Hacktoberfest 2026: the issues maintainers tagged for October, open and beginner-friendly. Browse Hacktoberfest issues

Tensor<T> slice pinning starts at the parent array instead of the slice

Closed Beginner friendly
  • Is something wrong?
#134,691 2 comments 0 reactions 0 assignees View on GitHub

Maintainers usually reply within 1 day

Nobody has claimed this yet.

Assessment

Difficulty
2/5
Estimated time
1-3 hours
Newbie friendliness
78/100
Issue type
Bug
Clarity
Clearly specified
Activity status
Active
Tech stack
csharp
Domain
backend

Research direction

Start with src/libraries/System.Numerics.Tensors/src/System/Numerics/Tensors/netcore/Tensor_1.cs, focusing on GetPinnableReference() and GetPinnedHandle() for sliced tensors. Run the TensorSlicePinning.cs file-based reproduction first, then verify that pinning begins at the logical slice offset and returns [11,12,21,22] with exit code 0.

Written by the indexing model from the issue text.

Description

area-System.Numerics.Tensors untriaged

Description

A contiguous Tensor<long> slice returns different values through its pinning APIs than through its indexer or tensor span. With System.Numerics.Tensors 10.0.9, GetPinnableReference() and GetPinnedHandle() start at the parent array rather than the beginning of the slice.

Scenario

I'm prototyping a .NET ONNX text-embedding pipeline. I want to use tensors through preprocessing, inference, and postprocessing, including passing contiguous slices to native inference without another data copy.

The shape is correct, but the pointer addresses the wrong part of the data. That means a native consumer can receive different inputs without a shape error. The repro below removes ONNX and the rest of my application. It only needs System.Numerics.Tensors.

Parent [3,2]       Logical slice [2,2], starting at [1,0]
[91,92]           excluded
[11,12]           [11,12]  <- expected pointer start
[21,22]           [21,22]

Observed pointer starts at 91, not 11.

Configuration

  • Windows x64
  • .NET SDK 10.0.401
  • Runtime .NET 10.0.12
  • NuGet System.Numerics.Tensors 10.0.9
  • JIT, net10.0; NativeAOT explicitly disabled

Reproduction Steps

You can find the runnable file-based app and captured output in this gist. The complete code is also included below.

Save the following as TensorSlicePinning.cs in a new directory outside a repository. File-based apps inherit project settings from parent directories.

#:package [email protected]
#:property TargetFramework=net10.0
#:property AllowUnsafeBlocks=true
#:property PublishAot=false

using System.Numerics.Tensors;

long[] values = [91, 92, 11, 12, 21, 22];
long[] expected = [11, 12, 21, 22];
var parent = Tensor.Create(values, [3, 2]);
var slice = parent.Slice([1, 0]);
Console.WriteLine($"Runtime: {System.Runtime.InteropServices.RuntimeInformation.FrameworkDescription}");
Console.WriteLine($"Shape: [{string.Join(",", slice.Lengths.ToArray())}]; strides: [{string.Join(",", slice.Strides.ToArray())}]");
Console.WriteLine($"Expected slice: [{string.Join(",", expected)}]");
Console.WriteLine($"Indexer first: {slice[0, 0]}");
Console.WriteLine($"Tensor span: [{string.Join(",", slice.AsTensorSpan().GetSpan([0, 0], 4).ToArray())}]");
Console.WriteLine($"Flattened: [{string.Join(",", slice.ToArray())}]");
Console.WriteLine($"Parent GetPinnableReference: {parent.GetPinnableReference()} (expected 91)");
long referenceFirst = slice.GetPinnableReference();
Console.WriteLine($"Slice GetPinnableReference: {referenceFirst} (expected 11)");
unsafe
{
    using var handle = slice.GetPinnedHandle();
    long[] pinned = new ReadOnlySpan<long>(handle.Pointer, 4).ToArray();
    Console.WriteLine($"Slice GetPinnedHandle: [{string.Join(",", pinned)}] (expected [{string.Join(",", expected)}])");
    bool matches = slice[0, 0] == 11 && slice.ToArray().SequenceEqual(expected)
        && slice.AsTensorSpan().GetSpan([0, 0], 4).SequenceEqual(expected)
        && referenceFirst == 11 && pinned.SequenceEqual(expected);
    Console.WriteLine(matches ? "PASS" : "FAIL: pinning does not start at the logical slice offset.");
    return matches ? 0 : 1;
}

dotnet run --file .\TensorSlicePinning.cs

Actual behavior

Compilation succeeds. Running the app produces the following output and exits with code 1:

Runtime: .NET 10.0.12
Shape: [2,2]; strides: [2,1]
Expected slice: [11,12,21,22]
Indexer first: 11
Tensor span: [11,12,21,22]
Flattened: [11,12,21,22]
Parent GetPinnableReference: 91 (expected 91)
Slice GetPinnableReference: 91 (expected 11)
Slice GetPinnedHandle: [91,92,11,12] (expected [11,12,21,22])
FAIL: pinning does not start at the logical slice offset.

Expected behavior

The reference and pointer should start at 11, the first element of the slice. Reading four elements should return [11,12,21,22], matching the indexer and tensor span. The app should exit with code 0.

Observation Expected Actual
Slice indexer, first element 11 11
Tensor-span and enumerated values [11,12,21,22] [11,12,21,22]
Slice GetPinnableReference() 11 91
Four values through slice GetPinnedHandle() [11,12,21,22] [91,92,11,12]
Process exit code 0 1

Other information

All pointer reads happen while the MemoryHandle is alive and stay within the allocated six-element array.

In the 10.0.9 source, Slice records _start, but GetPinnableReference() appears to return the array's first element without applying that offset. GetPinnedHandle() uses that reference. This looks related to the behavior above.

I've reproduced this on the configuration listed above. I haven't checked other versions or platforms.

Dominant language
C#
Stars
18.3k
Forks
5.6k
Avg merge
2d 22h
Merged PRs (30d)
614

Getting set up

Open in Codespaces

Starts the project's dev container in your browser, under your own GitHub account.

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

More from dotnet/runtime

All issues in dotnet/runtime

Similar issues

More C# issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.