diff --git a/cpp/apps/CMakeLists.txt b/cpp/apps/CMakeLists.txt
index 1d12f58cc24..0beebfc9881 100644
--- a/cpp/apps/CMakeLists.txt
+++ b/cpp/apps/CMakeLists.txt
@@ -141,4 +141,8 @@ if (BUILD_GUI)
     open3d_add_app_gui(Open3DViewer Open3D Open3DViewer)
 endif()
 
-open3d_add_app_common(OfflineReconstruction OfflineReconstruction OfflineReconstruction)
+if (BUILD_SYCL_MODULE)
+    open3d_add_app_common(OfflineReconstruction OfflineReconstruction OfflineReconstruction Open3D::3rdparty_sycl)
+else()
+    open3d_add_app_common(OfflineReconstruction OfflineReconstruction OfflineReconstruction)
+endif()
diff --git a/cpp/benchmarks/CMakeLists.txt b/cpp/benchmarks/CMakeLists.txt
index 89178ce97cd..3d7e1f38833 100644
--- a/cpp/benchmarks/CMakeLists.txt
+++ b/cpp/benchmarks/CMakeLists.txt
@@ -14,6 +14,9 @@ if (BUILD_CUDA_MODULE)
     find_package(CUDAToolkit REQUIRED)
     target_link_libraries(benchmarks PRIVATE CUDA::cudart)
 endif()
+if (BUILD_SYCL_MODULE)
+    target_link_libraries(benchmarks PRIVATE Open3D::3rdparty_sycl)
+endif()
 
 open3d_show_and_abort_on_warning(benchmarks)
 open3d_set_global_properties(benchmarks)
diff --git a/docker/test2.png b/docker/test2.png
new file mode 100644
index 00000000000..f03f508be40
Binary files /dev/null and b/docker/test2.png differ
diff --git a/docs/sycl.rst b/docs/sycl.rst
index bfb047cddbf..4e9c20ed446 100644
--- a/docs/sycl.rst
+++ b/docs/sycl.rst
@@ -19,16 +19,15 @@ compiler option) CMake variables. See the `compiler documentation
 <https://github.com/intel/llvm/blob/sycl/sycl/doc/UsersManual.md>`_ for
 information about building for specific hardware.
 
-Current status:
----------------
+Enabled features:
+-----------------
 
-Most Tensor API operations and tensor geometry operations without custom kernels
+Many Tensor API operations and Tensor Geometry operations without custom kernels
 can now be offloaded to SYCL devices. In addition, HW accelerated raycasting
 queries in :ref:`RayCastingScene` are also supported. You will get an error if
-an operation is not supported. We are working on adding more custom kernels and
-improving performance. The implementation is tested on Linux on Intel integrated
-and discrete GPUs. Currently, a single GPU (`SYCL:0`, if available) and the CPU
-(`SYCL:1` if a GPU is available, else `SYCL:0`) are supported.
+an operation is not supported. The implementation is tested on Linux on Intel
+integrated and discrete GPUs. Currently, a single GPU (`SYCL:0`, if available)
+and the CPU (`SYCL:1` if a GPU is available, else `SYCL:0`) are supported.
 
 Installation:
 -------------