With embarrass of the weakness in real-time requirements of ECC (Elliptic Curve Cryptograph). Streaming ECC with optimization was proposed, including streaming ECC for parallelization using thousands of threads as well as stream optimization with best utilization of memory hierarchy of CUDA-enable GPU. Experimental results show that prototype of ECC implemented using CUDA (Computing Unified Device Architecture) on NVIDIA’s GTX280 could achieve as high as 66×speedup than CPU counterpart available. Furthermore, above proposed techniques including streaming for parallelization and optimizations with memory hierarchy could be generalized for other streaming architectures.
