[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.

example/ck_tile/03_gemm/run_gemm_example.inc

@@ -30,7 +30,6 @@ auto calculate_rtol_atol(const ck_tile::index_t K,
         ck_tile::get_relative_threshold<CDataType, CDataType, CDataType>(kbatch);
     const auto atol_split_k = ck_tile::get_absolute_threshold<CDataType, CDataType, CDataType>(
         max_accumulated_value, kbatch);
-    // Use higher threshold
     return ck_tile::make_tuple(std::max(rtol, rtol_split_k), std::max(atol, atol_split_k));
 }
 
@@ -369,7 +368,9 @@ int run_gemm_example_with_layouts(ck_tile::ArgParser& arg_parser,
 
     std::size_t flop = std::size_t(2) * M * N * K;
     std::size_t num_byte =
-        sizeof(ADataType) * M * K + sizeof(BDataType) * N * K + sizeof(CDataType) * M * N;
+        sizeof(ADataType) * M * K / ck_tile::numeric_traits<ADataType>::PackedSize +
+        sizeof(BDataType) * N * K / ck_tile::numeric_traits<BDataType>::PackedSize +
+        sizeof(CDataType) * M * N;
     float tflops     = static_cast<float>(flop) / 1.E9 / ave_time;
     float gb_per_sec = num_byte / 1.E6 / ave_time;