[AArch64] Missed optimization converting vector fmul(sitofp(x), 0.5) -> scvtf(x, 2)

[efriedma-quic]

In #91924, we removed the existing combine for vector fdiv(sitofp(x), 2) -> scvtf(x, 2) etc. because it would almost never trigger: instcombine converts all fdiv operations by powers of two to fmul. However, the idea behind the optimization is still useful: for the given conversion operations, we can eliminate the multiply. We should try to restore this optimization so it works on vector inputs. (This optimization is already implemented for scalar inputs.)
https://godbolt.org/z/c8qx7v4b4

Marking this good-first-issue because the code that was removed can be used as a starting point for this.

[llvmbot]

@llvm/issue-subscribers-backend-aarch64

Author: Eli Friedman (efriedma-quic)

In #91924, we removed the existing combine for vector `fdiv(sitofp(x), 2) -> scvtf(x, 2)` etc. because it would almost never trigger: instcombine converts all fdiv operations by powers of two to fmul. However, the idea behind the optimization is still useful: for the given conversion operations, we can eliminate the multiply. We should try to restore this optimization so it works on vector inputs. (This optimization is already implemented for scalar inputs.)

Marking this good-first-issue because the code that was removed can be used as a starting point for this.

[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: Eli Friedman (efriedma-quic)

In #91924, we removed the existing combine for vector `fdiv(sitofp(x), 2) -> scvtf(x, 2)` etc. because it would almost never trigger: instcombine converts all fdiv operations by powers of two to fmul. However, the idea behind the optimization is still useful: for the given conversion operations, we can eliminate the multiply. We should try to restore this optimization so it works on vector inputs. (This optimization is already implemented for scalar inputs.)

Marking this good-first-issue because the code that was removed can be used as a starting point for this.

[zjaffal]

can i get assigned to this issue?

[braw-lee]

@zjaffal are you still working on this,
or can I try it?

[braw-lee]

@efriedma-quic
hi, i was looking to work on this issue

"This optimization is already implemented for scalar inputs"
can you point to this pls? I can't seem to find it

[vortex73]

Looks fun. Could I be assigned if this is still an issue?

[dipeshs809]

Hi @efriedma-quic I would like to work on this, can you assign this to me ?

[jph-13]

I took a stab at this one. PR 141480