Add is pattern benchmarks

[jjonescz]

Follow up on #84961 (comment).

Results of IsPatternLeadingSlashBenchmarks

BenchmarkDotNet v0.16.0-preview.1, Windows 11 (10.0.26100.9106/24H2/2024Update/HudsonValley) (Hyper-V)
Intel Xeon Platinum 8370C CPU 2.80GHz (Max: 2.79GHz), 1 CPU, 16 logical and 8 physical cores
Memory: 63.95 GB Total, 16.03 GB Available
.NET SDK 11.0.100-preview.7.26381.103
  [Host]   : .NET 10.0.11 (10.0.11, 10.0.1126.37416), X64 RyuJIT x86-64-v4
  ShortRun : .NET 10.0.11 (10.0.11, 10.0.1126.37416), X64 RyuJIT x86-64-v4

Job=ShortRun  EvaluateOverhead=False  IterationCount=3  
LaunchCount=1  WarmupCount=3  

Method	Input	Mean	Error	StdDev	Ratio	RatioSD
Comparisons	Empty	4.081 ns	6.2051 ns	0.3401 ns	1.00	0.00
Pattern	Empty	3.537 ns	0.9690 ns	0.0531 ns	0.87	0.07
Comparisons	Slash	4.010 ns	1.4465 ns	0.0793 ns	1.00	0.00
Pattern	Slash	4.241 ns	0.7739 ns	0.0424 ns	1.06	0.02
Comparisons	Backslash	3.593 ns	2.5362 ns	0.1390 ns	1.00	0.00
Pattern	Backslash	4.629 ns	0.6133 ns	0.0336 ns	1.29	0.04
Comparisons	Other	4.184 ns	3.6089 ns	0.1978 ns	1.00	0.00
Pattern	Other	4.678 ns	3.8583 ns	0.2115 ns	1.12	0.06

Microsoft Reviewers: Open in CodeFlow

[azure-pipelines]

Azure Pipelines:
Successfully started running 2 pipeline(s).
There may be pipelines that require an authorized user to comment /azp run to run.

[jjonescz]

📝 I guess the pattern cases will become equivalent to the comparison cases when the compiler used to build these is updated to include the fix, but perhaps these benchmarks are still useful as a sort of regression gate?

[Copilot]

Pull request overview

This PR adds new micro-benchmarks under src/Tools/Benchmarks for small is-pattern scenarios and updates the compiler’s LocalRewriter_IsPatternOperator comment to reference an in-repo benchmark source instead of an external PR link.

Changes:

• Added IsPatternBenchmarks.cs containing benchmarks comparing comparison-based implementations vs is-pattern implementations for a few small-pattern scenarios.
• Updated the documentation comment for MaxTestsForInvertedLinearSequence to point at the new benchmark file.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 2 comments.

File	Description
src/Tools/Benchmarks/IsPatternBenchmarks.cs	Introduces new BenchmarkDotNet benchmarks for small is-pattern cases (leading slash and small byte-pattern sets).
src/Compilers/CSharp/Portable/Lowering/LocalRewriter/LocalRewriter_IsPatternOperator.cs	Updates the explanatory comment for the inverted-linear-sequence threshold to reference in-repo benchmark code.

Suppressed comments (1)

src/Tools/Benchmarks/IsPatternBenchmarks.cs:107

• The baseline implementation compares a byte (Value) to char literals. Aligning the baseline with the pattern version by using byte constants makes the comparison more directly about the lowering strategy and avoids implicit integral promotions.

    public bool Comparisons() => Guard && (Value == '+' || Value == '.' || Value == 'E' || Value == 'e');

[Copilot]

Pull request overview

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

Suppressed comments (1)

Previously missed (1) — in code that hasn't changed since the last review.

src/Tools/Benchmarks/IsPatternBenchmarks.cs:34

• Avoid the null-forgiving initializer here; the benchmark can start with a non-null default and still be overwritten in GlobalSetup. This removes a nullability suppression and makes the type safe if the benchmark is ever instantiated outside BenchmarkDotNet’s setup flow.

    private string _path = null!;

[azure-pipelines]

Azure Pipelines:
Successfully started running 2 pipeline(s).
There may be pipelines that require an authorized user to comment /azp run to run.