[communities] GGF Proposal Submission

[spamidig at ncsa.uiuc.edu]

proposers_name: Sudhakar Pamidighantam 
 
affiliation: University of Illinois at Urbana Champaign 

email: spamidig at ncsa.uiuc.edu 

proposed_title: GridChem a Computational Chemistry Grid  

session_type: Tutorial 

proposed_duration: Half a day 

target_audience: Managers, Users( Computational Chemists), Technical Experts 

num_attendees: 2 to 3 Presenters 

abstract: This tutorial will provide an overview of the organization of GridChem, a The “Computational Chemistry Grid” (CCG) is a virtual organization that provides access to high performance computing resources for computational chemistry through a desktop client. The CCG environment is based on a three-tiered architecture that includes the client, a grid middleware server and a set of geographically distributed, high-end computational resources. 

This tutorial will focus on two aspects of the computational chemistry grid. 1. how to interface to the CCG using the desktop client, the user computational chemists perspective. 2. How to integrate new resources, both hardware, scientific computational applications and related services into the computational chemistry grid, the resource providers’ perspective.  The desktop client, called GridChem, provides a simple and intuitive interface incorporating chemistry functionality that computational chemists need to conduct their research, including pre- and post-processing tools.   With its extensible interface, other popular quantum chemistry packages can be integrated.  This comprehensive tutorial will cover tasks from login/authentication to post-processing and visualization. The resource management and integration involves defining the resources, hardware, applications and services and interfaces between these and implementing the integration. The resource definition, disco!
very and utilization by the client will also be described. The grid services require a robust failsafe framework for production quality usability  and a redundant middleware server implementation will be described.  

 

synopsis:  Tutorial Goals
The “Computational Chemistry Grid” (CCG) is a virtual organization that provides access to high performance computing resources for computational chemistry through a desktop client.  This organization provides community allocations of CPU time based on the available resources as well as serves existing user allocations at distributed HPC resources that are integrated into the organization using a simple web browser based interface.  

The CCG environment is based on a three-tiered architecture that includes a client, a grid middleware server and a set of distributed, high-end computational resources. The desktop client provides a simple and intuitive interface incorporating chemistry functionality that computational chemists need to conduct their research, including pre- and post-processing tools. Contained within the desktop client are modules for authentication via SSH, Kerberos and Globus certifi-cates, input preparation, resource discovery and  selection, job submission and monitoring.  Cur-rent chemistry software supported by GridChem includes Gaussian03, GAMESS, Molpro and NWChem.  With its extensible interface, other popular quantum chemistry packages can be inte-grated.

The goals of this tutorial are two fold  1. To engage computational chemists ( traditional/novices) to use grid resources at their fingertips 2. To engage production grid providers to collaborate in establishing application specific cyber environments for specific communities. More specifically, this tutorial would:

•	Describe the Computational Chemistry Grid project and what software, hardware and personnel have been committed
•	Cover, in detail, the installation, capabilities and functions of the desktop interface
•	Show through use of examples the ability to submit multiple quantum chemistry applications to the computational chemistry “Grid”. Monitor and analyze results from previously submitted jobs using job management tools in the client.
•	Describe the grid services layer and middleware utilization
•	Describe the Resource database and discovery technologies and new resource in-tegration requirements
•	Provide user management tools description including, allocation, consulting and usage reporting capabilities


 
 Relevency and currency of computational chemistry grids
The field of computational chemistry has had an extensive impact in the areas of drug design, materials science, environment chemistry and chemical engineering.  Currently, due to the fun-damental nature of matter several other faculties  such as mechanical engineers and geologists and atmospheric scientists are resorting to computing properties of matter of interest at a molecu-lar level and the need for resources and for cyber environments has increased tremendously. It is also one of the largest consumers of CPU cycles among local, regional state and national supercomputer centers.  The Computational Chemistry Grid offers this community a desktop interface to diverse computational resources, relieving the scientists of much of the burden of interfacing to these complex systems and managing job submission and data management.  In addition, the resources available through the Computational Chemistry Grid combined together can provide computational power far beyond what rese!
arch chemists have at their own individual organizations.

General Description Of Tutorial Content
The tutorial is comprised of material organized into sections, based on the first level of the out-line given later in this document.  Printable versions of the material will be available, as well as electronic versions of all sample input/output files.  The material will cover both organizational overview topics as well as step-by-step description of using the desktop client.  

Fully explained exercises and all supporting files will be provided to highlight key concepts or procedures with GridChem.


