diff --git a/Common/include/linear_algebra/CMatrixVectorProduct.hpp b/Common/include/linear_algebra/CMatrixVectorProduct.hpp
index 878bb132b98..e6baca2acf1 100644
--- a/Common/include/linear_algebra/CMatrixVectorProduct.hpp
+++ b/Common/include/linear_algebra/CMatrixVectorProduct.hpp
@@ -82,7 +82,11 @@ class CSysMatrixVectorProduct final : public CMatrixVectorProduct<ScalarType> {
    */
   inline CSysMatrixVectorProduct(const CSysMatrix<ScalarType>& matrix_ref, CGeometry* geometry_ref,
                                  const CConfig* config_ref)
-      : matrix(matrix_ref), geometry(geometry_ref), config(config_ref) {}
+      : matrix(matrix_ref), geometry(geometry_ref), config(config_ref) {
+#ifdef HAVE_CUDA
+    if (config->GetCUDA()) matrix.HtDTransfer();
+#endif
+  }
 
   /*!
    * \note This class cannot be default constructed as that would leave us with invalid pointers.
diff --git a/Common/src/linear_algebra/CSysMatrixGPU.cu b/Common/src/linear_algebra/CSysMatrixGPU.cu
index 90389264ed3..7e0c81ca54f 100644
--- a/Common/src/linear_algebra/CSysMatrixGPU.cu
+++ b/Common/src/linear_algebra/CSysMatrixGPU.cu
@@ -70,7 +70,6 @@ void CSysMatrix<ScalarType>::GPUMatrixVectorProduct(const CSysVector<ScalarType>
    ScalarType* d_vec = vec.GetDevicePointer();
    ScalarType* d_prod = prod.GetDevicePointer();
 
-   HtDTransfer();
    vec.HtDTransfer();
    prod.GPUSetVal(0.0);