Review: Remove unnecessary empty lines

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

@@ -72,22 +72,16 @@ struct BlockwiseGemmDpp_k0mk1_k0nk1_m0n0m1n1m2n2
     __device__ static auto CalculateAThreadOriginDataIndex()
     {
         const auto wave_idx  = GetWaveIdx();
-
         const auto waveId_m  = wave_idx[I0];
-
         const auto dpp_a_idx = dpp_gemm.CalculateAThreadOriginDataIndex();
-
         return make_tuple(0, waveId_m, dpp_a_idx[I1], KPerThread * dpp_a_idx[I0]);
     }
 
     __device__ static auto CalculateBThreadOriginDataIndex()
     {
         const auto wave_idx  = GetWaveIdx();
-
         const auto waveId_n  = wave_idx[I1];
-
         const auto dpp_b_idx = dpp_gemm.CalculateBThreadOriginDataIndex();
-
         return make_tuple(0, waveId_n, dpp_b_idx[I1], KPerThread * dpp_b_idx[I0]);
     }
 
@@ -95,10 +89,8 @@ struct BlockwiseGemmDpp_k0mk1_k0nk1_m0n0m1n1m2n2
     __device__ static auto CalculateCThreadOriginDataIndex(Number<m0>, Number<n0>)
     {
         const auto wave_idx = GetWaveIdx();
-
         const auto waveId_m = wave_idx[I0];
         const auto waveId_n = wave_idx[I1];
-
         const auto blk_idx  = dpp_gemm.GetBeginOfThreadBlk();
 
         constexpr auto mrepeat_mwave_MPerDpp_to_m_adaptor = make_single_stage_tensor_adaptor(
@@ -141,7 +133,6 @@ struct BlockwiseGemmDpp_k0mk1_k0nk1_m0n0m1n1m2n2
     __host__ __device__ static constexpr auto GetCThreadDescriptor_M0_N0_M1_N1_M2_N2()
     {
         constexpr auto c_m0_m1_m2_n_tblk_lens = dpp_gemm.GetCMNThreadBlkLengths();
-
         constexpr auto M0                     = c_m0_m1_m2_n_tblk_lens[I0];
         constexpr auto N                      = c_m0_m1_m2_n_tblk_lens[I1];
 
@@ -152,7 +143,6 @@ struct BlockwiseGemmDpp_k0mk1_k0nk1_m0n0m1n1m2n2
     __host__ __device__ static constexpr auto GetCThreadDescriptor_G_M0_N0_M1_N1_M2_N2()
     {
         constexpr auto c_m0_m1_m2_n_tblk_lens = dpp_gemm.GetCMNThreadBlkLengths();
-
         constexpr auto M0                     = c_m0_m1_m2_n_tblk_lens[I0];
         constexpr auto N                      = c_m0_m1_m2_n_tblk_lens[I1];
 

include/ck/tensor_operation/gpu/grid/gridwise_gemm_dpp.hpp

@@ -243,14 +243,11 @@ struct GridwiseGemm_k0mk1_k0nk1_mn_dpp
     {
         // LDS allocation for A and B: be careful of alignment
         constexpr auto a_block_desc_k0_m_k1 = GetABlockDescriptor_K0PerBlock_MPerBlock_K1();
-
         constexpr auto b_block_desc_k0_n_k1 = GetBBlockDescriptor_K0PerBlock_NPerBlock_K1();
-
         constexpr auto max_lds_align        = K1;
 
         constexpr auto a_block_space_size_aligned =
             math::integer_least_multiple(a_block_desc_k0_m_k1.GetElementSpaceSize(), max_lds_align);
-
         constexpr auto b_block_space_size_aligned =
             math::integer_least_multiple(b_block_desc_k0_n_k1.GetElementSpaceSize(), max_lds_align);
 
@@ -539,7 +536,6 @@ struct GridwiseGemm_k0mk1_k0nk1_mn_dpp
         // HACK: this force m/n_block_data_idx_on_grid into SGPR
         const index_t m_block_data_idx_on_grid =
             __builtin_amdgcn_readfirstlane(block_work_idx[I0] * MPerBlock);
-
         const index_t n_block_data_idx_on_grid =
             __builtin_amdgcn_readfirstlane(block_work_idx[I1] * NPerBlock);
 
@@ -547,7 +543,6 @@ struct GridwiseGemm_k0mk1_k0nk1_mn_dpp
 
         // A matrix in LDS memory, dst of blockwise copy
         constexpr auto a_block_desc_k0_m_k1 = GetABlockDescriptor_K0PerBlock_MPerBlock_K1();
-
         // B matrix in LDS memory, dst of blockwise copy
         constexpr auto b_block_desc_k0_n_k1 = GetBBlockDescriptor_K0PerBlock_NPerBlock_K1();