator replaces 500K+ lines of generated code

[A] ComparisonOpTests.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[A] CumSumComprehensiveTests.cs

feat: IL kernel migration for reductions, scans, and math ops

[A] DtypeCoverageTests.cs

feat(api): complete kernel API audit with NumPy 2.x alignment fix: sum axis reduction for broadcast arrays + NEP50 test fixes (6 more OpenBugs)

[A] DtypePromotionTests.cs

fix: test assertion bugs and API mismatches in PowerEdgeCaseTests and ClipEdgeCaseTests

[A] EdgeCaseTests.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code fix: comprehensive OpenBugs fixes (45 tests fixed, 108→63 failures) fix: empty array handling for std/var, cumsum refactor removing 4K lines Regen

[A] EmptyArrayReductionTests.cs

feat: NumPy 2.4.2 alignment investigation with bug fixes

[A] EmptyAxisReductionTests.cs

fix: comprehensive bug fixes from parallel agent battle-testing

[A] IndexingEdgeCaseTests.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[A] KernelMisalignmentTests.cs

feat(api): complete kernel API audit with NumPy 2.x alignment test: add [Misaligned] tests documenting NumPy behavioral differences

[A] LinearAlgebraTests.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code fix: implement np.dot(1D, 2D) - treats 1D as row vector test: add comprehensive np.dot(1D, 2D) battle tests test: update tests for bug fixes and NEP50 alignment

[A] ManipulationEdgeCaseTests.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[A] NanReductionTests.cs

feat(kernel): add SIMD axis reduction kernels and fix NaN edge cases feat: complete IL kernel migration - ATan2, ClipNDArray, NaN reductions

[A] NonContiguousTests.cs

feat(api): complete kernel API audit with NumPy 2.x alignment

[A] NumpyAlignmentBugTests.cs

feat: NumPy 2.4.2 alignment investigation with bug fixes fix: comprehensive OpenBugs fixes (45 tests fixed, 108→63 failures) fix: implement np.dot(1D, 2D) - treats 1D as row vector fix: kernel day bug fixes and SIMD enhancements fix: sum axis reduction for broadcast arrays + NEP50 test fixes (6 more OpenBugs) refactor: proper NumPy-aligned implementations replacing hacks test: update tests for bug fixes and NEP50 alignment

[A] ReductionOpTests.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code fix: test assertion bugs and API mismatches in PowerEdgeCaseTests and ClipEdgeCaseTests

[A] ShiftOpTests.cs

feat(kernel): complete ILKernelGenerator coverage with SIMD optimizations

[A] SimdOptimizationTests.cs

fix: comprehensive OpenBugs fixes (45 tests fixed, 108→63 failures) test: add comprehensive tests for SIMD optimizations and NumPy compatibility

[A] SlicedArrayOpTests.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[A] TypePromotionTests.cs

feat: NumPy 2.4.2 alignment investigation with bug fixes

[A] UnaryOpTests.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[A] UnarySpecialValuesTests.cs

fix: test assertion bugs and API mismatches in PowerEdgeCaseTests and ClipEdgeCaseTests

[A] VarStdComprehensiveTests.cs

feat: IL kernel migration for reductions, scans, and math ops

📁 test/NumSharp.UnitTest/Backends/

[M] NDArray.Base.MemoryLeakTest.cs

refactor: remove Parallel.For from MemoryLeakTest

[M] NDArray.Base.Test.cs

fix: test assertion bugs and API mismatches in PowerEdgeCaseTests and ClipEdgeCaseTests

📁 test/NumSharp.UnitTest/Backends/Unmanaged/Math/Reduction/

[M] ReduceAddTests.cs

fix: extend keepdims fix to all reduction operations fix: keepdims returns correct shape for element-wise reductions test: update tests for bug fixes and NEP50 alignment

📁 test/NumSharp.UnitTest/Backends/Unmanaged/Math/

[M] np.prod.tests.cs

test: move Bug 75/76 fix tests to proper test files

[M] np_divide_tests.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

📁 test/NumSharp.UnitTest/Backends/Unmanaged/

[M] UmanagedArrayTests.cs

feat(SIMD): dynamic vector width detection for IL kernels

📁 test/NumSharp.UnitTest/Casting/

[A] ScalarConversionTests.cs

fix: resolve 6 OpenBugs (3 fixed, 3 verified already working)

