Methods and apparatus to perform direct cache access in multiple core processors are described. In an embodiment, data corresponding to a direct cache access request is stored in a storage unit and a corresponding read request is generated. Other embodiments are also described.