Co-author(s) :

Naghashyan Zaven

Abstract:

Cluster job management system was designed for organizing parallel access to the cluster by different users. The management and efficient exploitation of clusters among users, applications and data continue to be a timeconsuming and challenging task. In many Linux-based clusters the jobs should be designed as programs that utilize MPI (message passing interface). MPI became de facto industry standard for the parallel computing environments. The most common programming languages that use MPI are C and C++. One of the main limitations of the system is the fact, that programs should be implemented using C++ programming language. The number of engineers with deep knowledge of C++ during the last decade stopped to grow in a way it was before. Modern programming trends made JavaScript as one of the most popular languages for majority of programmers. Initially JavaScript was designed to be used within web browsers. Node.Js is an environment that allows the programs written in JavaScript to run in the server environment. It supports special APIs that allow JavaScript programs to communicate with the operating system and external resources. Those APIs are mostly designed as asynchronous functions. MPI-Node.js Bridge is designed and implemented to allow Node.Js programs to be executed within Arm-cluster computational environment and utilize MPI functionality that allows the simplicity and efficient use of the available computing resources, as well as, making the system usable for vast majority of new users that have JavaScript programming skills.

Identifier:

oai:noad.sci.am:135810

Language:

English

URL:

Additional Information:

mikayelg@gmail.com ; znaghash@gmail.com

Affiliation:

Institute for Informatics and Automation Problems

Country:

Armenia

Year:

2017

Time period:

September25-29

Conference title:

11th International Conference on Computer Science and Information Technologies CSIT 2017

Place:

Yerevan

Participation type:

oral