Balancing Scalar and Vector Execution on GPU Architectures | IEEE Conference Publication | IEEE Xplore