A Memory-Level Parallelism Aware Fetch Policy for SMT Processors | IEEE Conference Publication | IEEE Xplore