https://community.intel.com/t5/Intel-Fortran-Compiler/Is-this-a-volatile-variable-problem/m-p/843399#M61368 <DIV>Hi all,</DIV> <DIV></DIV> <DIV>I have a large program that works perfectly well in debug mode but fails in release mode. The problem appears to be the closing of a file (on unit 1) but I'm unable to find out where this happens because it doesn't happen in debug mode.Could this be the result ofa volatile variable? I've heard about them but never met one before - how do I go about tracking it down?</DIV> <DIV></DIV> <DIV>Any suggestions?</DIV> <DIV></DIV> <DIV>With many thanks in advance,</DIV> <DIV></DIV> <DIV>Mike</DIV> Tue, 15 Nov 2005 15:58:28 GMT michael_green 2005-11-15T15:58:28Z https://community.intel.com/t5/Intel-Fortran-Compiler/Is-this-a-volatile-variable-problem/m-p/843399#M61368 <DIV>Hi all,</DIV> <DIV></DIV> <DIV>I have a large program that works perfectly well in debug mode but fails in release mode. The problem appears to be the closing of a file (on unit 1) but I'm unable to find out where this happens because it doesn't happen in debug mode.Could this be the result ofa volatile variable? I've heard about them but never met one before - how do I go about tracking it down?</DIV> <DIV></DIV> <DIV>Any suggestions?</DIV> <DIV></DIV> <DIV>With many thanks in advance,</DIV> <DIV></DIV> <DIV>Mike</DIV> Tue, 15 Nov 2005 15:58:28 GMT https://community.intel.com/t5/Intel-Fortran-Compiler/Is-this-a-volatile-variable-problem/m-p/843399#M61368 michael_green 2005-11-15T15:58:28Z https://community.intel.com/t5/Intel-Fortran-Compiler/Is-this-a-volatile-variable-problem/m-p/843400#M61369 Actually, this could be a problem of <B>uninitialized</B> variable -- "volatile" is something entirely different (VOLATILE keyword declares that the variable is changeable outside of compiler's view, like a shared-memory mapped variable or a port value, and prevents certain optimization).<BR /><BR />You <B>can</B> debug the release configuration, if you set "debug info" to "full" and rebuild. The stepping through the code can be kind of funny, and many variables are unreadable as they're stored in registers, but you still can have basic control what happens. <BR /><BR />It can also happen, though, that it's not reproducible even in this case. It could help, though, to try to reproduce the exact conditions -- insert a call to DebugBreak() (use DFWIN) somewhere in your startup code, then <B>run</B> the program without debugger (Ctrl+F5). DebugBreak statement will cause the just-in-time debugger to attach itself to the process.<BR /><BR />If that still doesn't help reproduce the problem, well...<BR /><BR /><A href="http://www.xeffor.com" target="_blank">Jugoslav</A> Tue, 15 Nov 2005 16:12:31 GMT https://community.intel.com/t5/Intel-Fortran-Compiler/Is-this-a-volatile-variable-problem/m-p/843400#M61369 Jugoslav_Dujic 2005-11-15T16:12:31Z