trivial refactors

include/ck/tensor_operation/gpu/block/blockwise_gemm_xdlops.hpp

@@ -701,9 +701,7 @@ struct BlockwiseGemmXdlops_v2
         const auto waveId_m = wave_idx[I0];
         const auto waveId_n = wave_idx[I1];
 
-        const auto tmp = xdlops_gemm.GetBeginOfThreadBlk(xdlops_i, blk_i);
-        const auto blk_idx =
-            TransposeC ? make_multi_index(tmp[I1], tmp[I0]) : make_multi_index(tmp[I0], tmp[I1]);
+        const auto blk_idx = xdlops_gemm.GetBeginOfThreadBlk(xdlops_i, blk_i);
 
         constexpr auto mrepeat_mwave_mperxdl_to_m_adaptor = make_single_stage_tensor_adaptor(
             make_tuple(make_unmerge_transform(make_tuple(MRepeat, MWaves, MPerXDL))),

include/ck/tensor_operation/gpu/warp/xdlops_gemm.hpp

@@ -819,7 +819,7 @@ struct XdlopsGemm
         index_t n_offset = blk_i * mfma_instr.n_per_blk + blk_td;
         index_t m_offset = xdlops_i * mfma_instr.m_per_blk + blk_id * mfma_instr.group_size;
 
-        return CIndex{m_offset, n_offset};
+        return TransposeC ? CIndex{n_offset, m_offset} : CIndex{m_offset, n_offset};
     }
 
     static constexpr auto mfma = MfmaSelector<base_type, MPerXdlops, NPerXdlops>{};