View source for Known issues caused by compiler bugs

'''''[[Intel Fortran Compiler|Previous]] | [[Fortran language resources|Next]] | [[Guide to compilers for GEOS-Chem]]'''''

#[[Supported compilers for GEOS-Chem]]
#[[GNU Fortran compiler|The GNU Fortran compiler (aka gfortran)]]
#[[Intel Fortran Compiler|The Intel Fortran compiler (aka ifort)]]
#<span style="color:blue">'''Known issues caused by compiler bugs'''</span>
#[[Fortran language resources]]


On this page we provide a list of known issues that are caused by bugs in the compiler used to build the GEOS-Chem executable file.  We have resolved most of these issues, or at least have workarounds for them.

Please also see [[Common_GEOS-Chem_error_messages|our ''Common GEOS-Chem error messages'' wiki page]] for a list of other commonly-encountered GEOS-Chem errors.

== GNU Fortran Compiler ==

The following GEOS-Chem issues have been traced to bugs (or configuration oddities) in the [[GNU Fortran compiler]]:

{| border=1 cellspacing=0 cellpadding=5 

|-bgcolor="#cccccc" align="center" valign="top"
!width="75px"|First reported
!width="75px"|<tt>gfortran</tt><br>version
!width="200px"|Module or routine
!width="475px"|Description
!width="275px"|Current Status

|-valign="top"
|[[GEOS-Chem v11-01|v11-01]]
|4.8.2
|<tt>ncdf_mod.F90</tt><br><tt>hco_diagn_mod.F90</tt>
|Not a compiler bug per se, the [[GNU Fortran compiler]] chokes when encountering subroutines (with <code>OPTIONAL</code> arguments) that are included in an <code>INTERFACE</code> block.  This seems to be caused by a strict adherence to the Fortran 2003 or Fortran 2008 standard. <span style="color:red">'''''NOTE: This same issue was also observed to occur when compiling GEOS-Chem with ifort 17.'''''</span>
*'''Reference:''' [[Intel_Fortran_Compiler#Cannot_compile_GEOS-Chem_v10-01_with_Intel_Fortran_Compiler_v17|Cannot compile GEOS-Chem v10-01 with Intel Fortran Compiler v17]]
|<span style="color:green">'''RESOLVED.'''</span>
*We have removed the offending code in [[GEOS-Chem v11-01]] and later versions

|-valign="top"
|[[GEOS-Chem v11-01|v11-01]]
|4.8.2
|<tt>hcox_paranox_mod.F90</tt>
|The <code>ASIND</code> function, which computes the inverse sine of an argument in degress, is not supported in Gfortran.
*'''Reference:''' [[GNU_Fortran_compiler#Now_us_ASIN_instead_of_obsolete_function_ASIND|Now use <code>ASIN</code> instead of obsolete function <code>ASIND</code>]]
|<span style="color:green">'''RESOLVED.'''</span>
*We convert the argument from degrees to radians, then use the standard <code>ASIN</code> function.

|-valign="top"
|[[GEOS-Chem v11-01|v11-01]]
|4.8.2
|<tt>ucx_mod.F</tt>
|The <code>SIND</code> function, which computes the sine of an argument in degress, is not supported in Gfortran.
*'''Reference:''' [[GNU_Fortran_compiler#Now_use_SIN_instead_of_obsolete_function_SIND|Now use <code>SIN</code> instead of obsolete function <code>SIND</code>]]
|<span style="color:green">'''RESOLVED.'''</span>
*We convert the argument from degrees to radians, then use the standard <code>SIN</code> function.

|-valign="top"
|[[GEOS-Chem v11-01|v11-01]]
|4.8.2
|<tt>isrpia.inc</tt>
|GNU Fortran v5 and higher does not allow you to place a <code>COMMON</code> block name in an <code>!$OMP THREADPRIVATE</code> statement.
*'''Reference:''' [[GNU_Fortran_compiler#THREADPRIVATE_error_in_GNU_Fortran_5_and_higher|<code>THREADPRIVATE</code> error in GNU Fortran 5 and higher]]
|<span style="color:green">'''RESOLVED.'''</span>
*This issue will be standardized in [[GEOS-Chem v11-02]].
*If you are using [[GEOS-Chem v11-01]], then you can [[ISORROPIA_II#Implement_ISORROPIA_v2.0_as_a_Fortran_module|apply the patch described here]].

