diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorDeviceSycl.h b/unsupported/Eigen/CXX11/src/Tensor/TensorDeviceSycl.h
index 96c95e294..44f4f33bc 100644
--- a/unsupported/Eigen/CXX11/src/Tensor/TensorDeviceSycl.h
+++ b/unsupported/Eigen/CXX11/src/Tensor/TensorDeviceSycl.h
@@ -218,7 +218,7 @@ struct SyclDevice {
     sycl_queue().submit([&](cl::sycl::handler &cgh) {
       auto src_acc =it1->second.template get_access<cl::sycl::access::mode::read, cl::sycl::access::target::global_buffer>(cgh);
       auto dst_acc =it2->second.template get_access<cl::sycl::access::mode::discard_write, cl::sycl::access::target::global_buffer>(cgh);
-      cgh.parallel_for(cl::sycl::nd_range<1>(cl::sycl::range<1>(GRange), cl::sycl::range<1>(tileSize)), MemCopyFunctor<T>(src_acc, dst_acc, rng, 0, offset));
+      cgh.parallel_for(cl::sycl::nd_range<1>(cl::sycl::range<1>(GRange), cl::sycl::range<1>(tileSize)), MemCopyFunctor<T>(src_acc, dst_acc, rng, i, offset));
     });
     asynchronousExec();
   }
@@ -249,7 +249,7 @@ struct SyclDevice {
     sycl_queue().submit([&](cl::sycl::handler &cgh) {
       auto src_acc= it->second.template get_access<cl::sycl::access::mode::read, cl::sycl::access::target::global_buffer>(cgh);
       auto dst_acc =dest_buf.template get_access<cl::sycl::access::mode::discard_write, cl::sycl::access::target::global_buffer>(cgh);
-      cgh.parallel_for( cl::sycl::nd_range<1>(cl::sycl::range<1>(GRange), cl::sycl::range<1>(tileSize)), MemCopyFunctor<T>(src_acc, dst_acc, rng, 0, 0));
+      cgh.parallel_for( cl::sycl::nd_range<1>(cl::sycl::range<1>(GRange), cl::sycl::range<1>(tileSize)), MemCopyFunctor<T>(src_acc, dst_acc, rng, 0, offset));
     });
     asynchronousExec();
   }