[rocm-libraries] ROCm/rocm-libraries#4594 (commit 1fce4cb)

[CK_TILE] MX GEMM non-preshuffled RCR layout

## Motivation

Implements a GEMM with MX scaling for fp4 and fp8 in non-preshuffled
layouts using async pipeline.

## Technical Details

<!-- Explain the changes along with any relevant GitHub links. -->

## Test Plan

<!-- Explain any relevant testing done to verify this PR. -->

## Test Result

<!-- Briefly summarize test outcomes. -->

## Submission Checklist

- [ ] Look over the contributing guidelines at
https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests.

include/ck_tile/core/container/tuple.hpp

@@ -287,6 +287,7 @@ struct tuple : impl::tuple_base<make_index_sequence<sizeof...(T)>, T...>
     template<index_t I> CK_TILE_HOST_DEVICE constexpr decltype(auto) operator[](number<I>)             { TP_COM_(); return get<I>(); }
     template<index_t I> CK_TILE_HOST_DEVICE constexpr decltype(auto) operator[](number<I>) const { TP_COM_(); return get<I>(); }
     template<index_t I> CK_TILE_HOST_DEVICE constexpr decltype(auto) operator()(number<I>)             { TP_COM_(); return get<I>(); }  // TODO: compatible
+    template<index_t I> CK_TILE_HOST_DEVICE constexpr decltype(auto) operator()(number<I>) const { TP_COM_(); return get<I>(); }
 
     // below function should be used under tuple_array<> type, no extra check will perform here
     template <typename Tx> CK_TILE_HOST_DEVICE constexpr decltype(auto) get_as()                            { return reinterpret_cast<tuple_array<Tx, size()>&>(*this); }

include/ck_tile/core/tensor/static_distributed_tensor.hpp

@@ -78,6 +78,7 @@ struct static_distributed_tensor
         constexpr auto sliced_thread_tensor_desc =
             make_naive_tensor_descriptor_packed(make_tuple(YSliceLengths...));
 
+        // divide element number by PackedSize to get the correct thread buffer size
         thread_buffer<DataType, sliced_thread_tensor_desc.get_element_space_size() / PackedSize>
             sliced_thread_data;