Researchers led by Peter Vincent from Imperial College London have made this year’s list of finalists for the Gordon Bell prize, with the backing of Piz Daint at the Swiss National Supercomputing Centre. The prize is awarded annually in November at SC, the world’s largest conference on supercomputing. It honors the success of scientists who are able to achieve very high efficiencies for their research codes running on the fastest supercomputer architectures currently available.
“Simulating fluid dynamic processes is still a challenge in engineering science. The research team from Imperial College developed new software that uses graphics processing units (GPUs) to better handle complex simulations of turbulence that are performed on large computer systems. Fluid dynamic simulations are essential, for example in aircraft construction, as the researchers emphasize in the abstract for the list of Gordon Bell finalists: “Accurate simulation of unsteady turbulent flow is critical for improved design of ‘greener’ aircraft that are quieter and more fuel-efficient.”
Piz Daint is a Cray XC30 capable of 7.8 petaflops. The hybrid system’s GPUs lend themselves well to code optimization. Vincent and his colleagues made use of this prior to testing their efficiency boost on the mighty 27 petaflop supercomputer Titan (a Cray XK7) at ORNL. With the PyFR software for simulating fluid dynamic processes, which is based on the Python programming language, the researchers attained 13.7 quadrillion compute operations per second – and with that, simulations more precise than ever before.