|-valign="top"
|[[GEOS-Chem v11-01|v11-01]]
|4.8.2
| -
|GNU Fortran 4.8 by default uses the DWARF-4 library.  The Totalview debugger needs either DWARF-2 or DWARF-3.
*'''Reference:''' [[GNU_Fortran_compiler#GNU_Fortran_4.8_and_higher_is_not_compatible_with_Totalview|GNU Fortran 4.8 and higher is incompatible with totalview]]
||<span style="color:red">'''UNRESOLVED.'''</span>
*Use the <tt>gdb</tt> debugger, or
*Use <tt>gfortran 4.4.7</tt>

|}

--[[User:Bmy|Bob Yantosca]] ([[User talk:Bmy|talk]]) 21:52, 20 January 2017 (UTC)

== Intel Fortran Compiler ==

The following issues have been traced to bugs in the Intel Fortran Compiler (aka <tt>ifort</tt>).

{| border=1 cellspacing=0 cellpadding=5 
|- bgcolor="#cccccc" align="center" valign="top"
!width="75px"|First reported
!width="75px"|<tt>ifort</tt><br>version
!width="200px"|Module or routine
!width="475px"|Description
!width="275px"|Current Status

|-valign="top"
|[[GEOS-Chem 12#12.8.1|12.8.1]]
|2019.0.3.199
|Somewhere in KPP?
|Causes an apparent optimization error in tropchem simulations, where OH and O3 go to zero.
*'''Reference:''' [https://github.com/geoschem/geos-chem/issues/317 geoschem/geos-chem #317]
|<span style="color:red">'''UPDATE REQUIRED.'''</span>
*Avoid using ifort 2019.0.3.199.  

|-valign="top"
|[[GEOS-Chem v10-01|v10-01]]
|17 (maybe 16 too)
|<tt>ncdf_mod.F90</tt><br><tt>hco_diagn_mod.F90</tt>
|Not a compiler bug per se, but Intel Fortran v17 chokes when encountering subroutines (with <code>OPTIONAL</code> arguments) that are included in an <code>INTERFACE</code> block.  This seems to be caused by a strict adherence to the Fortran 2003 or Fortran 2008 standard, unlike prior Intel compiler versions.  <span style="color:red">'''''NOTE: This same issue was also observed to occur when compiling GEOS-Chem with GNU Fortran.'''''</span>
*'''Reference:''' [[Intel_Fortran_Compiler#Cannot_compile_GEOS-Chem_v10-01_with_Intel_Fortran_Compiler_v17|Cannot compile GEOS-Chem v10-01 with Intel Fortran Compiler v17]]
|<span style="color:red">'''UPDATE REQUIRED.'''</span>
*Update to [[GEOS-Chem v11-01]], in which we have removed the offending code.

|-valign="top"
|[[GEOS-Chem v10-01|v10-01]]
|15.0.0<br>and newer
|<tt>regrid_a2a_mod.F90</tt>
|A [[Common_GEOS-Chem_error_messages#Segmentation_faults|segmentation fault error]] occurs when compiling GEOS-Chem with <tt>BOUNDS=yes DEBUG=no</tt> (which sets <tt>-check bounds -O2</tt>).

[https://software.intel.com/en-us/forums/topic/585242 Bob Y. reported the issue to Intel Customer Service].  They acknowledge that the issue occurs but have not issued a fix in <tt>ifort 15</tt>.  
*'''Reference:''' [[HEMCO#IFORT_15_error_when_using_array-out-of-bounds_error_checking|IFORT 15 error when using array-out-of-bounds error checking]]
|<span style="color:darkorange">'''WORKAROUND.'''</span>
*[[GCST]] recommends always compiling with <tt>DEBUG=y BOUNDS=y</tt> if you want to check for array out-of-bounds errors.

|-valign="top"
|[[GEOS-Chem v10-01|v10-01]]
|12<br>13.0.079<br>and newer
|<tt>hcoio_dataread_mod.F90</tt>
|Apparently a HEMCO pointer variable became disassociated without any good reason.
*'''Reference:''' [[HEMCO#IFORT 13/IFORT 14 segmentation fault error|IFORT 13/IFORT 14 segmentation fault error]]
|<span style="color:darkorange">'''WORKAROUND.'''</span>
*Seb Eastham recommends putting in a print statement (as described in the reference), which seems to alleviate the error.
*This workaround will be added to [[GEOS-Chem v11-01#v11-01g|v11-01g]].

|-valign="top"
|[[GEOS-Chem v10-01#v10-01h|v10-01h]]
|11.1.069<br>and newer
|<tt>regrid_a2a_mod.F90</tt>
|An optimization error was setting an array (that held the sines of latitude at the N and S edges of each grid box) to zeroes.  This propagated to the <code>MAP_A2A</code> regridding routines, thus causing GEOS-Chem to exit abnormally.

