From: Peter Freddolino (petefred_at_umich.edu)
Date: Sat Dec 05 2015 - 17:57:22 CST
Brian, many thanks for catching this, and Jason, thanks for the heads up on likely cause.
Brian, any luck in fixing this yet? I will take a look tonight if there is not a complete fix already. This needs to get patched and distributed as soon as possible.
> On Dec 5, 2015, at 8:54 AM, Jason Swails <jason.swails_at_gmail.com> wrote:
> On Thu, Dec 3, 2015 at 5:26 PM, Brian Radak <brian.radak.accts_at_gmail.com> wrote:
> Yes? But in a different way. Now LJAvgA and LJAvgB are extremely large and the correction becomes effectively infinite. I guess that means the overflow is in the double now?
> It is *very* hard to overflow a double. You need a number on the order of 10^300 to do that. Double overflow is unlikely.
> Jason M. Swails
> Rutgers University
> Postdoctoral Researcher
This archive was generated by hypermail 2.1.6 : Thu Dec 31 2015 - 23:22:17 CST