Method for the pairwise processing of elements in data arrays for many-task computing in a hybrid cloud
Seminars
Laboratory of Information Technologies
Date and Time: Wednesday, 20 April 2022, at 10:00 AM
Venue: Conference Hall, Laboratory of Information Technologies; online on Webex,
Seminar topic: “Method for the pairwise processing of elements in data arrays for many-task computing in a hybrid cloud”
Speaker: I. V. Bobyleva (Samara National Research University)
Abstract:
The dissertation is devoted to the development of a method for organizing efficient computing using hybrid cloud systems in data processing problems. The method is based on the joint application of Cole’s algorithmic skeletons and a variant of Hewitt’s actor model to describe many-task computations. The proposed method is investigated by the example of algorithms for the pairwise processing of data arrays according to the principle of a round-robin tournament. Computational experiments demonstrate the effectiveness of software implementations of these algorithms on various types of modern computer architectures that represent hybrid cloud resources.
The dissertation is devoted to the development of a method for organizing efficient computing using hybrid cloud systems in data processing problems. The method is based on the joint application of Cole’s algorithmic skeletons and a variant of Hewitt’s actor model to describe many-task computations. The proposed method is investigated by the example of algorithms for the pairwise processing of data arrays according to the principle of a round-robin tournament. Computational experiments demonstrate the effectiveness of software implementations of these algorithms on various types of modern computer architectures that represent hybrid cloud resources.
(based on the PhD thesis)