Built for Science – The TITAN Supercomputer


This computing beast is all set to redefine the limits of scientific calculations, creating a whole new era of “Accelerated Computing”. 

It was way back in 2009 that the Department of Energy’s (DOE) Oak Ridge National Laboratory was gifted with a supercomputer-the ‘Jaguar’. This supercomputer proved to be the then most powerful supercomputer in the world. Jaguar couldn’t hold that title for long for a Chinese system snatched it only a year after. However the U.S got the title back earlier this year, credit to the DOE’s Sequoia system. The history repeating itself, the title finally falls in the name ORNL only, this time the credit goes to the revamped version of the predecessor Jaguar. The transformed system was renamed to be the Titan, a name worthy of this miraculous machine.

This computing beast is all set to redefine the limits of scientific calculations, creating a whole new era of “Accelerated Computing”. With a mind-boggling peak performance of 20 petaflops (that is equal to 2000 trillion calculations per second).The update to the Jaguar made the machine (Titan) as much as 10 times more powerful and also enabled it to overcome space and power constraints. It’s truly remarkable to observe that the concept behind such efficient upgradation is very basic in nature. The processing unit is the key transformation that the machine has gone through. Titan uses a combination of CPUs and GPU (Graphics Processing Unit – a family of processors meant for gaming operations). It is this synchronized combination of the CPUs and GPUs that provides Titan a cutting edge over the others in existence. Such a combination requires less power than the native CPU-only systems. With an additive processing unit at service, Titan can now deliver numerous scientific calculations with greater speed and accuracy.

This 97 million dollar beast is developed by Cray Incorporation. The mega computer uses 200 cabinets covering around 404 sqm(4352 sq ft) with Cray XK7 blades which use AMD Opteron 6274 CPUs and NvidiaTesla K20 GPUs. There are 18,688 nodes (2 per blade), each containing a 16 core CPU with 32GB of RAM and a GPU with 6GB. The total number of processor cores is 299,008 and total amount of RAM is over 710 Terabytes. Titan draws 12.7 MW, 2 MW more than Jaguar did, but it is almost ten times as fast in terms of floating point calculations.

When technology underwent a transition from the historic single core architecture to the multi-core systems, the researchers around the world also had to alter their problem solving approach. They had to break their calculations into smaller problems that could be parceled out separately to different processing cores, an approach referred to as parallel computing. Titan has been developed with the insights that are pushing parallelism even farther by allowing researchers to continue dividing their larger problems into even smaller parcels. Because of this unparalleled synchronized combination of the CPUs and GPUs Titan is well able to handle hundreds of calculations simultaneously and enables the researchers to carry out scientific researches with greater speed and accuracy.

When you got a biggy like the Titan in hand, preparing the user base for it is as complicated a task as to design it. Titan is open to select projects while ORNL and Cray work through the process of finalizing the projects to be undertaken. However the major access to Titan in the coming year will be in the hands of the Department of Energy’s Innovative and Novel Computational Impact on Theory and Experiment program, also known as INCITE. Moreover the supercomputer will be employed in the following research studies.

WL-LSMS

Illuminating the role of material disorder, statistics, and fluctuations in nanoscale materials and systems.

LAMMPS

A molecular description of membrane fusion, one of the most common ways for molecules to enter or exit living cells.

S3D

Understanding turbulent combustion through direct numerical simulation with complex chemistry.

CAM-SE

Answering questions about specific climate change adaptation and mitigation scenarios; realistically represent features like precipitation patterns / statistics and tropical storms.

NRDF

Radiation transport – important in astrophysics, laser fusion, combustion, atmospheric dynamics, and medical imaging – computed on AMR grids.

Denovo

Discrete ordinates radiation transport calculations that can be used in a variety of nuclear energy and technology applications.

The verdict:

The final word on this Debut is that Titan has opened up a brand new route for the upcoming supercomputing developments. The Titan is the biggest hope for scientists to achieve the exascale of computing speed. With the official announcement due next month, the Titan will be crowned the “Fastest Supercomputer in Existance”.

Infographics credit:    www.olcf.ornl.gov

Sagar

google plus

Born and brought up in Gorakhpur, U.P , presently persuing my graduation at Assam University, silchar , Asam. My strength is turning technical mumbo-jumbo into easy to understand language. This is also my interest and passion and sums up what is at the heart of my ability to be the kind of technical writer you need.


Posted By
Sagar

2 November, 2012

7

0

SHARES