It turned out that passing an array reference to the sine function<br>(e.g. <code>SIN( YDGR(I,J,L) )</code>) was the culprit.
*'''Reference:''' [[Regridding in GEOS-Chem#Fix for optimization error discovered in v10-01h|Fix for optimization error discovered in v10-01h]]
|<span style="color:green">'''RESOLVED.'''</span>
*Rewriting the call to the <code>SIN()</code> function (as described in the reference) fixed this error.

|-valign="top"
|[[GEOS-Chem v9-01-03|v9-01-03]]
|12.1.3.293
|<tt>strat_chem_mod.F90</tt>
|GEOS-Chem nested-grid simulations exited with this error, which occurred in the stratospheric chemistry module:
<tt>catastrophic error: **Internal compiler error: segmentation violation signal raised</tt>

[[GCST]] discovered that an undefined variable (i.e. only used for global simulations but not nested-grid simulations) was being accessed within a parallel loop  <tt>Ifort</tt> should have thrown a compile-time error, but didn't.
*'''Reference:''' [[Intel_Fortran_Compiler#Bug_fix_for_GEOS-Chem_compiled_with_Intel_Fortran_Compiler_12|Bug fix for GEOS-Chem compiled with Intel Fortran Compiler 12]]
|<span style="color:green">'''RESOLVED.'''</span>
*We rewrote an <code>#ifdef</code> block to prevent the undefined variable from appearing during nested-grid simulations.
*This was issued as a [[GEOS-Chem v9-01-03|v9-01-03]] post-release patch on 05 Oct 2012.

|-valign="top"
|[[GEOS-Chem v9-01-02|v9-01-02]]
|12.1.3.293<br>on Mac OSX
| -
|The default compilation resulted in this error:<br>
<tt>ld: library not found for -lcrt1.10.6.o</tt>
*'''Reference:''' [[Intel_Fortran_Compiler#Error_compiling_with_IFORT_12_and_Mac_OS_X|Error compiling with IFORT 12 and Mac OS X]]
|<span style="color:green">'''RESOLVED.'''</span>
*As described in the reference, a certain path has to be added to the <tt>LD_LIBRARY_PATH</tt> variable.

|-valign="top"
|[[GEOS-Chem v8-02-03|v8-02-03]]
|11.0.074
|<tt>partition.F</tt>
|GEOS-Chem would stop with an abnormal exit in <tt>partition.F</tt>.  Also, the spatial pattern of species as computed by SMVGEAR (especially NOx) was totally wrong.

It turned out that the earlier builds of <tt>ifort 11</tt> (i.e. prior to <tt>ifort 11.1.058</tt>) produced numerically-unstable code.  Other non-GEOS-Chem users had reported this as well.
*'''Reference:''' [[Intel_Fortran_Compiler#Problems_with_IFORT_11.0.xxx|Problems with IFORT_11.0.xxx]]
|<span style="color:green">'''RESOLVED.'''</span>
*Moving to <tt>ifort 11.1.058</tt> or higher solved the problem.

|-valign="top"
|[[GEOS-Chem Adjoint|v8-02-03 adjoint]]
|11.1.069
| -
|An apparent linker error produced this error message:<br><tt>OMP abort: Initializing libguide.so, but we but found libguide.so already initialized</tt>.
*'''Reference:''' [[Intel Fortran Compiler#Incompatibility between IFORT 11 and OS version|Incompatibility between IFORT 11 and OS version]]
|<span style="color:green">'''RESOLVED.'''</span>
*Updating the OS version seems to have fixed this error.

|}

--[[User:Bmy|Bob Yantosca]] ([[User talk:Bmy|talk]]) 21:29, 20 January 2017 (UTC)