diff --git a/bench/tensors/tensor_benchmarks.h b/bench/tensors/tensor_benchmarks.h
index 325026113..3a640ede4 100644
--- a/bench/tensors/tensor_benchmarks.h
+++ b/bench/tensors/tensor_benchmarks.h
@@ -539,6 +539,11 @@ for (int iter = 0; iter < 10; ++iter) {
     if (Eigen::internal::is_same<Device, Eigen::GpuDevice>::value) {
       device_.synchronize();
     }
+#elif defined(EIGEN_USE_SYCL)
+    if (Eigen::internal::is_same<Device, Eigen::SyclDevice>::value) {
+      device_.synchronize();
+    }
+
 #endif
     StopBenchmarkTiming();
     SetBenchmarkFlopsProcessed(num_items);
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorConvolutionSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorConvolutionSycl.h
index 5db16d559..2e6021b1e 100644
--- a/unsupported/Eigen/CXX11/src/Tensor/TensorConvolutionSycl.h
+++ b/unsupported/Eigen/CXX11/src/Tensor/TensorConvolutionSycl.h
@@ -425,6 +425,7 @@ struct TensorEvaluator<const TensorConvolutionOp<Indices, InputArgType, KernelAr
         }
       }
     });
+    m_device.asynchronousExec();
   }
 
   EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE CoeffReturnType coeff(Index index) const