Add ScatterStore and GatherLoad to SVE microbenchmark

[ylpoonlg]

Performance Results

Run on Neoverse-V2

Method	Size	Mean	Error	StdDev	Median	Min	Max	Allocated
Scalar	15	6.283 ns	0.0110 ns	0.0103 ns	6.282 ns	6.270 ns	6.304 ns	-
SveScatterStore	15	4.492 ns	0.0076 ns	0.0071 ns	4.491 ns	4.482 ns	4.507 ns	-
Scalar	127	47.276 ns	0.0668 ns	0.0592 ns	47.291 ns	47.157 ns	47.368 ns	-
SveScatterStore	127	40.656 ns	0.3856 ns	0.3220 ns	40.743 ns	39.585 ns	40.765 ns	-
Scalar	527	179.332 ns	0.1554 ns	0.1213 ns	179.324 ns	179.081 ns	179.593 ns	-
SveScatterStore	527	156.546 ns	0.0273 ns	0.0228 ns	156.542 ns	156.519 ns	156.603 ns	-
Scalar	10015	4,222.043 ns	6.5129 ns	6.0921 ns	4,222.449 ns	4,209.463 ns	4,231.155 ns	-
SveScatterStore	10015	3,395.879 ns	3.7978 ns	3.3667 ns	3,396.018 ns	3,391.456 ns	3,401.467 ns	-

Method	Size	Mean	Error	StdDev	Median	Min	Max	Allocated
Scalar	15	4.542 ns	0.0030 ns	0.0023 ns	4.542 ns	4.539 ns	4.546 ns	-
SveGatherLoad	15	4.096 ns	0.0015 ns	0.0013 ns	4.096 ns	4.094 ns	4.097 ns	-
Scalar	127	62.573 ns	0.1186 ns	0.1109 ns	62.609 ns	62.371 ns	62.702 ns	-
SveGatherLoad	127	32.776 ns	0.2165 ns	0.2026 ns	32.664 ns	32.573 ns	33.097 ns	-
Scalar	527	307.239 ns	0.2782 ns	0.2323 ns	307.237 ns	306.579 ns	307.554 ns	-
SveGatherLoad	527	132.341 ns	0.8452 ns	0.7906 ns	132.432 ns	131.385 ns	134.162 ns	-
Scalar	10015	6,383.405 ns	86.2813 ns	67.3627 ns	6,397.476 ns	6,268.075 ns	6,469.938 ns	-
SveGatherLoad	10015	2,607.590 ns	1.1626 ns	1.0306 ns	2,607.389 ns	2,606.123 ns	2,609.474 ns	-

cc @dotnet/arm64-contrib @SwapnilGaikwad @LoopedBard3

[Copilot]

Pull request overview

Adds two new Arm SVE-focused microbenchmarks to measure gather-load and scatter-store patterns on Arm64 hardware with SVE support, consistent with the existing src/benchmarks/micro/sve benchmark suite.

Changes:

• Introduce ScatterStore benchmark with scalar and SVE scatter-store implementations.
• Introduce GatherLoad benchmark with scalar and SVE gather-load implementations plus result verification scaffolding.

Reviewed changes

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

File	Description
src/benchmarks/micro/sve/ScatterStore.cs	New SVE benchmark comparing scalar indexed stores vs SVE scatter stores.
src/benchmarks/micro/sve/GatherLoad.cs	New SVE benchmark comparing scalar indexed loads+sum vs SVE gather loads and vector reduction.

---

💡 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 new tests!