 Part1. Description of Client Demonstration or Exercises User Computational Chemist’s Perspective

We propose to offer hands on exercises with tutorial attendants to reinforce concepts and capa-bilities covered in this tutorial.  The tutorial presenters will work closely with the attendees to cover the following exercise topics:

1.	Client Software Installation
2.	Authenticating
3.	Building Jobs
a.	General Visualization Interface
b.	Gaussian 03 Graphical Interface
4.	Submitting jobs
a.	To Specific Compute Resources
b.	To General Computational Chemistry Grid Through The Condor Interface
5.	Post-Processing Results

Since a key component of GridChem is the desktop client application that is the interface to the Computational Chemistry Grid, people who sign up for the workshop will be encouraged to bring their personal laptops to the tutorial so that they will leave the tutorial with the software in-stalled on their laptop.   Some demo allocations will be available for the audience to try some of the functionality if network connectivity is available.

 Detailed Outline
I.	Computational Chemistry Grid Overview
a.	Partner Institutions
b.	Desktop Client
c.	Middleware Services ( See Part 2 also)
d.	Computational Resources
e.	Data Storage Resources
II.	Software Installation 
a.	System Requirements
b.	Client Installation And Execution
III.	Authentication/Authorization And Allocations
a.	Policies
i.	Access
ii.	Authentication
b.	Allocation Process
IV.	Interface Fundamentals 
a.	Authentication
b.	Creating And Submitting Jobs
c.	Managing Jobs
d.	Post-Processing Results
V.	Chemistry Applications 
a.	Gaussian 03, Gamess, Molpro, Nwchem
b.	Aces Ii, Q-Chem
VI.	Pre-Processor/Molecule Builder 
a.	Graphical Interface For Gaussian 03
b.	Nanocad
VII.	Job Manager 
a.	Submitting Jobs
b.	Monitoring Jobs
c.	Retrieving Job Logs And Output
VIII.	Post Processing/Visualization 
a.	Viewing Application Output Files
b.	Nanocad Molecular Editor
IX.	Technical Support 
a.	On-Line Training Material
i.	User Syntax Checking
ii.	Explanation Of Package Features
b.	Technical Support System
i.	Scope Of Technical Support
ii.	PCS Trouble Ticket System

This component may take 2/3 of the time and the rest of the 1/3 will be used for part 2 detailed below.

Part2. Description of Resource and Grid Management. Resource providers Perspective  


X.	Resource Definitions
XI.	Resource discovery and presentation
XII.	Scientific Application Grid Resource Interfaces
XIII.	Resource Integration
XIV.	Automated job submissions
XV.	Middleware server organization, redundancy failover mechanisms
XVI.	Application deployment and maintenance
XVII.	Storage requirements
XVIII.	Community allocations and security
XIX.	Current Status And Future Directions
A. Current Status and utilization
     a. Users and their experience
     b. Initial lessons in collaboration and distributed/Grid production
     c. Application and System issues
     d. Bugs and Error notifications

 B. Planned Activities 
     a.  Middleware upgrades and Web services 
     b.  Metascheduling and optimization
     c.  Benchmarking of scientific applications
     d.  Reporting to Funding Agencies
     e.  Interactions with other virtual organizations

Outreach Activities:
Computational Chemistry Grid started with a team (EOTS) dedicated to providing education, outreach, training and services. This team is engaged in providing documentation, tutorials and posters, presentations and workshops. We have already delivered a tutorial,  an Access Grid demonstration and workshop at Supercomputing 2005 conference. We have presented our ex-periences in previous GGF conference  and disseminated at American Chemical Society and CCGrid conferences. Access Grid events were advertised locally to computational chemistry community at all participating sites. Several national and international resource provides have noticed and shown interest in collaborations at various levels. We are currently working on gen-erating a draft memorandum of  understanding to use in such collaborations.

Schedule :
CCG has just entered a friendly user period and slowly we are encouraging some early adapters to use the resources. Several traditional users have been granted allocations. We will plan to go to full production early in 2006 and gradually improve the seamlessness of Grid Usability and expand the resources and  application pool including post processing tools and provide reusable data from existing applications in others. 
 

tech_requirements: Access Grid will be useful to widen the audience, Good network connectivity is required live demonstration. 

prereq_participants: The first part of the tutorial will require familiarity with computational chemistry. Particularly the G03 GUI part will require some familiarity with the G03 application. 

advertise_suggestion: AccessGrid Scheduling/Announcements ( in case AG session is available). Communicating with local computational chemists and those among the attendees.
Production application Grid providers and Science gateways provides need to be specifically contacted. Supercomputer centers which are potential resource providers will have to be intimated of the event.