Congestion
发表于 2025-3-30 11:58:23
http://reply.papertrans.cn/16/1598/159759/159759_51.png
进取心
发表于 2025-3-30 14:51:39
http://reply.papertrans.cn/16/1598/159759/159759_52.png
阉割
发表于 2025-3-30 18:56:37
https://doi.org/10.1007/978-981-19-2169-8t) scheme of SM4 is proposed by exploiting the structure property. Moreover, this scheme is further improved by introducing the page-locked memory and CUDA streams. The results show that: SM4 optimized parallel implementation under GPU can obtain with a speed-up ratio of 89, and the throughput can reach up to 31.41 Gbps.