.NET Performance: ToList vs ToArray

.NET 9 brings significant improvements to LINQ’s ToArray method, both in terms of speed, as well as memory allocation. To achieve this, the new implementation makes use of new runtime features such as InlineArray and ArrayPool to significantly reduce memory allocations and improve speed, especially when dealing with an IEnumerable<T> of unknown length. If you’re curious about the implementation, you can check out the PR on GitHub.

.NET 8 vs .NET 9

First, let’s look at the ToArray performance improvement between .NET 8 and .NET 9:

| Method  | Runtime  | Count  | Mean          | Ratio | Allocated | Alloc Ratio |
|-------- |--------- |------- |--------------:|------:|----------:|------------:|
| ToArray | .NET 8.0 | 10     |     115.61 ns |  1.00 |     256 B |        1.00 |
| ToArray | .NET 9.0 | 10     |      71.91 ns |  0.62 |     104 B |        0.41 |
|         |          |        |               |       |           |             |
| ToArray | .NET 8.0 | 1000   |   3,209.52 ns |  1.00 |    8536 B |        1.00 |
| ToArray | .NET 9.0 | 1000   |   2,625.86 ns |  0.82 |    4064 B |        0.48 |
|         |          |        |               |       |           |             |
| ToArray | .NET 8.0 | 100000 | 545,642.89 ns |  1.00 |  925132 B |        1.00 |
| ToArray | .NET 9.0 | 100000 | 362,780.53 ns |  0.67 |  400148 B |        0.43 |

BenchmarkDotNet v0.14.0, Windows 11 (10.0.26100.2605)
AMD Zen 2, 1 CPU, 8 logical and 4 physical cores
  .NET 8.0 : .NET 8.0.11 (8.0.1124.51707), X64 RyuJIT AVX2
  .NET 9.0 : .NET 9.0.0 (9.0.24.52809), X64 RyuJIT AVX2Code language: Bash (bash)

While the results vary slightly depending on the number of items in the iterator, you still get an up to 38% reduction in CPU time and a whooping 57% in allocated memory. That’s quite good!

ToList vs ToArray

Now, traditionally ToList was faster when dealing with collections of unknown sizes because it was skipping that final array allocation, while ToArray was more efficient in terms of memory usage. Starting with .NET 9, that is no longer the case, with ToArray being more efficient from both perspectives.

| Method  | Runtime  | Count  | Mean          | Ratio | Allocated | Alloc Ratio |
|-------- |--------- |------- |--------------:|------:|----------:|------------:|
| ToList  | .NET 9.0 | 10     |      81.44 ns |  1.00 |     256 B |        1.00 |
| ToArray | .NET 9.0 | 10     |      71.91 ns |  0.88 |     104 B |        0.40 |
|         |          |        |               |       |           |             |
| ToList  | .NET 9.0 | 1000   |   2,942.87 ns |  1.00 |    8464 B |        1.00 |
| ToArray | .NET 9.0 | 1000   |   2,625.86 ns |  0.89 |    4064 B |        0.48 |
|         |          |        |               |       |           |             |
| ToList  | .NET 9.0 | 100000 | 494,497.60 ns |  1.00 | 1049112 B |        1.00 |
| ToArray | .NET 9.0 | 100000 | 362,780.53 ns |  0.73 |  400148 B |        0.38 |

BenchmarkDotNet v0.14.0, Windows 11 (10.0.26100.2605)
AMD Zen 2, 1 CPU, 8 logical and 4 physical cores
  .NET 8.0 : .NET 8.0.11 (8.0.1124.51707), X64 RyuJIT AVX2
  .NET 9.0 : .NET 9.0.0 (9.0.24.52809), X64 RyuJIT AVX2Code language: Bash (bash)

Overall we’re looking at a 10 to 30% speed difference and around 60% less memory being allocated, depending on the number of elements.

Running the benchmark

You can run this benchmark yourself if you’d like using BenchmarkDotNet.

[SimpleJob(RuntimeMoniker.Net80, baseline: true)]
[SimpleJob(RuntimeMoniker.Net90)]
[MemoryDiagnoser(false)]
[HideColumns("Job", "Error", "StdDev", "RatioSD")]
public class Benchmarks
{
    [Params(10, 1000, 100000)]
    public int Count;

    [Benchmark]
    public List<int> ToList() => GetItems(Count).ToList();

    [Benchmark]
    public int[] ToArray() => GetItems(Count).ToArray();

    private IEnumerable<int> GetItems(int count)
    {
        for (int i = 0; i < count; i++)
        {
            yield return 1;
        }
    }
}
Code language: C# (cs)

BenchmarkRunner.Run<Benchmarks>();
Code language: C# (cs)