[FEA]: Add dialect macros from upstream

[jamesETsmith]

Is this a duplicate?

• I confirmed there appear to be no duplicate issues for this request.

Is your feature request related to a problem? Please describe.

When compiling DGL on theRock based images, I'm seeing the following problems which looks like an incompatibility between rocthrust (#define THRUST_VERSION 200805) and libhipcxx 2.7.0. For more info about the env setup, see the very rough dockerfile I'm using here (lots of things are commented out because it's still a work in progress just fyi)

In file included from /workspace/graphbolt/src/cuda/extension/gpu_cache.cu:22:
In file included from /workspace/graphbolt/src/cuda/extension/../common.h:29:
In file included from /opt/venv/lib/python3.12/site-packages/torch/include/ATen/hip/HIPEvent.h:5:
In file included from /opt/venv/lib/python3.12/site-packages/torch/include/ATen/hip/HIPContext.h:7:
In file included from /opt/venv/lib/python3.12/site-packages/torch/include/ATen/Context.h:4:
In file included from /opt/venv/lib/python3.12/site-packages/torch/include/ATen/CPUGeneratorImpl.h:3:
In file included from /opt/venv/lib/python3.12/site-packages/torch/include/ATen/core/Generator.h:18:
In file included from /opt/venv/lib/python3.12/site-packages/torch/include/c10/core/GeneratorImpl.h:8:
In file included from /opt/venv/lib/python3.12/site-packages/torch/include/c10/core/TensorImpl.h:11:
In file included from /opt/venv/lib/python3.12/site-packages/torch/include/c10/core/ScalarType.h:13:
In file included from /opt/venv/lib/python3.12/site-packages/torch/include/c10/util/complex.h:9:
In file included from /opt/venv/lib/python3.12/site-packages/_rocm_sdk_devel/include/thrust/complex.h:25:
In file included from /opt/venv/lib/python3.12/site-packages/_rocm_sdk_devel/include/thrust/detail/config.h:22:
In file included from /opt/venv/lib/python3.12/site-packages/_rocm_sdk_devel/include/thrust/detail/config/config.h:53:
/opt/venv/lib/python3.12/site-packages/_rocm_sdk_devel/include/thrust/detail/config/rtti.h:20:12: fatal error: 'cuda/std/__cccl/rtti.h' file not found
   20 | #  include _THRUST_STD_INCLUDE(__cccl/rtti.h)
      |            ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
/opt/venv/lib/python3.12/site-packages/_rocm_sdk_devel/include/thrust/detail/config/libcxx.h:41:41: note: expanded from macro '_THRUST_STD_INCLUDE'
   41 | #  define _THRUST_STD_INCLUDE(LIB)      <cuda/std/LIB>
      |                                         ^~~~~~~~~~~~~~
<scratch space>:146:1: note: expanded from here
  146 | <cuda/std/__cccl/rtti.h>
      | ^~~~~~~~~~~~~~~~~~~~~~~~

Describe the solution you'd like

This looks like it was added a while ago in CCCL (NVIDIA/cccl#2578), when do you think we can pull these changes in libhipcxx?

Describe alternatives you've considered

No response

Additional context

No response

[obersteiner]

This incompatibility with rocthrust was fixed in ROCm 7.1.1+.