implement variants of cuda_invoke for all three types of launch configs

src/pystencilssfg/composer/basic_composer.py

src/pystencilssfg/ir/call_tree.py

src/pystencilssfg/ir/postprocessing.py

src/pystencilssfg/lang/cpp/std_tuple.py

src/pystencilssfg/lang/cuda.py

src/pystencilssfg/lang/expressions.py