[Jiterpreter] Add support for TryGetHashCode intrinsic

[AustinWise]

#80520 added support in the Jiterpreter for the intrinsic that underlies RuntimeHelpers.GetHashCode. This PR adds similar support for the intrinsic that underlies RuntimeHelpers.TryGetHashCode.

For reference RuntimeHelpers.TryGetHashCode was added in #80059. It is currently only used in ConditionalWeakTable.TryGetValue.

Testing

./dotnet.sh build /t:Test src/libraries/System.Runtime/tests /p:TargetOS=browser /p:TargetArchitecture=wasm /p:Configuration=Release

[ghost]

Tagging subscribers to this area: @BrzVlad, @kotlarmilos
See info in area-owners.md if you want to be subscribed.

Issue Details

---

#80520 added support in the Jiterpreter for the intrinsic that underlies RuntimeHelpers.GetHashCode. This PR adds similar support for the intrinsic that underlies RuntimeHelpers.TryGetHashCode.

For reference RuntimeHelpers.TryGetHashCode was added in #80059. It is currently only used in ConditionalWeakTable.TryGetValue.

Author:	AustinWise
Assignees:	-
Labels:	area-Codegen-Interpreter-mono, community-contribution
Milestone:	-

[AustinWise]

@kg is there a documented way to test and benchmark Jiterpreter changes? I had BenchmarkDotNet project that defined some custom WASM jobs, but it appears that no longer works.

[kg]

@kg is there a documented way to test and benchmark Jiterpreter changes? I had BenchmarkDotNet project that defined some custom WASM jobs, but it appears that no longer works.

It should work the same as anything else, you can manually adjust the tiering thresholds for it so that traces compile sooner (either by passing a runtime arg, or editing the defaults in options-def.h).

To set a runtime arg you can pass it from JS when initializing the runtime.

[AustinWise]

I found the incantation to make my benchmark work again:

./dotnet.sh build -p:TargetOS=browser -p:TargetArchitecture=wasm -c Release src/mono/wasm/Wasm.Build.Tests /t:InstallWorkloadUsingArtifacts

This creates the dotnet-latest directory containing a .NET SDK that can be used with BenchmarkDotNet.

Anyways, the results of running the benchmark:

BenchmarkDotNet=v0.13.3, OS=ubuntu 22.04
AMD Ryzen Threadripper PRO 3955WX 16-Cores, 1 CPU, 32 logical and 16 physical cores
.NET SDK=7.0.102
  [Host]     : .NET 7.0.2 (7.0.222.60605), X64 RyuJIT AVX2
  PR         : .NET Core (Mono) 8.0.0-dev, Wasm AOT
  merge-base : .NET Core (Mono) 8.0.0-dev, Wasm AOT

Runtime=Wasm  IterationCount=15  LaunchCount=2  
WarmupCount=10  
V8 version 11.1.92

Method	Job	Toolchain	NumberOfObjects	Mean	Error	StdDev	Ratio	RatioSD
TryGetNonExistentValue	PR	Wasm: PR	1000	132.9 ns	0.83 ns	1.21 ns	1.01	0.01
TryGetNonExistentValue	merge-base	Wasm: merge-base	1000	131.0 ns	1.07 ns	1.56 ns	1.00	0.00
TryGetExistingValue	PR	Wasm: PR	1000	198.4 ns	2.41 ns	3.61 ns	0.90	0.02
TryGetExistingValue	merge-base	Wasm: merge-base	1000	219.8 ns	0.96 ns	1.41 ns	1.00	0.00

I don't know for sure if the Jiterpreter is even being run by this benchmark, but it does seem to improve performance. Running it a second time has similar results, so I'm assuming it is not a fluke caused by differing numbers of has collisions run-to-run.

[kg]

You'll need to rebase this now to resolve a conflict (sorry! 🙇‍♀️), it shouldn't be too hard to do though. The jiterpreter code just got rearranged and split into two files, but copying your work over should be a simple bit of copy-paste.

[AustinWise]

I resolved the merge conflicts and the tests pass.

[kg]

I resolved the merge conflicts and the tests pass.

Great! I'll make sure we're cool adding this (I don't see any problems) and get it merged.

[kg]

Thank you for your contribution! 🎉