clang should directly consume launch_bounds attribute for HIP

[arsenm]

Clang has the launch_bounds attribute, which is not implemented for AMDGPU. Instead, the HIP headers work around this by implementing this with separate macros

The directly consumed attribute is silently ignored

[llvmbot]

@llvm/issue-subscribers-clang-codegen

Author: Matt Arsenault (arsenm)

Clang has the [launch_bounds attribute](https://clang.llvm.org/docs/AttributeReference.html#launch-bounds), which is not implemented for AMDGPU. Instead, the HIP headers work around this by [implementing this with separate macros ](https://github.com/ROCm/clr/blob/933aa1d3a7bc4e4a2b4cfb2ad7e4c40df0b8ae61/hipamd/include/hip/amd_detail/amd_hip_runtime.h#L176)

The directly consumed attribute is silently ignored

[llvmbot]

@llvm/issue-subscribers-backend-amdgpu

Author: Matt Arsenault (arsenm)

Clang has the [launch_bounds attribute](https://clang.llvm.org/docs/AttributeReference.html#launch-bounds), which is not implemented for AMDGPU. Instead, the HIP headers work around this by [implementing this with separate macros ](https://github.com/ROCm/clr/blob/933aa1d3a7bc4e4a2b4cfb2ad7e4c40df0b8ae61/hipamd/include/hip/amd_detail/amd_hip_runtime.h#L176)

The directly consumed attribute is silently ignored