Currently we have two ways of defining intrinsic functions in JIT, for better consistency + the fact that the inliner can only recognize NamedIntrinsics we should convert all of the old ones to NamedIntrinsics just like in this PR: #52156.

• CORINFO_INTRINSIC_Array_Get
• CORINFO_INTRINSIC_Array_Address
• CORINFO_INTRINSIC_Array_Set
• CORINFO_INTRINSIC_InitializeArray
• CORINFO_INTRINSIC_RTH_GetValueInternal
• CORINFO_INTRINSIC_Object_GetType
• CORINFO_INTRINSIC_StubHelpers_GetStubContext
• CORINFO_INTRINSIC_StubHelpers_GetStubContextAddr
• CORINFO_INTRINSIC_StubHelpers_NextCallReturnAddress
• CORINFO_INTRINSIC_InterlockedAdd32
• CORINFO_INTRINSIC_InterlockedAdd64
• CORINFO_INTRINSIC_InterlockedXAdd32
• CORINFO_INTRINSIC_InterlockedXAdd64
• CORINFO_INTRINSIC_InterlockedXchg32
• CORINFO_INTRINSIC_InterlockedXchg64
• CORINFO_INTRINSIC_InterlockedCmpXchg32
• CORINFO_INTRINSIC_InterlockedCmpXchg64
• CORINFO_INTRINSIC_MemoryBarrier
• CORINFO_INTRINSIC_MemoryBarrierLoad
• CORINFO_INTRINSIC_ByReference_Ctor
• CORINFO_INTRINSIC_ByReference_Value
• CORINFO_INTRINSIC_GetRawHandle

The most important (popular) ones are CORINFO_INTRINSIC_Object_GetType and Interlocked*

NOTE: Possible crossgen(2) issue: #52156 (comment)
NOTE2 (unrelated): We might want to implement a sort of a binary search instead of that long chain of strcmp for NamedIntrinsics