<table border="1" cellspacing="0" cellpadding="8">
    <tr>
        <th>Issue</th>
        <td>
            <a href=https://github.com/llvm/llvm-project/issues/65600>65600</a>
        </td>
    </tr>

    <tr>
        <th>Summary</th>
        <td>
            VLA / VLS  options for RVV auto-vectorization seems not work
        </td>
    </tr>

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

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

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

<pre>
    I tried some examples in [Auto-Vectorization in LLVM](https://llvm.org/docs/Vectorizers.html)  and I found using `VLA / VLS ` options for RVV auto-vectorization, the assembly code generated didn't change. Does it normal for LLVM? or I miss something? 

I use: 

- `-march=rv64gcv -mllvm --riscv-v-vector-bits-min=256` for generating VLS code
- `-march=rv64gcv -mllvm --scalable-vectorization=on` for generating VLA code

you can see https://godbolt.org/z/7YqWxnTzh.



</pre>
<img width="1px" height="1px" alt="" src="http://email.email.llvm.org/o/eJyMU8tu2zAQ_BrqspBBUy_roINTQYAB99IWLnqkyJXEhiJdknLifH1BxUGdoIdcVhCJHc7MznLv1WgQG1I8kKJN-BIm65rzYh6vIumtvDYHCE6hBG9nBHzm81mjB2WAFA_7Jdj0hCJYp154UNbEi-Px9JUULWG7KYSzJ9mesI6wTuvLvLFuJKyTVnjCurdWdH4zhVkTVgNwI-EAg12MhMUrMwIp6em4B8I6OB2_x1-w5_iah8E6-HY6AY9MLvdMCPsCYULg3uPc6ysIKxFGNOh4QAlSSUNYFUBM3Iy4gdZGXQGMdTPXK_KqJOvAOjjArLxfXQiTMmM8JrQldP9aD7B4JNn-3WEauaYzd2IiWesuZT6KC6RzdALS1CkvLunlxjvtVfDprAzJWlaUUWXkcGMcfYjio4pPgXvBNe81fnAla635L_T-Dvq1Xu0CghvwiPB-kqOVvdXhNswXwrrq15-fz-bHy7S5R3itiWwyWWc1T7DZlnVOa8ponkxNz1k_ULor66rkgrKMD1WPKDgbdih3u0Q1jLKM1rTasqzM8w3bMpnJodpWQ4G5zElOceZKb96ylSjvF2zKoqQ00bxH7ddsM2bwCdZLwliMumtiT9ovoyc51coH_w8lqKCxuQ_dJxIXjZo9GBvgybrHZHG6-eCbCtPSb4Sdb-tw-6RnZ3-jCIR1K8W4G6uEvwEAAP__Kl4pag">