Oh, I don't think at all that this method of operation (Tomasulo?) is remotely close to a real dataflow in terms of efficiency and complexity. But as I wrote, dataflow is another story ;)