CudaDMA: optimizing GPU memory bandwidth via warp specialization. (2011)

by M Bauer, H Cook, B Khailany
Venue:In Proceedings of 2011 International Conference for High Performance Computing, Networking, Storage and Analysis, SC ’11.