Showing papers by "Azerbaijan State Oil Academy published in 2018"

Convergence of hpc and ai: two directions of connection

Abstract: *Correspondence: Nigar Ismayilova, Department of General and Applied Mathematics, HPC Advance Research Center, Azerbaijan State Oil and Industry University, nigar.ismailova@asoiu. edu.az Abstract This paper examines the role of HPC systems in the solution of the most AI problems, on the other hand, assesses the impact of the application of AI methods on the resolution of different tasks in distributed systems. The findings from the literature review illustrate how these two main fields of science and information technologies can be converged together for the achievement of the goals in designing and developing of the intelligent agents with the high level of intelligence, also for the solution of optimization problems in distributed systems with the increasing complexity.

ExaMig Matrix: Process Migration based on Matrix Definition of Selecting Destination in Distributed Exascale Environments

Abstract: In traditional computing system, load balancer, interim selecting the process, determine the destination computing node based on describing Indicators process status. In distributed Exascale computing system, due to the possibility of occurrence of a dynamic and interactive nature in execution time, it is possible. That the chosen destination computing node affected with dynamic and interactive nature so cannot be considered as a destination in process migration. This paper, by changing management approach in process migration. Consider process as an abstract element on the target computing node and calculates the impact of the factors the parameters affecting the process. Considering the above factors make process migration manager able to create sets of computational node that can be considered as destination computing node. In the event of a dynamic and interactive nature, in each element of the set, the process migration management, consider the effects of the factors affecting the activity of the process management and then re-weighs the computing element which make the above set. Using this mechanism allow the process migration management in case of dynamic and interactive nature occurrence in destination able to decide about changing on global activity execution so it is not necessary to recall load balancer manager in order to choose destination computing node. These subject louds to decrease execution time of process migration activity in distributed Exascale computing system.