benchmarks: Improve benchmark for better precision and run time.

[jeanplevesque]

GitHub Issue: #

Proposed Changes

• Bug fix
• Feature
• Code style update (formatting)
• Refactoring (no functional changes, no api changes)
• Build or CI related changes
• Documentation content changes
• Other, please describe: Benchmarks

What is the current behavior?

• Some benchmarks don't follow best practices from the BenchmarkDotNet documentation, which invalidates the results.
  • IterationSetup is used on iterations that last less than 100 ms.
• The result table is bloated.

Method	Job	InvocationCount	UnrollFactor	Mean	Error	StdDev	Median	Gen0	Gen1	Gen2	Allocated
CreateViewModel	DefaultJob	Default	16	1.280 μs	0.0255 μs	0.0615 μs	1.268 μs	0.1488	0.1450	0.0019	936 B
CreateViewModel_WithExplicitName	DefaultJob	Default	16	1.193 μs	0.0236 μs	0.0419 μs	1.188 μs	0.1488	0.1469	-	936 B
ReadProperty_Unresolved	Job-PVCBIR	1	1	2.810 μs	0.0599 μs	0.1376 μs	2.800 μs	-	-	-	856 B
ReadProperty_Resolved	Job-PVCBIR	1	1	1.515 μs	0.0342 μs	0.0941 μs	1.500 μs	-	-	-	720 B
SetProperty_Unresolved	Job-PVCBIR	1	1	19.909 μs	0.4667 μs	1.3389 μs	19.800 μs	-	-	-	2592 B
SetProperty_Resolved	Job-PVCBIR	1	1	1.831 μs	0.0401 μs	0.0763 μs	1.800 μs	-	-	-	656 B
DisposeViewModel	Job-PVCBIR	1	1	1.335 μs	0.0318 μs	0.0880 μs	1.300 μs	-	-	-	600 B
SetProperty_Unresolved_WithListener	Job-PVCBIR	1	1	21.604 μs	0.4888 μs	1.3627 μs	21.600 μs	-	-	-	2616 B
SetProperty_Resolved_WithListener	Job-PVCBIR	1	1	2.625 μs	0.0730 μs	0.2047 μs	2.600 μs	-	-	-	680 B
DisposeViewModel_WithListener	Job-PVCBIR	1	1	1.417 μs	0.0344 μs	0.0943 μs	1.400 μs	-	-	-	600 B

What is the new behavior?

• Commits with the benchmarks conventional commit type are now allowed.
• IterationSetup is used on iterations that last more than 100 ms in conjuction with OperationsPerInvoke.
• Benchmarks about the ViewModelBase class and the extension methods on IViewModel are now better separated.
• The results table is lean.

Method	Mean	Gen0	Gen1	Allocated
CreateViewModel	623.60 ns	0.0858	0.0849	544 B
CreateViewModel_WithExplicitName	587.77 ns	0.0858	0.0849	544 B
DisposeViewModel	62.59 ns	-	-	-
GetFromValue_Unresolved	427.69 ns	0.0405	0.0401	256 B
GetFromObservable_Unresolved	544.88 ns	0.0477	0.0467	304 B
GetFromValue_Resolved	46.09 ns	0.0191	-	120 B
GetFromObservable_Resolved	52.72 ns	0.0204	-	128 B
Set_Unresolved	1,210.68 ns	0.1106	0.0286	695 B
Set_Resolved	30.67 ns	0.0038	-	24 B

Impact on version

• Major (Public API was modified.)
  • Public constructs (class, struct, delegate, enum, etc.) were removed or renamed.
  • Public members were removed or renamed.
  • Public method signatures were changed or renamed.
• Minor (Public API was extended.)
  • Public constructs, members, or overloads were added.
• Patch (Public API was unchanged.)
  • A bug in behavior was fixed.
  • Documentation was changed.
• None (The library is unchanged.)
  • Only code under the build folder was changed.
  • Only code under the .github folder was changed.
  • Only code in the Benchmarks project was changed.

Checklist

Please check that your PR fulfills the following requirements:

• Documentation has been added/updated.
• Automated Unit / Integration tests for the changes have been added/updated.
• Updated BREAKING_CHANGES.md (if you introduced a breaking change).
• Your conventional commits are aligned with the Impact on version section.

Other information