Accelerate GPU Concurrent Kernel Execution by Mitigating Memory Pipeline Stalls | IEEE Conference Publication | IEEE Xplore