LA-LLC: Inter-Core Locality-Aware Last-Level Cache to Exploit Many-to-Many Traffic in GPGPUs | IEEE Journals & Magazine | IEEE Xplore