Integrated grid environment for massive distributed computing in neuroscience
Андрій Олександрович Сальніков, Роман Левченко, Oleksandr O. Sudakov · 2011
Automation methods for huge amount of computations in grid environment and their applications in Ukrainian National Grid (UNG) are described. Proposed methods are based on asynchronous job submission and control. Methods has solved the problems of job data staging in, description preparing, job submission, job status monitoring and staging out results. Methods are implemented for building integrated environment for investigations in neuroscience. Running thousands non-interacting jobs with different input data sets primarily suitable for grid computing. Using portal web-interface, you can handle those thousands of jobs in “one-click”. The first application results in neuroscience are obtained. Simulations covers Kuramoto phenomeno-logical phase oscillators model and Hodgkin-Huxley-Katz realistic neurons model. Proposed integrated environment can be extended to solve other scientific problems.