Statistically tests should be able to compare methods only among a specific runtime in multi-framework bench

[tats-u]

Method	Runtime	Input	Count	Mean	Error	StdDev	Ratio	MannWhitney(5%)	RatioSD	Gen0	Gen1	Gen2	Allocated	Alloc Ratio
RepeatWithCounter	.NET 8.0	👍	3	22.51 ns	0.477 ns	0.423 ns	1.55	Slower	0.03	0.0204	-	-	128 B	3.20
RepeatNoCounter	.NET 8.0	👍	3	14.51 ns	0.219 ns	0.194 ns	1.00	Base	0.00	0.0063	-	-	40 B	1.00
RepeatPowerShell	.NET 8.0	👍	3	16.76 ns	0.158 ns	0.132 ns	1.16	Slower	0.02	0.0063	-	-	40 B	1.00
RepeatDoubleBlockSize	.NET 8.0	👍	3	17.74 ns	0.304 ns	0.455 ns	1.22	Slower	0.04	0.0063	-	-	40 B	1.00
RepeatWithCounter	NativeAOT 8.0	👍	3	27.16 ns	0.508 ns	0.475 ns	1.87	Slower	0.02	0.0179	-	-	112 B	2.80
RepeatNoCounter	NativeAOT 8.0	👍	3	17.75 ns	0.074 ns	0.058 ns	1.22	Slower	0.02	0.0063	-	-	40 B	1.00
RepeatPowerShell	NativeAOT 8.0	👍	3	19.38 ns	0.435 ns	0.565 ns	1.34	Slower	0.04	0.0063	-	-	40 B	1.00
RepeatDoubleBlockSize	NativeAOT 8.0	👍	3	20.51 ns	0.142 ns	0.133 ns	1.41	Slower	0.02	0.0063	-	-	40 B	1.00

For example, we should be able to compare 4 methods in each runtime respectly:

Method	Runtime	MannWhitney(5%)
RepeatWithCounter	NET 8.0	Slower
RepeatNoCounter	.NET 8.0	Base
RepeatPowerShell	.NET 8.0	Slower
RepeatDoubleBlockSize	.NET 8.0	Slower
RepeatWithCounter	NativeAOT 8.0	? (Slower?)
RepeatNoCounter	NativeAOT 8.0	Base
RepeatPowerShell	NativeAOT 8.0	? (Slower?)
RepeatDoubleBlockSize	NativeAOT 8.0	? (Slower?)