<table border="1" cellspacing="0" cellpadding="8">
    <tr>
        <th>Issue</th>
        <td>
            <a href=https://github.com/llvm/llvm-project/issues/91584>91584</a>
        </td>
    </tr>

    <tr>
        <th>Summary</th>
        <td>
            [mlir][quant] Bump up the MaxStorageBits from 32 to 64.
        </td>
    </tr>

    <tr>
      <th>Labels</th>
      <td>
            mlir
      </td>
    </tr>

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

    <tr>
      <th>Reporter</th>
      <td>
          dansuh17
      </td>
    </tr>
</table>

<pre>
    64-bit storage type is often used for certain settings like on-device or when dealing with models that are quality-sensitive. For example, a TFLite micro kernel supports 64-bit quantized types for 16x8 quantized operations.

Currently the Quant dialect allows up to 32 bit storage bits: [MaxStorageBits](https://github.com/llvm/llvm-project/blob/b903badd73a2467fdd4e363231f2bf9b0704b546/mlir/include/mlir/Dialect/Quant/QuantTypes.h#L55). Bump this limit up to 64.
</pre>
<img width="1px" height="1px" alt="" src="http://email.email.llvm.org/o/eJyUks1urDgQhZ_GbEqNoLD5WbBIp8UqsxhNXsDgovHEYGIXSfo-_RV0ouQu78YlVdnW-eocHaO9LkStUGehLoneePKhNXqJ25RXSe_NrS3lqbcMkX3QVwK-rQQ2gh-ZFtgiGRh9gIECa7tAJGa7XCM4-0Lgl5OhNzsQ-ADvEy1gSDu7XOHd8gSzN-Qi8KQZdCB43bSzfDtFWqJl-0YpdD4Afeh5dSTwETQ8d0-WCWY7BA8vFBZyELd19YEjfGp93fTC9heZQ208BOblR_1j4FcKmq1fYiqyi8ge7ufjFgIt7G7AE8G_-3UwVjsaGLRz_j3CtgJ7KBB-bqW3HEXxAEKd_9Ef_927572pLgLriXnd5wI7gd3V8rT16eBngZ1zb1_ltAb_Pw0ssOud7_fSZEWvjakKjbKsRmMkFWWBRT5iPzZ9VmWyV7IU2M3OBoGdXQa3GfpuXO7qBXYHzVd93heTTgKLJ6UENimct3kFnuxu3Wz5k7OUaWLawjRFoxNq8ypXqEols2Rq67zJxlyZOiccColVXmqSWGBRj1hhk9gWM5SZypocs1rJlCppqCbEcRyqygxCZjRr69IdP_XhmtgYN2qbXNUycbonF49wIt5xcI9paI9t9ds1Cpk5Gzl-f8CW3RHo44G6CHU-XBfqckfcuSaCP22CMfh5N_UTeQuu_WvTDu1RYHfI_x0AAP__W2EbHw">