Award winning pilchard
29 April 2013
An academic paper by a University of Sydney information technologies engineer has been recognised by his international peers as one of the most influential contributions to reconfigurable computing in the past two decades.
Professor Leong says the concept of reconfigurable computing has existed since the 1960s, when it was proposed that a computer was made of a standard processor and an array of reconfigurable hardware.
Reconfigurable computing is the application of field programmable gate array (FPGA) devices to computational problems, he states.
Reconfigurable computing can be used to accelerate computationally intensive applications such as those found in signal processing, cryptography and computational finance. Although reconfigurable computing offered enormous computational resources, they caused a bottleneck between the FPGA and the host computer, says Professor Leong.
His landmark paper published in 2001 and titled Pilchard - a reconfigurable computing platform with memory slot interface overcame this problem.
"We utilised the memory interface rather than the standard PCI interface. This provided an easy to use, low cost and high performance platform for reconfigurable computing," says Professor Leong.
The keen fisherman named his device 'Pilchard' because it was designed to be a small, inexpensive and plentiful resource like the fish.
The 21st International Symposium on Field-Programmable Custom Computing Machines (FCCM) in California has identified the 25 most significant papers from its first 20 years.
The papers capture pioneering work that has clarified the landscape, provided key tools, opened areas of research, resolved serious problems, illuminated difficult issues, and illustrated innovative ways to use FPGAs and other reconfigurable computing devices.
Since the publication of his groundbreaking research, Professor Leong's work has focused on the development of novel hardware and software solutions to problems that range from derivative pricing and hedging of foreign exchange risk to the treatment of cardiac disorders.
He says much of his work involves developing solutions to problems constrained by processing speed or power consumption.
Nowadays he often uses field-programmable gate arrays (FPGAs) as his 'canvas', as they enable the development of problem-specific architectures that are orders of magnitude more efficient than microprocessor-based solutions.
"Among the applications I've pioneered have been FPGA-based financial derivative pricing systems. A number of banks have since developed their own implementations based on these systems, and have found that they offer much lower power consumption than competing technologies.
"The overall goal of my research is to benefit society through improved high-performance custom hardware and parallel software."
|Follow University of Sydney Media on Twitter|
Media enquiries: Victoria Hollick, 9351 2579, 0401 711 361, firstname.lastname@example.org