[NVPTX] Port code to `llvm/lib/CodeGen/SelectionDAG/*`

[justinfargnoli]

Some of NVPTX's custom operation lowering and custom DAG combines do not rely on NVPTXISD nodes. A majority of these operation lowerings and DAG combines can be ported to live in LegalizeDAG.cpp or DAGCombiner.cpp respectively.

This will increase the coverage we have on this code and enable other targets to benefit from the work we've done.

For example,

• Custom Lowering
  • bitcast
  • concat_vector
  • int -> fp / fp -> int
• Custom DAG Combines
  • rem

[llvmbot]

Hi!

This issue may be a good introductory issue for people new to working on LLVM. If you would like to work on this issue, your first steps are:

1. Check that no other contributor has already been assigned to this issue. If you believe that no one is actually working on it despite an assignment, ping the person. After one week without a response, the assignee may be changed.
2. In the comments of this issue, request for it to be assigned to you, or just create a pull request after following the steps below. Mention this issue in the description of the pull request.
3. Fix the issue locally.
4. Run the test suite locally. Remember that the subdirectories under test/ create fine-grained testing targets, so you can e.g. use make check-clang-ast to only run Clang's AST tests.
5. Create a Git commit.
6. Run git clang-format HEAD~1 to format your changes.
7. Open a pull request to the upstream repository on GitHub. Detailed instructions can be found in GitHub's documentation. Mention this issue in the description of the pull request.

If you have any further questions about this issue, don't hesitate to ask via a comment in the thread below.

[llvmbot]

@llvm/issue-subscribers-good-first-issue

Author: Justin Fargnoli (justinfargnoli)

Some of NVPTX's custom operation lowering and custom DAG combines do not rely on [`NVPTXISD` nodes](https://github.com/llvm/llvm-project/blob/6dceb0e34ed3dd4be72d211abb8c9c447bd57735/llvm/lib/Target/NVPTX/NVPTXISelLowering.h#L22). A majority of these operation lowerings and DAG combines can be ported to live in [`LegalizeDAG.cpp`](https://github.com/llvm/llvm-project/blob/main/llvm/lib/CodeGen/SelectionDAG/LegalizeDAG.cpp) or [`DAGCombiner.cpp`](https://github.com/llvm/llvm-project/blob/main/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp) respectively.

This will increase the coverage we have on this code and enable other targets to benefit from the work we've done.

For example,

• Custom Lowering
  • bitcast
  • concat_vector
  • int -> fp / fp -> int
• Custom DAG Combines
  • rem

[actinks]

I am interested in this issue, please assign it to me. Thank you @justinfargnoli @EugeneZelenko

[justinfargnoli]

@QiYueFeiXue, I'd recommend claiming a particular function in NVPTXISelLowering.cpp. This way, we can parallelize this task by having multiple people work on different functions.

[AidanGoldfarb]

I would like to work on int -> fp / fp -> int, could it be assigned to me?

[Artem-B]

Regarding int/fp conversion, #74827 which introduced that code explicitly mentions that There doesn't seem to be a way to express this promotion directly within the legalization framework, so fallback on Custom lowering.

This one may potentially be a bit more involved than it looks on the surface.