Floating point errors with Redhat linux kernel 2.2.14-5.0smp

["David F. Green" <dfgreen ^at^ lms.mit.edu>]

Dear CCLers,
 I'd just like to warn anyone using RedHat Linux with kernel version
 2.2.14-5.0smp
 (the version which ships with RedHat 6.2).  We have discovered that
 occasional
 errors in floating point calculations, resulting in a NaN, can occur in
 random way
 when using this kernel.  Upgrading to 2.2.14-12smp fixes the problem.
 The problem
 may be related to occasional system crashes on heavily loaded systems
 running
 this version of the kernel reported on the RedHat website.  I would
 strongly suggest
 that anyone using 2.2.14-5.0smp upgrades to 2.2.14-12smp.  If you are
 running
 RedHat 6.2, and have not made any updates, you will be using
 2.2.14-5.0.  If you
 have anymore questions about this, feel free to email me.
 --
 ************************************************
 David F. Green
 Department of Chemistry
 Massachusetts Institute of Technology
 77 Massachusetts Ave., Rm. 6-133
 Cambridge, MA 02139, USA
 E-mail: dfgreen ^at^ lms.mit.edu
 Phone:  (617) 258-6229
 Fax:    (617) 253-7030
 ************************************************