Electrical energy aware parallel and distributed computing

Recently, the world has faced the problem – growing of electric energy consumption. It is not the only one problem. The human race used and uses non-regenerating land resources very carelessly. A very god example is the production of electricity. There are several methods of the production of electricity: thermal power plants, nuclear power plants, hydropower plants, wind power plants, and solar batteries. And only three of them use regenerating resources, namely only wind power, hydropower plants, and solar batteries. All the other electric power generation methods use not only non-regenerating resources, but also polluting and dangerous to the environment. It is necessary to reduce the power consumption of equipment. To this end, we need to reduce the scale of power generation and make electric power generation less influential on people and environment. Electric power consumption is also applied to parallel and distributed computing. The number of supercomputers, clusters, grid and cloud computers is growing up in the world every year. All of these resources require a lot of electricity. Same applies to the most powerful computer in the world, \"K computer\". It requires 12 659.89 kW of electricity when running at full power. That much power would be sufficient to maintain about 60 000 usual households. As a result, the term green computing has emerged. San Murugesan defines the field of green computing as “the study and practice of designing, manufacturing, using, and disposing of computers, servers and associated subsystems – such as monitors, printers, storage devices and networking and communications systems – efficiently and effectively with minimal or no impact on the environment”. The dissertation presents the green computing theme, which has not got much attention in Lithuania. Here parallel computing is presented, a combinatorial optimization branch and bound algorithm is proposed for the optimization of truss structures, and grid computing and clusters are overviewed. A method for parallel execution of tasks on grids, where there is no access to the MPI, is proposed in the dissertation. As well in addition, a way to reduce power consumption for parallel computers is proposed and experimentally investigated