[Clang] VectorExprEvaluator::VisitCallExpr / InterpretBuiltin - add MMX/SSE/AVX/AVX512 PMADDWD/PMADDUBSW intrinsics to be used in constexpr

[RKSimon]

The PMADDWD/PMADDUBSW intrinsics can all be used in constexpr with suitable handling of the __builtin_ia32_pmadd* builtins inside VectorExprEvaluator::VisitCallExpr and InterBuiltin.cpp similar to #152540

_mm_madd_pi16
_mm_madd_epi16 _mm256_madd_epi16 _mm512_madd_epi16
_mm_mask_madd_epi16 _mm256_mask_madd_epi16 _mm512_mask_madd_epi16
_mm_maskz_madd_epi16 _mm256_maskz_madd_epi16 _mm512_maskz_madd_epi16

_mm_maddubs_pi16
_mm_maddubs_epi16 _mm256_maddubs_epi16 _mm512_maddubs_epi16
_mm_mask_maddubs_epi16 _mm256_mask_maddubs_epi16 _mm512_mask_maddubs_epi16
_mm_maskz_maddubs_epi16 _mm256_maskz_maddubs_epi16 _mm512_maskz_maddubs_epi16

The constant folding doesn't map to an APInt operation - you should handle it locally, plus its a horizontal pairwise addition so that will need handling as well:

  // Split Lo/Hi elements pairs, extend and add together.
  // PMADDWD(X,Y) =
  // add(mul(sext(lhs[0]),sext(rhs[0])),mul(sext(lhs[1]),sext(rhs[1])))
  // PMADDUBSW(X,Y) =
  // sadd_sat(mul(zext(lhs[0]),sext(rhs[0])),mul(zext(lhs[1]),sext(rhs[1])))

We might need to consider adjusting the PMADDUBSW builtins to take an unsigned lhs argument as well.

[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 is working on this issue. If someone is assigned to the issue or claimed to be working on it, ping the person. After one week without a response, the assignee may be changed.
2. Leave a comment indicating that you are working on the issue, 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-backend-x86

Author: Simon Pilgrim (RKSimon)

The PMADDWD/PMADDUBSW intrinsics can all be used in constexpr with suitable handling of the __builtin_ia32_pmadd* builtins inside VectorExprEvaluator::VisitCallExpr similar to https://github.com//pull/152540 ``` _mm_madd_pi16 _mm_madd_epi16 _mm256_madd_epi16 _mm512_madd_epi16 _mm_mask_madd_epi16 _mm256_mask_madd_epi16 _mm512_mask_madd_epi16 _mm_maskz_madd_epi16 _mm256_maskz_madd_epi16 _mm512_maskz_madd_epi16

_mm_maddubs_pi16
_mm_maddubs_epi16 _mm256_maddubs_epi16 _mm512_maddubs_epi16
_mm_mask_maddubs_epi16 _mm256_mask_maddubs_epi16 _mm512_mask_maddubs_epi16
_mm_maskz_maddubs_epi16 _mm256_maskz_maddubs_epi16 _mm512_maskz_maddubs_epi16

The constant folding doesn't map to an APInt operation - you should handle it locally, plus its a horizontal pairwise addition so that will need handling as well:

// Split Lo/Hi elements pairs, extend and add together.
// PMADDWD(X,Y) =
// add(mul(sext(lhs[0]),sext(rhs[0])),mul(sext(lhs[1]),sext(rhs[1])))
// PMADDUBSW(X,Y) =
// sadd_sat(mul(zext(lhs[0]),sext(rhs[0])),mul(zext(lhs[1]),sext(rhs[1])))

We might need to consider adjusting the PMADDUBSW builtins to take an unsigned lhs argument as well.
</details>

[llvmbot]

@llvm/issue-subscribers-clang-frontend

Author: Simon Pilgrim (RKSimon)

The PMADDWD/PMADDUBSW intrinsics can all be used in constexpr with suitable handling of the __builtin_ia32_pmadd* builtins inside VectorExprEvaluator::VisitCallExpr similar to https://github.com//pull/152540 ``` _mm_madd_pi16 _mm_madd_epi16 _mm256_madd_epi16 _mm512_madd_epi16 _mm_mask_madd_epi16 _mm256_mask_madd_epi16 _mm512_mask_madd_epi16 _mm_maskz_madd_epi16 _mm256_maskz_madd_epi16 _mm512_maskz_madd_epi16

_mm_maddubs_pi16
_mm_maddubs_epi16 _mm256_maddubs_epi16 _mm512_maddubs_epi16
_mm_mask_maddubs_epi16 _mm256_mask_maddubs_epi16 _mm512_mask_maddubs_epi16
_mm_maskz_maddubs_epi16 _mm256_maskz_maddubs_epi16 _mm512_maskz_maddubs_epi16

The constant folding doesn't map to an APInt operation - you should handle it locally, plus its a horizontal pairwise addition so that will need handling as well:

// Split Lo/Hi elements pairs, extend and add together.
// PMADDWD(X,Y) =
// add(mul(sext(lhs[0]),sext(rhs[0])),mul(sext(lhs[1]),sext(rhs[1])))
// PMADDUBSW(X,Y) =
// sadd_sat(mul(zext(lhs[0]),sext(rhs[0])),mul(zext(lhs[1]),sext(rhs[1])))

We might need to consider adjusting the PMADDUBSW builtins to take an unsigned lhs argument as well.
</details>

[llvmbot]

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

Author: Simon Pilgrim (RKSimon)

The PMADDWD/PMADDUBSW intrinsics can all be used in constexpr with suitable handling of the __builtin_ia32_pmadd* builtins inside VectorExprEvaluator::VisitCallExpr similar to https://github.com//pull/152540 ``` _mm_madd_pi16 _mm_madd_epi16 _mm256_madd_epi16 _mm512_madd_epi16 _mm_mask_madd_epi16 _mm256_mask_madd_epi16 _mm512_mask_madd_epi16 _mm_maskz_madd_epi16 _mm256_maskz_madd_epi16 _mm512_maskz_madd_epi16

_mm_maddubs_pi16
_mm_maddubs_epi16 _mm256_maddubs_epi16 _mm512_maddubs_epi16
_mm_mask_maddubs_epi16 _mm256_mask_maddubs_epi16 _mm512_mask_maddubs_epi16
_mm_maskz_maddubs_epi16 _mm256_maskz_maddubs_epi16 _mm512_maskz_maddubs_epi16

The constant folding doesn't map to an APInt operation - you should handle it locally, plus its a horizontal pairwise addition so that will need handling as well:

// Split Lo/Hi elements pairs, extend and add together.
// PMADDWD(X,Y) =
// add(mul(sext(lhs[0]),sext(rhs[0])),mul(sext(lhs[1]),sext(rhs[1])))
// PMADDUBSW(X,Y) =
// sadd_sat(mul(zext(lhs[0]),sext(rhs[0])),mul(zext(lhs[1]),sext(rhs[1])))

We might need to consider adjusting the PMADDUBSW builtins to take an unsigned lhs argument as well.
</details>

[markbhasawut]

Hi @RKSimon, I can work on this one. Please assign this to me.

[RKSimon]

@markbhasawut ping - any progress?

[markbhasawut]

@RKSimon Sorry, I was busy last week. I'm still working on it.