luisquintanilla opened a new pull request, #379: URL: https://github.com/apache/arrow-dotnet/pull/379
## What's Changed Adds a new `Apache.Arrow.Compute` project with SIMD-accelerated Sum/Min/Max/Mean aggregation kernels for `PrimitiveArray<T>`, backed by `System.Numerics.Tensors` (`TensorPrimitives`). When `NullCount == 0` the kernels dispatch to `TensorPrimitives` for one vectorized pass over the contiguous `Values` span; with nulls they fall back to a validity-aware scalar loop. The project is AOT-compatible with scalar fallbacks for older target frameworks, and pulls in `System.Numerics.Tensors` via `Directory.Packages.props`. Includes an xUnit test project (`Apache.Arrow.Compute.Tests`, 11 tests) covering int32, int64, float, and double, with and without nulls, and the empty-array case. The new projects are added to `Apache.Arrow.sln` and the test to `Apache.Arrow.Tests.slnf`. This implements the aggregation-kernel portion of #375; the elementwise kernels and `Tensor<T>` wrapper described there remain as follow-ups. Part of #375. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
