segmentation fault in OutputProcessor.f90

[jeff....@pnnl.gov]

I'm using gfortran 4.4.7 on RHEL5 with the energyplus-8.1.0.008-source distribution.  Using the attached example file 1ZoneUncontrolled.idf, I get a segmentation fault after seeing the console output "Reporting Surfaces".  Using valgrind, I can see the following problem:

==24476== Memcheck, a memory error detector
==24476== Copyright (C) 2002-2009, and GNU GPL'd, by Julian Seward et al.
==24476== Using Valgrind-3.5.0 and LibVEX; rerun with -h for copyright info
==24476== Command: ../EnergyPlus
==24476==
 EnergyPlus Starting
 EnergyPlus, Portable Version V8.1.0.008, YMD=2014.01.10 10:50
 Processing Data Dictionary
 Processing Input File
 Initializing Simulation
 Reporting Surfaces
==24476== Conditional jump or move depends on uninitialised value(s)
==24476==    at 0x12FAF24: producerddmdd_ (OutputProcessor.f90:8809)
==24476==    by 0x1D9046C: __simulationmanager_MOD_

managesimulation (SimulationManager.f90:264)
==24476==    by 0x94A389: MAIN__ (EnergyPlus.f90:421)
==24476==    by 0x2343849: main (in /home/d3n000/fpgi/energyplus-8.1.0.008-source/EnergyPlus)
==24476==

Unfortunately, I lack the skill to try and fix this and it's currently keeping me from using EnergyPlus.  Please let me know how to fix this. Thanks.

[Neal Kruis]

Jeff,

I can confirm this problem on my Fedora machine. I don't know what exactly is causing this problem, but I have found that if I change the optimization flag "-O" to "-O2" from the "CMAKE_Fortran_FLAGS_RELEASE" variable in the CMakeLists.txt file, this problem does not occur. This is interesting because this is not the first time we've had a problem with the optimization flags, and from what I can tell, what works might be dependent on the operating system.

Please give this a try and let me know if you are having any trouble with it. I will continue looking into this issue to see if I can't come up with a more general fix across architectures.

Thanks,

Neal