Performance of Hash Algorithms on GPUs for Use in Blockchain

The main cryptographic primitives in blockchain networks are hashing functions that are designed to form short and unpredictable digests for the message entered. In blockchain networks, hashing is used to build linked block lists, which provide safe and secure storage of important information in a distributed repository. The peculiarity of the hash search problem in blockchain networks allows applying the maximum parallelization of calculations, what good are multithreaded graphics processors (GPUs). In this paper, we explore the performance of GOST 34.311, STRIBOG, KECCAK, SHA2, RIPEMD160, Blake2b, and Whirlpool cryptographic hashing algorithms. HashCat software and various GPUs were used for comparative analysis of efficiency. GPUs were used: Geforce 740M 2GB; Geforce GTX1050ti 4GB; Rx580 Aorus 4GB; Rx580 Sapphire Pulse 8GB; Sapphire Vega 56 8GB.