Improve GPGPU Front-end Efficiency Via Inter-Warp Instruction Sharing | IEEE Conference Publication | IEEE Xplore