Fix SIMD intrinsics handling in crossgen2 (dotnet/coreclr#27853)
* Fix SIMD intrinsics handling in crossgen2 Crossgen2 was compiling methods that call SIMD intrinsics (System.Numerics.Vector<T>). This is not correct, as the size of the vector is a runtime specific detail - e.g. when running on devices without SSE2 support, the size is 4 and when running on devices with SSE2 support, the size is 8. This fixes runtime errors in 6 coreclr pri 0 tests. * Reflect PR feedback Create a IsVectorOfT method and use it at all places where we were previously checking the namespace and type name of Vector<T> Commit migrated from https://github.com/dotnet/coreclr/commit/0f9814cc4e25ee4060d8f3ae2305210aef94f1bc
Showing
想要评论请 注册 或 登录