Do we still need [AggressiveOptimization] on so many methods?

[stephentoub]

dotnet/coreclr#22191 added a bunch of [AggressiveOptimization] attributes to methods throughout SpanHelpers and elsewhere that make heavy use of intrinsics, e.g.

runtime/src/libraries/System.Private.CoreLib/src/System/SpanHelpers.Byte.cs

Lines 337 to 338 in 2a01ceb

	[MethodImpl(MethodImplOptions.AggressiveOptimization)]
	public static bool Contains(ref byte searchSpace, byte value, int length)

After all of the inlining improvements made in .NET 6 and 7, are these still necessary and helping more than they hurt?

category:design
theme:basic-cq
skill-level:beginner
cost:small
impact:small

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

Issue Details

---

dotnet/coreclr#22191 added a bunch of [AggressiveOptimization] attributes to methods throughout SpanHelpers and elsewhere that make heavy use of intrinsics, e.g.

runtime/src/libraries/System.Private.CoreLib/src/System/SpanHelpers.Byte.cs

Lines 337 to 338 in 2a01ceb

	[MethodImpl(MethodImplOptions.AggressiveOptimization)]
	public static bool Contains(ref byte searchSpace, byte value, int length)

After all of the inlining improvements made in .NET 6 and 7, are these still necessary and helping more than they hurt?

Author:	stephentoub
Assignees:	-
Labels:	area-CodeGen-coreclr
Milestone:	7.0.0

[jkotas]

These [AggressiveOptimization] attributes were added as a workaround for #9120

[EgorBo]

I think we can remove them at least from SpanHelpers because SPC.dll is always prejitted (R2R) so e.g. SpanHelpers will never see Tier0 and will be properly instrumented and optimized for current app (in Dynamic PGO mode), e.g. if current app mostly works with large inputs we will re-order "handle small-inputs" blocks, etc.

However, we might need to check that SpanHelpers is actually fully prejitted due to presence of Vector<T> (see #67398 (comment))

[EgorBo]

SpanHelpers members which refuse to precompile (because they use Vector<T> without Vector128 and Vector256 paths)

