Problems with v10.1

[abenson]

Anyone else having problems with v10.1? A code that was working perfectly with v10.0 now produces floating exceptions. I haven't been able to pin these down yet because adding in additional output to attempt to figure out what's happening changes the error (which now occurs at another point in my code!). If I can make this reproducable I'll post an example.

$ ifort -V
Intel Fortran Compiler for applications running on Intel 64, Version 10.1 Build 20070913 Package ID: l_fc_p_10.1.008

Running on RHEL4, kernel 2.6.9-42.0.8.ELsmp, glibc-2.3.4

(I've seen the same error under other configurations also.)

[jimdempseyatthecove]

If the location of error moves within your code by inserting diognostic prints into the code then the error is likely due to placement of the data or code. Placement sensitivity often have as a root cause the use ofuninitialized variables. The unintended re-use of a global variable (e.g. TEMP in common) could also cause problems but is often not sensitive to placement. A seperate placement sensitivity problem has to do with templating a named common differently in different subroutine/funcitons. Values being stored at different addresses by one subroutine as observed by another subroutine (manifested in a NaN or DivBy0).

Are you by chance assuming variables are automatically initialized to 0?

Jim Dempsey