@home with Windows AzureAs you may know, my colleagues Brian Hitney and Peter Laudati and I have put together @Home with Windows Azure, a great way to get your feet wet with Windows Azure and help medical science by deploying Stanford’s Folding@home client to the cloud using your free, 90-day Windows Azure trial account.

Some of the participants were drawn to the effort having participated in many of these distributed computing projects in the past and were naturally intrigued by the ability to get some free compute cycles for their favorite projects.  The nature of our original distribution of the code wasn’t really conducive to these “professional folders,” so we’ve made a few updates to the code, as Brian describes in his blog post.

The changes we’ve made are summarized below and will appeal primarily to those that are already familiar with the Folding@home project and have participated in it as part of other teams.

  • You can now specify a different team number (Windows Azure is team 184157) so your contribution is attributed to your current Folding@home team,
  • You can specify a passkey,
  • The default Folding@home client application (which we distribute) is now the SMP version, which generally yields more points even when run in a single core VM,
  • We’ve uploaded pre-compiled package files for both a small VM instance (single core) and a extra-large VM instance (8 core) to utilize the SMP client.  Keep in mind that with the trial account, you can a single core 7x24 for each month of your 3-month trial (roughly 750 hours of compute time), but running an 8-core instance will exhaust that 750 hours in just under four days.  During that four days, you’ll likely accrue more points than a month worth of folding with one CPU, but your account will be suspended when the 750 hours is exceeded, and you’ll have to redeploy the application when the meter resets for the subsequent months of your trial period.  See Brian’s post for a bit more insight into all of this and why it’s worth the effort!

By the way, the last of our webcast series related to @Home with Windows Azure is this coming Thursday.  We’ll be talking about scaling patterns for the cloud and hope you’ll be able to join us.