Is "micronode" a new GPU concept, or did they mean "microcode" in the bulletin?