<table border="1" cellspacing="0" cellpadding="8">
    <tr>
        <th>Issue</th>
        <td>
            <a href=https://github.com/llvm/llvm-project/issues/98733>98733</a>
        </td>
    </tr>

    <tr>
        <th>Summary</th>
        <td>
            [aarch64] use sabd to optimize the float abs
        </td>
    </tr>

    <tr>
      <th>Labels</th>
      <td>
            question
      </td>
    </tr>

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

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

<pre>
    * test : https://gcc.godbolt.org/z/Gohebqbqe
```
float foo (float a, float b) {
    return abs (a - b);
}
```

the aarch64 target has **sabd** to address the **abs**, and S registers is alias to V registers, so would it possible to use **sabd** address the scalar float type of abs ?
</pre>
<img width="1px" height="1px" alt="" src="http://email.email.llvm.org/o/eJxsUt2OmzwQfZrhZrSRMwYcLrjIfhHfA1TqvY0NuHLirMdstfv0lQFpq7YSwjOjc-b3aGY_P5zroXmF5lbpNS8x9e-TnabKRPvRA10xO84I8opLzk8GeQUagIZ5HE9ztCaGfIppBho-gYb_4-LMm3lzIG4grtCK49vcKUSdcYoRgS67o4H-w900QB2Cet2xiIjJ5TU9UBsuBI0vGwbkAQF1-2eZ_Z8Xh1qncWlrzDrNLuOiS6Ir0JW1sbuFOaK2NjlmLJQ9qg3vRmlPPyx-w-Rmz9klRs-og9dcqN-_4gXKEX_GNVj0GZ-R2ZvgCmxl93fl38vyqINOxybyx9NhnPbB5VDZXtpOdrpy_VmRuLR1q2S19PXZ2ItqzqJum9po07YdjXRuzl2rJtdQ5XsSVAt1lqJrVCNPtVXKdUaJtpaCtIBauLv24RTC-73csfLMq-u7i5KyCtq4wJs8iN5Wx9nHBxAVsaS-UF7MOjPUInjO_JUk-xw2WR0HgOa2raAMX9YRn9nf_afbRj-EYLhaU-j_UJnPy2pOY7wDDSX98bw8U_zhxgw0bA0z0LD3_N7TrwAAAP__dd7dTg">