diff --git a/src/core/impl/tensor.cpp b/src/core/impl/tensor.cpp index ee01694073a0301dda3c4cca6c5558f1bf66e110..6bc4d950e84e20eea1fa4dc381d98a27f21bf17a 100644 --- a/src/core/impl/tensor.cpp +++ b/src/core/impl/tensor.cpp @@ -65,7 +65,7 @@ void noncont_tensor_copy( } DeviceTensorND tmp; tmp.copy_from(src); - dest.copy_from_fixlayout(tmp); + dest.copy_from_fixlayout(tmp).sync(); } //! implement non-contiguous h2d copy