ALEXANDRIA, Va., July 16 -- United States Patent no. 12,670,357, issued on June 30, was assigned to Intel Corp. (Santa Clara, Calif.).
"Near memory sparse matrix computation in deep neural network" was invented by Srivatsa Rangachar Srinivasa (Hillsboro, Ore.), Jainaveen Sundaram Priya (Hillsboro, Ore.), Bradley A. Jackson (Lake Oswego, Ore.), Ambili Vengallur (Bangalore, India), Dileep John Kurian (Bangalore, India) and Tanay Karnik (Portland, Ore.).
According to the abstract* released by the U.S. Patent & Trademark Office: "A DNN accelerator includes a multiplication controller controlling whether to perform matrix computation based on weight values. The multiplication controller reads a weight matrix from a WRAM in the DNN accelerator ...