📁 test/NumSharp.UnitTest/Creation/

[M] NpBroadcastFromNumPyTests.cs

test: update tests for bug fixes and NEP50 alignment

[M] np.concatenate.Test.cs

fix: comprehensive OpenBugs fixes (45 tests fixed, 108→63 failures)

📁 test/NumSharp.UnitTest/Extensions/

[M] ndarray.argsort.Test.cs

fix: comprehensive OpenBugs fixes (45 tests fixed, 108→63 failures)

📁 test/NumSharp.UnitTest/

[A] GlobalUsings.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

📁 test/NumSharp.UnitTest/Indexing/

[A] NonzeroTests.cs

fix: test assertion bugs and API mismatches in PowerEdgeCaseTests and ClipEdgeCaseTests

[A] np_nonzero_edge_cases.cs

test: add comprehensive tests for SIMD optimizations and NumPy compatibility

[A] np_nonzero_strided_tests.cs

feat: IL kernel migration for reductions, scans, and math ops

📁 test/NumSharp.UnitTest/LinearAlgebra/

[A] np.dot.BattleTest.cs

test(dot): add comprehensive battle tests for np.dot NumPy alignment

[M] np.dot.Test.cs

feat: complete IL kernel migration batch 2 - Dot.NDMD, CumSum axis, Shift, Var/Std SIMD

[A] np.matmul.BattleTest.cs

test(linalg): add battle tests for np.matmul and np.outer

[M] np.matmul.Test.cs

fix: np.matmul broadcasting crash with >2D arrays

[A] np.outer.BattleTest.cs

test(linalg): add battle tests for np.matmul and np.outer

📁 test/NumSharp.UnitTest/Logic/

[M] NEP50.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

[A] TypePromotionTests.cs

fix: test assertion bugs and API mismatches in PowerEdgeCaseTests and ClipEdgeCaseTests

[M] np.allclose.Test.cs

chore: cleanup dead code and fix IsClose/All/Any

[M] np.any.Test.cs

chore: cleanup dead code and fix IsClose/All/Any

[A] np.comparison.Test.cs

feat(api): complete kernel API audit with NumPy 2.x alignment fix: test assertion bugs and API mismatches in PowerEdgeCaseTests and ClipEdgeCaseTests

[M] np.isclose.Test.cs

chore: cleanup dead code and fix IsClose/All/Any

[M] np.isfinite.Test.cs

refactor: proper NumPy-aligned implementations replacing hacks

[A] np.isinf.Test.cs

fix: test assertion bugs and API mismatches in PowerEdgeCaseTests and ClipEdgeCaseTests

[M] np.isnan.Test.cs

refactor: proper NumPy-aligned implementations replacing hacks

[M] np_all_axis_Test.cs

chore: cleanup dead code and fix IsClose/All/Any

📁 test/NumSharp.UnitTest/Manipulation/

[A] NDArray.astype.Truncation.Test.cs

test: comprehensive edge case battle-testing for recent fixes

[M] NDArray.flat.Test.cs

fix(tests): correct IsBroadcasted expectations in broadcast_arrays tests

[M] NDArray.unique.Test.cs

fix(unique): sort unique values to match NumPy behavior

[A] ReshapeScalarTests.cs

fix: resolve 6 OpenBugs (3 fixed, 3 verified already working)

[M] np.repeat.Test.cs

refactor: proper NumPy-aligned implementations replacing hacks

[A] np.unique.EdgeCases.Test.cs

fix: comprehensive bug fixes from parallel agent battle-testing test: comprehensive edge case battle-testing for recent fixes

📁 test/NumSharp.UnitTest/Math/

[M] NDArray.Absolute.Test.cs

test: update tests for bug fixes and NEP50 alignment

[A] NDArray.cumprod.Test.cs

perf: CumProd, MethodInfo cache, Integer Abs/Sign bitwise, Vector512

[M] NDArray.cumsum.Test.cs

test: move Bug 75/76 fix tests to proper test files test: update tests for bug fixes and NEP50 alignment

[M] NDArray.negative.Test.cs

test: update tests for bug fixes and NEP50 alignment

[M] NDArray.positive.Test.cs

test: update tests for bug fixes and NEP50 alignment

[M] NdArray.Convolve.Test.cs

fix: comprehensive OpenBugs fixes (45 tests fixed, 108→63 failures)

[A] SignDtypeTests.cs

fix: resolve 6 OpenBugs (3 fixed, 3 verified already working)

[A] np.minimum.Test.cs

fix: comprehensive OpenBugs fixes (45 tests fixed, 108→63 failures)

