<table border="1" cellspacing="0" cellpadding="8">
    <tr>
        <th>Issue</th>
        <td>
            <a href=https://github.com/llvm/llvm-project/issues/90847>90847</a>
        </td>
    </tr>

    <tr>
        <th>Summary</th>
        <td>
            Regression in codegen for some SIMD operations
        </td>
    </tr>

    <tr>
      <th>Labels</th>
      <td>
            new issue
      </td>
    </tr>

    <tr>
      <th>Assignees</th>
      <td>
      </td>
    </tr>

    <tr>
      <th>Reporter</th>
      <td>
          jfalcou
      </td>
    </tr>
</table>

<pre>
    While checking for code generation quality of EVE, our SIMD wrapper, I encountered this issue between clang 15 and 18 

https://godbolt.org/z/daxdhvbTx

After a [short discussion on Discord](https://discord.com/channels/636084430946959380/636732781086638081/1235396822483206206), we ended up with those more streamlined repro:

[Sources](https://godbolt.org/z/8vqqPGzsq)
[Generated IR 1][https://godbolt.org/z/dxbzePxMo)
[Generated IR 2](https://godbolt.org/z/9dPTf7oad)
</pre>
<img width="1px" height="1px" alt="" src="http://email.email.llvm.org/o/eJyMk09vqzoQxT-N2YwaGRuIvWDRvJSqi0pVW723NvYA7nPsxDZN2k9_RdKr2_tP925AGjRnzhx-o1Kyo0dsSb0h9bZQc55CbF8G5XSYiz6Yt_a_yToEPaH-3_oRhhBBB4Mwoseosg0eDrNyNr9BGODm3xvC_oEwR3i6u9_CMar9HuNSuwP0Osw-Y0QDebIJbEozQo_5iOhBO-VHKGtQ3kApgNAtodeX55TzPhF-TVhHWDcG0weXVyGOhHXvhHVGncz02j-fPjddDxkjKCD1Jk0hZjA26TmlxXPwsLVJh2hIvSVMfD_AXD6tdNgR1ulJeY8uEdY1vKGiqjiVVSNryQW9FNecrUVJRdNwQUVJWFcyXnPZCMYqwRltGG0Ik0sQRwT0Bg3MezjaPEGeQkLYhYiQckS1c9ajgYj7GBZLn1Yi9eYpzFFj-pXtn3MRr4fDw-17OiyzvyrcXn4dGrh7hHIRqjd_DPjUv-PD6T78Voj9nSNpHp6HdVCGMFmYlhvJpSqwLddl1VRClqKYWl1RXjdKG6WHUslBSYqca0mZqOma14VtGWUVrSkrS1bWYrUWCkuUQ1X1sue8JhXFnbJu5dzrbhlfnGFrJRXVunCqR5fO2DPm8XghkbBlhyK2S89VP4-JVNTZlNM3lWyzw_YRx4gXkKw_38OI_nwbKezwgn7Yf9xHKubo2h9ysXma-w--FvGP19U-hhfUmbDubGlh7mz5SwAAAP__PkskHw">