Comparison and optimization of gpu performances on ethereum mining

Bitcoin, which emerged in 2008 with the article "Bitcoin: A Peer-to-Peer Electronic Cash System", is increasing its popularity day by day by adding new terms to scientific literature. Developed as an alternative cryptocurrency, Ethereum is a system that allows for anonymous and inexpensive asset transfer between the opposite ends without the need for a central server or a checkpoint. Bitcoin and alternative crypto coins are the most common use of blockchain structure. Crypto coins are realized by mining which is the solution of a challenging algorithm. Any miner in the system uses the graphics card and processor power in hand to obtain this solution. As the processors of the GPUs are more efficient in floating point calculations, the performances of the various graphics cards used in Ethereum mining will be focused on in this study. The model and memory quantities of the graphics cards used are as follows: 4 GB AMD R9 290, 4 GB AMD R9 380, 8 GB AMD R9 390, 4 GB AMD RX 560, 4 GB AMD RX 470, 4 GB AMD RX 480, 4 GB AMD RX 570, 4 GB AMD RX 580, 6 GB NVIDIA GTX1060, 8 GB NVIDIA GTX1070, 8 GB NVIDIA GTX1080. The most important aim of this study is to apply optimization methods in the Ethash summary algorithm which is specific for Ethereum mining. For this purpose, improvements have been made to enable AMD and NVIDIA chipset graphics cards to solve faster algorithms with lower power consumption. When the graphics cards were operated with the factory default settings, none of them could profit from the current exchange rates. Reducing the power consumption along with increasing the speed of the algorithm has been an important factor in the transition to the profit zone. As a result of the studies, a performance increase of 30% is observed. Also, the electrical energy consumed by the graphics cards used for Ethereum mining is reduced by 20 to 30 percent. Keywords: Ethereum, Blockchain, Cryptocurrency mining, GPU mining.