Exploiting Memory-Level Parallelism in Reconfigurable Accelerators | IEEE Conference Publication | IEEE Xplore