Query/Perf: Compile identifier lambdas passed to PopulateIncludeCollection rather than inline

[maumar]

Part of the bigger perf issue: #35053

PopulateIncludeCollection (as well as couple other methods) take a number of delegate arguments. For scenarios with significant number of entities we see significant perf improvement when these delegates are compiled (like we used to do in EF8), rather than inlined.

Relevant shaper fragment in EF8

ShaperProcessingExpressionVisitor.PopulateIncludeCollection<SalesOrderHeader, SalesOrderDetail>(
	collectionId: 0, 
	queryContext: queryContext, 
	dbDataReader: dataReader, 
	resultCoordinator: resultCoordinator, 
	parentIdentifier: Func<QueryContext, DbDataReader, object[]>, 
	outerIdentifier: Func<QueryContext, DbDataReader, object[]>, 
	selfIdentifier: Func<QueryContext, DbDataReader, object[]>, 
       (...)

and the same fragment in EF9:

ShaperProcessingExpressionVisitor.PopulateIncludeCollection<SalesOrderHeader, SalesOrderDetail>(
	collectionId: 0, 
	queryContext: queryContext, 
	dbDataReader: dataReader, 
	resultCoordinator: resultCoordinator, 
	parentIdentifier: (queryContext, dataReader) => new object[]
	{ 
		(object)(int?)dataReader.GetInt32(0), 
		(object)(int?)dataReader.GetInt32(26) 
	}, 
	outerIdentifier: (queryContext, dataReader) => new object[]
	{ 
		(object)(int?)dataReader.GetInt32(0), 
		(object)(int?)dataReader.GetInt32(26) 
	}, 
	selfIdentifier: (queryContext, dataReader) => new object[]
	{ 
		(object)dataReader.IsDBNull(33) ? default(int?) : (int?)dataReader.GetInt32(33), 
		(object)dataReader.IsDBNull(34) ? default(int?) : (int?)dataReader.GetInt32(34) 
	}, 
       (...)

Benchmark results:

before the fix (but already with invoke fix and no interpretation)

Method	Async	Mean	Error	StdDev	Op/s	Gen0	Gen1	Allocated
MultiInclue	False	455.1 ms	8.94 ms	10.29 ms	2.197	11000.0000	6000.0000	67.92 MB
MultiInclue	True	435.4 ms	1.77 ms	1.66 ms	2.297	11000.0000	6000.0000	67.92 MB

after the fix:

Method	Async	Mean	Error	StdDev	Op/s	Gen0	Gen1	Allocated
MultiInclue	False	363.3 ms	6.72 ms	6.29 ms	2.752	9000.0000	3000.0000	58.51 MB
MultiInclue	True	351.9 ms	2.08 ms	1.74 ms	2.842	9000.0000	3000.0000	58.51 MB

This gets us pretty close to the EF8 numbers which were:

Method	Async	Mean	Error	StdDev	Op/s	Gen0	Gen1	Allocated
MultiInclue	False	297.1 ms	1.47 ms	1.30 ms	3.365	8000.0000	6000.0000	52.4 MB
MultiInclue	True	290.2 ms	3.76 ms	3.52 ms	3.446	8500.0000	6000.0000	52.4 MB

[maumar]

re-openinig for potential patch