Info: Method `[S.P.CoreLib]System.SpanHelpers.Fill<__Canon>(__Canon&,native uint,__Canon)` was not compiled because `[S.P.CoreLib]System.Numerics.Vector`1<uint8>` requires runtime JIT
Info: Method `[S.P.CoreLib]System.SpanHelpers.UnalignedCountVector(char&)` was not compiled because `This function is using SIMD intrinsics, their size is machine specific` requires runtime JIT
Info: Method `[S.P.CoreLib]System.SpanHelpers.GetCharVectorSpanLength(native int,native int)` was not compiled because `This function is using SIMD intrinsics, their size is machine specific` requires runtime JIT
Info: Method `[S.P.CoreLib]System.SpanHelpers.UnalignedCountVectorFromEnd(uint8&,int32)` was not compiled because `This function is using SIMD intrinsics, their size is machine specific` requires runtime JIT
Info: Method `[S.P.CoreLib]System.SpanHelpers.UnalignedCountVector(uint8&)` was not compiled because `This function is using SIMD intrinsics, their size is machine specific` requires runtime JIT
Info: Method `[S.P.CoreLib]System.SpanHelpers.GetByteVectorSpanLength(native uint,int32)` was not compiled because `This function is using SIMD intrinsics, their size is machine specific` requires runtime JIT
Info: Method `[S.P.CoreLib]System.SpanHelpers.LastIndexOfAny(uint8&,uint8,uint8,uint8,int32)` was not compiled because `[S.P.CoreLib]System.Numerics.Vector`1<uint8>` requires runtime JIT
Info: Method `[S.P.CoreLib]System.SpanHelpers.LastIndexOfAny(uint8&,uint8,uint8,int32)` was not compiled because `[S.P.CoreLib]System.Numerics.Vector`1<uint8>` requires runtime JIT
Info: Method `[S.P.CoreLib]System.SpanHelpers.Fill<char>(char&,native uint,char)` was not compiled because `[S.P.CoreLib]System.Numerics.Vector`1<uint8>` requires runtime JIT
Info: Method `[S.P.CoreLib]System.SpanHelpers.Fill<GuidResult>(GuidResult&,native uint,GuidResult)` was not compiled because `[S.P.CoreLib]System.Numerics.Vector`1<uint8>` requires runtime JIT
Info: Method `[S.P.CoreLib]System.SpanHelpers.Fill<native int>(native int&,native uint,native int)` was not compiled because `[S.P.CoreLib]System.Numerics.Vector`1<uint8>` requires runtime JIT
Info: Method `[S.P.CoreLib]System.SpanHelpers.Fill<int32>(int32&,native uint,int32)` was not compiled because `[S.P.CoreLib]System.Numerics.Vector`1<uint8>` requires runtime JIT
Info: Method `[S.P.CoreLib]System.SpanHelpers.Fill<float32>(float32&,native uint,float32)` was not compiled because `[S.P.CoreLib]System.Numerics.Vector`1<uint8>` requires runtime JIT
Info: Method `[S.P.CoreLib]System.SpanHelpers.Fill<uint16>(uint16&,native uint,uint16)` was not compiled because `[S.P.CoreLib]System.Numerics.Vector`1<uint8>` requires runtime JIT
Info: Method `[S.P.CoreLib]System.SpanHelpers.Fill<uint64>(uint64&,native uint,uint64)` was not compiled because `[S.P.CoreLib]System.Numerics.Vector`1<uint8>` requires runtime JIT
Info: Method `[S.P.CoreLib]System.SpanHelpers.Fill<uint32>(uint32&,native uint,uint32)` was not compiled because `[S.P.CoreLib]System.Numerics.Vector`1<uint8>` requires runtime JIT
Info: Method `[S.P.CoreLib]System.SpanHelpers.Fill<int64>(int64&,native uint,int64)` was not compiled because `[S.P.CoreLib]System.Numerics.Vector`1<uint8>` requires runtime JIT
Info: Method `[S.P.CoreLib]System.SpanHelpers.Fill<int16>(int16&,native uint,int16)` was not compiled because `[S.P.CoreLib]System.Numerics.Vector`1<uint8>` requires runtime JIT
Info: Method `[S.P.CoreLib]System.SpanHelpers.IndexOfValueType<int32>(int32&,int32,int32)` was not compiled because `[S.P.CoreLib]System.Numerics.Vector`1<int32>` requires runtime JIT
Info: Method `[S.P.CoreLib]System.SpanHelpers.IndexOfValueType<int64>(int64&,int64,int32)` was not compiled because `[S.P.CoreLib]System.Numerics.Vector`1<int64>` requires runtime JIT
Info: Method `[S.P.CoreLib]System.SpanHelpers.Fill<KeyValuePair`2<__Canon,__Canon>>(KeyValuePair`2<__Canon,__Canon>&,native uint,KeyValuePair`2<__Canon,__Canon>)` was not compiled because `[S.P.CoreLib]System.Numerics.Vector`1<uint8>` requires runtime JIT
Info: Method `[S.P.CoreLib]System.SpanHelpers.Fill<CalendarId>(CalendarId&,native uint,CalendarId)` was not compiled because `[S.P.CoreLib]System.Numerics.Vector`1<uint8>` requires runtime JIT
Info: Method `[S.P.CoreLib]System.SpanHelpers.Fill<float64>(float64&,native uint,float64)` was not compiled because `[S.P.CoreLib]System.Numerics.Vector`1<uint8>` requires runtime JIT
Info: Method `[S.P.CoreLib]System.SpanHelpers.Fill<KeyValuePair`2<SessionInfo,bool>>(KeyValuePair`2<SessionInfo,bool>&,native uint,KeyValuePair`2<SessionInfo,bool>)` was not compiled because `[S.P.CoreLib]System.Numerics.Vector`1<uint8>` requires runtime JIT
Info: Method `[S.P.CoreLib]System.SpanHelpers.Fill<KeyValuePair`2<int32,__Canon>>(KeyValuePair`2<int32,__Canon>&,native uint,KeyValuePair`2<int32,__Canon>)` was not compiled because `[S.P.CoreLib]System.Numerics.Vector`1<uint8>` requires runtime JIT
Info: Method `[S.P.CoreLib]System.SpanHelpers.Fill<SessionInfo>(SessionInfo&,native uint,SessionInfo)` was not compiled because `[S.P.CoreLib]System.Numerics.Vector`1<uint8>` requires runtime JIT

[EgorBo]

Moving to 8.0.0 as it depends on the planned change to make crossgen avx2-aware by default.

[EgorBo]

I assume we can close this one as we now have only 3 [AO] in BCL:

1. AsyncTaskMethodBuilder<TResult>.AwaitUnsafeOnCompleted the comment states that it allocates in T0 (presumably, when R2R doesn't hande a specific generic instance)
2. (and 3) CastHelper's StelemRef and LdelemaRef -- these have to be [AO] because VM special case them to be direct calls (so there is no call counting in between) - we can make them indirect just like 99% of other methods but VSadov claimed it is needed for perf.