Over at Popular Science, Kathryn Doyle writes that the Sequoia supercomputer is being tasked with a simulation of the human heat that looks down to the cellular level and predicts how a heart would respond to particular drugs.
Modeling programs like Cardioid approximate the heart by breaking the organ down into units: The smaller the units, the higher the resolution and the more accurate the approximation of a real heart. Before Cardioid, the best resolution anyone could get was 0.2 mm in each direction, but Cardioid can run at 0.1 mm. “People typically have run these simulations for tens of heartbeats, but we are able to run for thousands of heartbeats on the full Sequoia, and at higher resolution,” said computational scientist Art Mirin of LLNL. “The simulation runs up to 300 times faster than was possible before.”
Fred Streitz, Dave Richards, and Art Mirin will reveal their results publicly for the first time at the SC12 in November, where the team is one of five finalists for the Gordon Bell Prize, given annually for highest scientific achievement in supercomputing.
Read the Full Story.