Add TCPChecksum to SVE microbenchmark

[ylpoonlg]

Performance Results

Run on Neoverse-V2

Method	Size	Mean	Error	StdDev	Median	Min	Max	Allocated
Scalar	527	105.68 ns	0.105 ns	0.088 ns	105.65 ns	105.62 ns	105.93 ns	-
Vector128TCPChecksum	527	13.82 ns	0.017 ns	0.013 ns	13.82 ns	13.79 ns	13.83 ns	-
SveTCPChecksum	527	26.27 ns	0.236 ns	0.220 ns	26.35 ns	25.87 ns	26.59 ns	-
Scalar	1234	213.40 ns	0.096 ns	0.080 ns	213.40 ns	213.28 ns	213.59 ns	-
Vector128TCPChecksum	1234	30.98 ns	0.393 ns	0.307 ns	30.88 ns	30.85 ns	31.94 ns	-
SveTCPChecksum	1234	58.24 ns	0.249 ns	0.233 ns	58.25 ns	57.77 ns	58.51 ns	-
Scalar	10015	2,036.12 ns	1.814 ns	1.697 ns	2,036.36 ns	2,033.20 ns	2,038.84 ns	-
Vector128TCPChecksum	10015	285.82 ns	0.885 ns	0.739 ns	285.58 ns	285.20 ns	287.46 ns	-
SveTCPChecksum	10015	475.23 ns	0.922 ns	0.770 ns	474.84 ns	474.68 ns	477.21 ns	-

cc @dotnet/arm64-contrib @SwapnilGaikwad @LoopedBard3

[LoopedBard3]

Looks good, thanks for the benchmarks!

[Copilot]

Pull request overview

Adds a new SVE-focused microbenchmark to measure TCP-style checksum computation and compare scalar vs AdvSimd (Vector128) vs SVE implementations, consistent with the existing src/benchmarks/micro/sve benchmark suite.

Changes:

• Introduces TCPChecksum benchmark with scalar, Vector128/AdvSimd, and SVE variants.
• Adds deterministic packet/length generation in GlobalSetup and scalar equivalence validation in GlobalCleanup.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

You can also share your feedback on Copilot code review. Take the survey.

[LoopedBard3]

Thanks for the tests!