Latest

Geek gods

COEP students, Shaizeen Aga, Swati Gour and Akshay Galande, win Microsoft’s Imagine Cup Egypt ‘09, the world’s premier student technology competition. Vishal Gangawane reports

Vishal Gangawane

Posted On Monday, June 15, 2009 at 03:27:03 AM

The world is getting faster and we want to write code which runs faster. Parallel computing is one way you can use to do so

— Shaizeen Aga

A team from Class of ’09 of College of Engineering Pune won the first prize of USD 5,000 (Rs 2,37,249) for their project on Parallel Computing at the Imagine Cup Egypt '09.

The Imagine Cup is billed as the ‘World’s Premier Student Technology Competition’ and is organised by Microsoft Corp.

The competition is a challenge to youngsters to yoke their technology skills to work at solving some of the world’s toughest problems. This is the seventh year of existence for the Imagine Cup.

The ‘imagine’ in the name comes from Microsoft’s challenge “Imagine a world where technology helps solve the toughest problems facing us today”.

This year, more than 3 lakh students from 100 countries registered to be part of the Imagine Cup 2009.

The Pune-based team named Biollel (the members coined it by mating Computational Biology and Parallel Computing) was placed first in the Parallel Computing section.

The three-member team includes Shaizeen Aga, Swati Gour and Akshay Galande, all of whom have just completed their engineering from College Of Engineering Pune, this May.

The three were in the same class at the Information Technology department and have worked together on many other projects before.

“We are very passionate about computer science, algorithms and data structures being our favourites. We read and discuss them quite often and we really enjoy doing that,” said Shaizeen.

All about drug design
Drug design is the process of finding new medications based on the knowledge of the infected individual, ie, biological target.

It is a complicated area of work that usually involves the construction of a phylogenetic tree.

A phylogenetic (phylum=class) tree is an evolutionary tree. It shows the evolutionary relationships among various biologicial species or other entities with a common ancestor.

Biollel used the Maximum Likelihood (MLE) Algorithm, which is most accurate for the task, but is rarely used because it is so time-consuming.

Instead, scientists use algorithms which work faster, but with less statistical accuracy.

So, to make the MLE algorithm work faster, Biollel used parallel implementation (a form of computing in which many calculations are carried out simultaneously).

They used multi-core platforms to make their code run faster, much like getting your colleagues to chip in when you want to to finish your work fast.

Biollel’s project dealt mainly with the parallel implementation of maximum likelihood method of phylogenetic tree construction algorithm for protein amino acid sequences using Microsoft's Task Parallel Library.

The project, according to the team, reduces the computational time for phylogenetic tree construction by efficiently implementing them in parallel so that we get performance benefits on multi-core platforms.

Phylogenetic Tree construction is crucial in drug design as well as protein structure prediction.

“In the final year of our engineering course, we are allowed to choose from a set of subjects to be our ‘elective’.

Computational Biology was one of them and it was going to be taught by faculty from Persistent Systems Ltd. The syllabus mentioned loads of algorithms and all three of us signed in for this course.

We had an amazing time interacting with the faculty and this led to our interests in this field.

Having studied various algorithms in the field of Computational Biology, we always marvelled at the vast amount of information this field incorporates in storing and analysing.

This holds the clue to solving many unsolved and tough problems related to life sciences,” said Aga.

The Imagine Cup has many other areas of competition like software design, embedded development, robotics algorithm among others.

So what made the team choose the parallel computing? “The world is getting faster. We want things to work fast and smart.

We as computer programmers want to write codes that run faster and are more efficient.

There are many ways we can achieve this, parallel computing being one of them. From single core now dual core machines are almost commonly manufactured.

We need to harness the power of multi-core platforms to make our code run faster,” said Shaizeen. She also pointed out the help she got from one particular quarter.

“We feel deep gratitude to the faculty from Persistent Systems Ltd for it was the interaction with them that led to our interests in this field. Also, without their guidance and support, this project would not have been possible.”

	Mail this article Print this article Translate this article	Rate me....
	Share Share Share Share

Contribute/ Share your Opinion

Latest

Geek gods

Contribute/ Share your Opinion

More