📁 test/NumSharp.UnitTest/NumPyPortedTests/

[A] ArgMaxArgMinEdgeCaseTests.cs

feat: IL kernel migration for reductions, scans, and math ops fix: correct assertion syntax and API usage in edge case tests

[A] ClipEdgeCaseTests.cs

feat: IL kernel migration for reductions, scans, and math ops fix: correct assertion syntax and API usage in edge case tests fix: test assertion bugs and API mismatches in PowerEdgeCaseTests and ClipEdgeCaseTests

[A] ClipNDArrayTests.cs

feat: complete IL kernel migration - ATan2, ClipNDArray, NaN reductions

[A] CumSumEdgeCaseTests.cs

feat: IL kernel migration for reductions, scans, and math ops fix: correct assertion syntax and API usage in edge case tests

[A] ModfEdgeCaseTests.cs

feat: IL kernel migration for reductions, scans, and math ops fix: correct assertion syntax and API usage in edge case tests

[A] NonzeroEdgeCaseTests.cs

feat: IL kernel migration for reductions, scans, and math ops fix: correct assertion syntax and API usage in edge case tests

[A] PowerEdgeCaseTests.cs

feat: IL kernel migration for reductions, scans, and math ops fix: correct assertion syntax and API usage in edge case tests fix: remove async from non-async test methods in PowerEdgeCaseTests fix: test assertion bugs and API mismatches in PowerEdgeCaseTests and ClipEdgeCaseTests

[A] VarStdEdgeCaseTests.cs

feat: IL kernel migration for reductions, scans, and math ops fix: correct assertion syntax and API usage in edge case tests

📁 test/NumSharp.UnitTest/

[M] OpenBugs.ApiAudit.cs

OpenBugs.ApiAudit.cs: test updates for int64 changes feat: NumPy 2.4.2 alignment investigation with bug fixes fix: empty array handling for std/var, cumsum refactor removing 4K lines Regen fix: resolve 6 OpenBugs (3 fixed, 3 verified already working) test: move Bug 75/76 fix tests to proper test files test: update tests for bug fixes and NEP50 alignment

[A] OpenBugs.ILKernelBattle.cs

fix: IL kernel battle-tested fixes for shift overflow and Dot non-contiguous arrays fix: implement np.dot(1D, 2D) - treats 1D as row vector

[M] OpenBugs.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code fix(unique): sort unique values to match NumPy behavior fix: comprehensive OpenBugs fixes (45 tests fixed, 108→63 failures) fix: sum axis reduction for broadcast arrays + NEP50 test fixes (6 more OpenBugs)

📁 test/NumSharp.UnitTest/Operations/

[A] EmptyArrayComparisonTests.cs

fix: comprehensive bug fixes from parallel agent battle-testing

📁 test/NumSharp.UnitTest/Random/

[M] np.random.choice.Test.cs

refactor(random): align parameter names and docs with NumPy API

[M] np.random.seed.Test.cs

refactor(random): align parameter names and docs with NumPy API

📁 test/NumSharp.UnitTest/RandomSampling/

[A] np.random.shuffle.NumPyAligned.Test.cs

fix(shuffle): align with NumPy legacy API (no axis parameter)

[M] np.random.shuffle.Test.cs

feat(shuffle): add axis parameter and fix NumPy alignment (closes #582)

📁 test/NumSharp.UnitTest/Selection/

[A] BooleanIndexing.BattleTests.cs

test: add comprehensive boolean indexing battle tests (76 tests)

[A] BooleanMaskingTests.cs

test: add comprehensive tests for SIMD optimizations and NumPy compatibility

[M] NDArray.AMax.Test.cs

feat: NumPy 2.4.2 alignment investigation with bug fixes

[M] NDArray.Indexing.Test.cs

fix(tests): correct IsBroadcasted expectations in broadcast_arrays tests

📁 test/NumSharp.UnitTest/Sorting/

[A] ArgsortNaNTests.cs

fix: test assertion bugs and API mismatches in PowerEdgeCaseTests and ClipEdgeCaseTests

📁 test/NumSharp.UnitTest/Statistics/

[M] NdArray.Mean.Test.cs

fix: keepdims returns correct shape for element-wise reductions

📁 test/NumSharp.UnitTest/Utilities/

[M] FluentExtension.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code

📁 test/NumSharp.UnitTest/View/

[M] Shape.OffsetParity.Tests.cs

feat: IL Kernel Generator replaces 500K+ lines of generated code