<table border="1" cellspacing="0" cellpadding="8">
    <tr>
        <th>Issue</th>
        <td>
            <a href=https://github.com/llvm/llvm-project/issues/78641>78641</a>
        </td>
    </tr>

    <tr>
        <th>Summary</th>
        <td>
            [flang] MOD 7x performance regression
        </td>
    </tr>

    <tr>
      <th>Labels</th>
      <td>
            performance,
            flang:runtime
      </td>
    </tr>

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

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

<pre>
    Polyhedron/ac slowed down by 3.7x after https://github.com/llvm/llvm-project/pull/77167
`_FortranAModReal8` itself slowed down by 7.4x

I wonder if it can use the previous approach for not so different `a` and `p`.
</pre>
<img width="1px" height="1px" alt="" src="http://email.email.llvm.org/o/eJyUkcGu2ywUhJ8Gb45-C7AD9sKL_I0sdXHVqi9QgX2w6SVgHfC9SZ--cpqqVXfdwNnMjGY-k7NfIuLATv8zKV-D2XNAYlKy06Uye1kTDW_fzetqyFc2zffhcwr3FWdKkcnRTJBDescZ5vQewd6hqfUNjCtIsJayZdacmRyZHBdf1t3WU7oyOYbw9uv7b6P0DafC5LjtITA5ai2UZvzC-Jkp_nVMVMjE80uav6AJHVMcfMkY3N_Zum5vT93j_QjvKc5I4B34ApOJsGeEsiJshG8-7RnMtlEy0wouEcRUICeYvXNIGAswxc2RZ-J83BtTvK7moZn7pjcVDkLzkxK9anm1DtpyK_rJCtOha53shUTbWiXcqXf91FR-kFy2XIhONqLlvJ44Nm0nrZC800pp1nK8Gh_qY5g60VL5nHccdKdaUQVjMeQnqg3JJbqaOOFBS35gUrpg4sKaM-2x-Cs-KdLwmNnuS2YtDz6X_Nu_-BIe9H9qTxd4-XQBfYM__IFwIczZp1jtFIZ_5vookQ-yR48fAQAA__9eXsjV">