Program Development Environment

GIDE - A Comprehensive Development Environment on GARUDA Grid

Grid computing has emerged as one of the key computing paradigms enabling large-scale and enhanced scientific endeavors to be carried out via collaborations on a global scale. GARUDA is a collaboration of Scientific and technological researchers on a nation wide grid comprising of computational nodes, mass storage and scientific instruments.

GIDE is an Integrated Development Environment across GARUDA Grid.

The Grid infrastructure includes various types of compute resources, storage resources, and special devices, interconnected through a high-speed communication channel that can be used to solve large scientific problems through collaborations. With a large number of scientists eager to use the Grid, the need of the hour is to provide a means through which the user can exploit the Grid without getting bogged by its complexities for developing, building, executing and debugging their Grid application.

GIDE helps an application developer with no prior knowledge on Grid Middleware (Globus), to carry out entire program development life cycle on Grid and maintain huge grid applications.

COMPONENTS OF GIDE

• Project Development & Management
• Resource Management
• Job Management
• Inbuilt source level Debugger
• Profiler
• Accounting
• Online context sensitive Help

SALIENT FEATURES OF GIDE
- User-friendly interface management components: Windows, Menus, Panels, Toolbars and Editor to navigate easily in the application.
- Platform/ Operating System independent: Supports Building, Execution and compilation options across different platforms (like Solaris, AIX, Linux) for serial and parallel applications.
- Enhances efficiency of the application and productivity of the application developers.

Project Development & Management

• Workspace concept allows navigation of files in huge applications efficiently
• Support multiple language (C, C++, Fortran languages) and multiple editor windows along with all standard editing features
• Compilation across different platforms of Grid

Resource Management

• Resource browser displays all the active resources on the Grid at any given point of time
• Dynamically updates the resource details
• Provides ‘Advanced Search’ option to resources based on different parameters

Job Management

• Job submission across GARUDA grid on selected resources just with mouse click
• Automation of ‘makefile’ and ‘rsl’ files generation before compiling and executing the grid application
• Redirection of input(s) / output(s) / error files as and when generated by the application

Source level Debugger
Comprehensive and flexible environment for debugging serial and parallel applications
Debugs step by step at source level and supports all standard debugging features

Profiler

• Graphically depicts the time consumed by all user defined functions or subroutines of the grid application
• Helps to identify the parts of the program to be modified to achieve better performance

Accounting
This helps to account the usage of various GARUDA resources like CPU, Memory, Virtual memory, wall clock time etc

Online context sensitive help at every stage of development

AVAILABILITY
GIDE is available on Submit node of the GARUDA Grid. GIDE can be invoked invoked either through VNC (Virtual Network Computing) Viewer or exporting the remote display.

GRIDHRA
Gridhra is a web based debugger and visualization tool for parallel applications on the Grid. Designed specifically to meet the scalability and heterogeneity concerns in a grid, this tool offers a unified interface for performing both correctness and performance debugging of grid-enabled parallel applications. Deployed as a web tool, Gridhra frees the user from the hassles of remote debugging and performance visualization of such applications. With a feature-rich graphical user interface, Gridhra masks all the complexities involved in debugging distributed applications.

Features

• Caters to the heterogeneity of the resources on the grid.
• Offers a rich GUI wizard for masking all the intricacies involved in generating the job description and submitting it.
• Supports Source Level Debugging of MPICH-G2 applications.
• Supports logical grouping of processes for issuing debugging commands collectively.
• Provides a visual depiction of the application flow, in coherence with the source code.
• Offers a comprehensive, visual depiction of the Call Graph sequence of the application, along with the time spent in each routine.
• Provides a Time Line representation of the communication events in the application.
• Provides communication -computation breakup, both process-wise and application-wise.
• Highlights the frequently used calls and time spent in each of them.

AVAILABILITY