[nativeaot][tests] Add Native AOT iOS perf scenarios

[kotlarmilos]

This PR updates the CI jobs to include the Native AOT iOS performance scenarios. The changes include building the NativeAOT iOS scenarios and running them, as well as updating the underlying scripts and project files to support it. Additionally, it separates the scenarios from the current Mono AOT sample measurements.

Additionally, it temporarily fixes regression to allow testing. The regression will be fixed on the main in a separate PR tracked in #80911.

Preliminary perf measurements for the HelloiOS app using Native AOT.

SOD - iOS HelloWorld Native AOT Zip Size

Metric	Average	Min	Max
SOD - iOS HelloWorld Native AOT Zip Size	1727305.000 bytes	1727305.000 bytes	1727305.000 bytes

SOD - iOS HelloWorld Native AOT .app

Metric	Average	Min	Max
SOD - iOS HelloWorld Native AOT .app	8075485.000 bytes	8075485.000 bytes	8075485.000 bytes
pub/nativeaot/HelloiOS.app/HelloiOS	8074496.000 bytes	8074496.000 bytes	8074496.000 bytes
pub/nativeaot/HelloiOS.app/Info.plist	981.000 bytes	981.000 bytes	981.000 bytes
pub/nativeaot/HelloiOS.app/PkgInfo	8.000 bytes	8.000 bytes	8.000 bytes

Device Startup - iOS HelloWorld Native AOT

Metric	Average	Min	Max
Generic Startup	38.600 ms	34.000 ms	41.000 ms

[ghost]

Tagging subscribers to 'os-ios': @steveisok, @akoeplinger
See info in area-owners.md if you want to be subscribed.

Issue Details

---

This PR updates the CI jobs to include the Native AOT iOS performance scenarios. The changes include building the NativeAOT iOS scenarios and running them, as well as updating the underlying scripts and project files to support it. Additionally, it separates the scenarios from the current Mono AOT sample measurements.

Additionally, it temporarily fixes regression to allow testing. The regression will be fixed on the main in a separate PR tracked in #80911.

Author:	kotlarmilos
Assignees:	kotlarmilos
Labels:	os-ios, area-NativeAOT-coreclr
Milestone:	8.0.0

[kotlarmilos]

This PR is ready for review. The performance results can be found in the dotnet-runtime-perf pipeline. The Mono performance runs have been expanded to include separate runs with and without symbols. Once this PR is merged, the following job names need to be updated in the dashboard.

Native AOT performance measurements:

• SOD - iOS HelloWorld Native AOT .app Size symbols
• SOD - iOS HelloWorld Native AOT .app Size nosymbols
• SOD - iOS HelloWorld Native AOT Zip Size symbols
• SOD - iOS HelloWorld Native AOT Zip Size nosymbols
• SOD - Device Startup - iOS Native AOT HelloWorld symbols
• SOD - Device Startup - iOS Native AOT HelloWorld nosymbols

Mono performance measurements:

• SOD - iOS HelloWorld Mono .app Size llvm symbols
• SOD - iOS HelloWorld Mono .app Size llvm nosymbols
• SOD - iOS HelloWorld Mono .app Size nollvm symbols
• SOD - iOS HelloWorld Mono .app Size nollvm nosymbols
• SOD - iOS HelloWorld Mono Zip Size llvm symbols
• SOD - iOS HelloWorld Mono Zip Size llvm nosymbols
• SOD - iOS HelloWorld Mono Zip Size nollvm symbols
• SOD - iOS HelloWorld Mono Zip Size nollvm nosymbols
• Device Startup - iOS Mono HelloWorld llvm symbols
• Device Startup - iOS Mono HelloWorld llvm nosymbols
• Device Startup - iOS Mono HelloWorld nollvm symbols
• Device Startup - iOS Mono HelloWorld nollvm nosymbols

@LoopedBard3 I will keep you posted.

[ivanpovazan]

Please take a look at my comment: https://github.com/dotnet/runtime/pull/85317/files#r1210389112

[ivanpovazan]

LGTM!
Just a small note on "symbol-stripped" scenarios, that in those cases we are stripping local symbols only via strip -x

[kotlarmilos]

LGTM! Just a small note on "symbol-stripped" scenarios, that in those cases we are stripping local symbols only via strip -x

Thanks! If we maintain consistency across all platforms and strip in the same way, it should be fine. If I recall correctly, we encountered an issue with global symbols stripping.