Skip to main content

High Performance & Distributed Computing

Introduction

 

Bernard Crossland Building, 18 Malone Road

 

The discipline of High Performance and Distributed Computing (HPDC) historically attempts to solve

computational problems that cannot be solved with the processing power and memory of a single computing

system. The parallel and distributed processing capacity offered by multiprocessors, clusters and

grids over the past decades has been instrumental in tackling computational challenges in science and

engineering. Since the first supercomputer developed in the 1960s, the HPDC discipline has laid the

foundations for making computing systems faster and a catalyst for accelerating scientific discovery.

 

In the 21st century, the foundations and methods of HPDC are more timely and relevant than ever.

Most computers in our homes and our planet are now sophisticated parallel processing engines. Our

desktops, our laptops, the embedded computers that power mobile phones, tablets, cars, airplanes and

many other artifacts that we use in our everyday lives, as well the servers that process the Exabytes

of data that humanity generates in datacentres, have many processor cores, specialised computational

accelerators and deep memory hierarchies. While parallel computing penetrates each and every computing

device, distributed computing interconnects billions of computing devices into a global computing

ecosystem. Emerging applications and services split their computation and data processing tasks between

mobile devices and datacentres that form computational clouds. Networks of distributed sensors

assume critical missions such as securing our power grids or monitoring physiological indicators in our

bodies. These sensors collect massive amounts of data which are processed in servers, to give meaningful

answers back to millions of users. Once a niche area, HPDC is now ubiquitous and spans the

entire computing ecosystem.

 

The mission of the HPDC Research Cluster, part of the School of EEECS at Queen's University of

Belfast, is to develop the theoretical foundations, algorithms, hardware and software that will make future

computing systems faster, more resilient, energy efficient and thus friendlier to the environment, sustainable,

and less expensive to operate. The Cluster is committed to making contributions of a foundational

nature, with impact that extends far and beyond the HPDC discipline and Computing Sciences.

The research practice in the cluster balances harmoniously theory, experimentation and deployment:

we develop solutions that we test in the field, on real parallel and distributed systems, some of them with

hardware components built in the lab by ourselves. We develop algorithms, programming languages and

concurrent execution environments that scale from a few up to millions of processing elements. We build

methods to improve the reliability and energy-efficiency of small- and large-scale distributed systems.

A key advantage of our methods is that they jointly consider application for the co-design of hardware

and system software. Our methods and systems are thus applicable to a variety of domains. We support

commercially relevant applications of high societal and economic value. In par with groundbreaking

fundamental and applied research, we are committed to training and mentoring students, early-career

and advanced researchers in the principles and practice of parallel and distributed computing. We provide

unique skill sets that are highly marketable in both industry and academia.

 

The Cluster has a remarkable and diverse research funding portfolio with a total value that exceeds

£14 million over the last three years. The funded research activities span the areas of many-core programming

languages and systems, programming models for Exascsale systems, energy accounting and

optimisation, micro-servers, datacentre monitoring and optimisation, middleware for large-scale data processing

and stream data processing. We further explore foundational aspects of networked distributed

systems. Our projects involve 26 academic partners, 16 industrial partners, and 4 supercomputing centres,

from 14 countries. Our team includes members with accolades such as Marie Curie Fellowship,

Newton Fellowship, NSF & DOE Career Award, EPSRC First Grant, IBM Faculty Award, and numerous

Best Paper Awards.

 

We invite you to explore our web pages and welcome contacts from academic or industrial partners

for exploring collaborations.

HPDC labs HPDC labs
HPDC labs HPDC labs
HPDC labs HPDC labs
HPDC labs HPDC labs

Home

News

Aug 29
CCGrid'14