Monday, July 6, 2009

Solar Simulations Rock New York

For the past year our research group has been working with the folks at the American Museum of Natural History in New York on their latest show Journey to the Stars which looks at stars from the Big Bang to the future end of our solar own sun. Ben Brown, my office-mate, did most of the work on our end and our simulations ended up producing about 3 minutes of the 25 minute show.

Journey to the Stars premiered last week in New York and was reviewed by both Scientific American and the New York Times. Scientific American's review has us particularly excited. They said:
"One highlight is a simulation of the interior of the sun, showing its convection and churning magnetic field. The demo came courtesy of Juri Toomre's group at the University of Colorado at Boulder, and required about 14 million hours of supercomputer time spread across four major U.S. supercomputing centers. Hundreds of billions of bytes of data were processed, all of which went into the visualization of the solar interior."

We'd love to make all of our science this accessible to the public, but even the 25 minutes of this show cost millions of dollars to produce. Hopefully, though, it will make its way to many other planetariums and help the public understand a little more of the wonder of astrophysics and the practical reasons to spend billions of dollars studying the sun.

And if your in New York, stop by and take a look. I hear it is a great show and they go to great pains to use real data/simulations and top-notch science as much as possible. You can watch the trailer or buy tickets here.


  1. That's really amazing Nick how many computational hours are needed for a short video. I know Dan Whalen is trying to model the life cycles of Pop III stars in 3d. It will take several years to get right and will similarly produce movies. But when the movies are done, you get some cool looking results. :)

  2. The stuff in the video took about 500,000 processor-hours because it was run at very high resolution but for a relatively short period of simulated time. Our normal simulations take anywhere from 100,000 to 5,000,000 processor-hours.

    I know of several research groups working on "star in a box" calculations. Everything that I've seen seems to indicate that they are still in a fairly primitive state, but they can do a much better job with realistic boundaries (like a vacuum at the outer edge of the star) than we can.


To add a link to text:
<a href="URL">Text</a>