diff --git a/include/ck/tensor_operation/gpu/device/impl/device_gemm_xdl_cshuffle_v3_b_preshuffle.hpp b/include/ck/tensor_operation/gpu/device/impl/device_gemm_xdl_cshuffle_v3_b_preshuffle.hpp
index cc2a55bea8..7ce78cd7c6 100644
--- a/include/ck/tensor_operation/gpu/device/impl/device_gemm_xdl_cshuffle_v3_b_preshuffle.hpp
+++ b/include/ck/tensor_operation/gpu/device/impl/device_gemm_xdl_cshuffle_v3_b_preshuffle.hpp
@@ -339,19 +339,21 @@ struct DeviceGemm_Xdl_CShuffleV3_BPreshuffle : public DeviceGemmV2BPreshuffle<AL
                     if(arg.KBatch > 1)
                     {
                         const auto kernel =
-                            kernel_gemm_xdl_cshuffle_v3<GridwiseGemm,
+                        kernel_gemm_xdl_cshuffle_v3_b_preshuffle<GridwiseGemm,
                                                         false,
                                                         InMemoryDataOperationEnum::AtomicAdd,
-                                                        minimum_occupancy>;
+                                                        minimum_occupancy,
+                                                        TailNumber::Odd>;
                         Run(kernel);
                     }
                     else
                     {
                         const auto kernel =
-                            kernel_gemm_xdl_cshuffle_v3<GridwiseGemm,
+                        kernel_gemm_xdl_cshuffle_v3_b_preshuffle<GridwiseGemm,
                                                         false,
                                                         InMemoryDataOperationEnum::Set,
-                                                        minimum_occupancy>;
+                                                        minimum_occupancy,
+                                                        TailNumber::Odd>;
                         Run(kernel);
                     }
                 }
diff --git a/include/ck/tensor_operation/gpu/thread/threadwise_tensor_slice_transfer.hpp b/include/ck/tensor_operation/gpu/thread/threadwise_tensor_slice_transfer.hpp
index 6eebed7319..37cbeeb3e9 100644
--- a/include/ck/tensor_operation/gpu/thread/threadwise_tensor_slice_transfer.hpp
+++ b/include/ck/tensor_operation/gpu/thread/threadwise_tensor_slice_transfer.hpp
@@ -377,14 +377,14 @@ struct ThreadwiseTensorSliceTransfer_v2
 
                     if constexpr(InvalidElementAsNaN)
                     {
-                        dst_buf(Number<dst_offset / PackedSize>{}) =
+                        dst_buf(Number<dst_offset>{}) =
                             is_src_valid
                                 ? type_convert<DstData>(src_vector.template AsType<SrcData>()[i])
                                 : NumericLimits<DstData>::QuietNaN();
                     }
                     else
                     {
-                        dst_buf(Number<dst_offset / PackedSize>{}) =
+                        dst_buf(Number<dst_offset>{}) =
                             type_convert<DstData>(src_vector.template AsType<SrcData>()[i]);
                     }
                 });
@@ -1619,7 +1619,7 @@ struct ThreadwiseTensorSliceTransfer_StaticToStatic
                     constexpr index_t src_offset = src_desc.CalculateOffset(
                         src_slice_origin_idx + idx_md + i * dst_scalar_step_in_vector);
 
-                    src_tmp_vector.template AsType<SrcData>()(i) = src_buf[Number<src_offset / PackedSize>{}];
+                    src_tmp_vector.template AsType<SrcData>()(i) = src_buf[Number<src_offset>{}];
                 });
 
                 // copy data from src_tmp_vector to dst_tmp_vector (data cast data from SrcData to