NVIDIA GPUs used to power H1N1 research
Researchers in the United Kingdom and Thailand are using NVIDIA Tesla GPUs to run simulations demonstrating how H1N1 mutations in a key influenza enzyme can lead to increased drug resistance.
The research is revealing a critical mechanism of resistance to existing anti-influenza medication. In the future, it may be possible to develop inhibitor drugs that can address these mutations and reduce the impact of epidemics.
"The four-node, eight-GPU cluster allowed us to quickly run and repeat a much larger number of complex simulations than otherwise would have been possible to perform," Dr. Christopher Woods of the University of Bristol said. "This enabled us to exhaustively explore all virus mutations of interest, building up a detailed picture that allowed us to quickly identify the key steps in the mechanism of resistance A comparable CPU-only system with 16 to 24 CPUs would have taken twice as long, at best. And, it would have been practically impossible for our team to monopolize all this time on the cluster, given the high demand for compute cycles by other researchers across the university."