[llvm] [AMDGPU] Remove wavefrontsize feature from GFX10+ (PR #98400)

[Stanislav Mekhanoshin via llvm-commits]

rampitec wrote:

Also note that clang and flang do the same thing by calling `AMDGPU::insertWaveSizeFeature()` from TargetParser.cpp.

https://github.com/llvm/llvm-project/pull/98400