The -fpen (Linux* and Mac OS*) or /fpe:n (Windows*) option allows some control over the results of floating-point exceptions.
The Intel Fortran Compiler or simply IFORT is a family of extremely powerful Fortran compiler for Intel hardware. It comes with many lucrative feature sets that make program optimization very easy. Most of these Fortran compilers can pre-process, compile, assemble, and link Fortran applications very effectively on both IA-32 and Intel 64 systems. Written by Aron 3 Comments Posted in My work, Scientific computing Tagged with El Capitan, Fortran, Ifort, Mac, OpenMPI, Phonopy, VASP July 28, 2012 Mountain Lions like Fortran A slow running laptop was a good excuse to play around with a fresh install of OSX 10.8 (Mountain Lion). I have a program that need to be installed with the 'gfortran/ifort' compiler. I have a Mac OSX 10.6.8 (Snow Leopard). I know I have gcc installed, because I installed my self (from the gcc package inside the Xcode installer, so I did not have to install the whole xcode), also after typing in the terminal 'which gcc' I get the directory. I have a program that need to be installed with the 'gfortran/ifort' compiler. I have a Mac OSX 10.6.8 (Snow Leopard). I know I have gcc installed, because I installed my self (from the gcc package inside the Xcode installer, so I did not have to install the whole xcode), also after typing in the terminal 'which gcc' I get the directory. Ifort x.o (Linux OS and Mac OS X) ifort x.obj (Windows OS) The following command compiles a.for, b.for, and c.for. It creates three temporary object files, then links the object files into an executable file named a.out (on Linux OS and Mac OS X) or a.exe (Windows OS): ifort a.for b.for c.for.
-fpe0 or /fpe:0 restricts floating-point exceptions by enabling the overflow, the divide-by-zero, and the invalid floating-point exceptions. The program will print an error message and abort if any of these exceptions occurs. If a floating underflow occurs, the result is set to zero and execution continues. This is called flush-to-zero. This option sets -IPF_fp_speculationstrict (Linux and Mac OS) or /QIPF_fp_speculationstrict (Windows) if no specific -IPF_fp_speculation or /QIPF_fp_speculation option is specified. The -fpe0 or /fpe:0 option sets -ftz (Linux and Mac OS) /Qftz (Windows). To get more detailed location information about where the exception occurred, use -traceback (Linux and Mac OS) or /traceback (Windows).
Note
On systems based on the IA-32 and Intel® 64 architectures , explicitly setting -fpe0 or /fpe:0 can degrade performance since the generated code stream must be synchronized after each floating-point instruction to allow for abrupt underflow fix-up.
-fpe1 or /fpe:1 restricts only floating-point underflow. Floating-point overflow, floating-point divide-by-zero, and floating-point invalid produce exceptional values (NaN and signed Infinities) and execution continues. If a floating-point underflow occurs, the result is set to zero and execution continues. The /fpe:1 option sets -ftz or /Qftz.
Note
On systems based on the IA-32 and Intel® 64 architectures , explicitly setting -fpe1 or /fpe:1 can degrade performance since the generated code stream must be synchronized after each floating-point instruction to allow for abrupt underflow fix-up.
-fpe3 or /fpe:3 is the default on all processors, which allows full floating-point exception behavior. Floating-point overflow, floating-point divide-by-zero, and floating-point invalid produce exceptional values (NaN and signed Infinities) and execution continues. Floating underflow is gradual: denormalized values are produced until the result becomes 0.
The -fpe or /fpe option enables exceptions in the Fortran main program only. The floating-point exception behavior set by the Fortran main program remains in effect throughout the execution of the entire program unless changed by the programmer. If the main program is not Fortran, the user can use the Fortran intrinsic FOR_SET_FPE to set the floating-point exception behavior.
When compiling different routines in a program separately, you should use the same value of n in -fpen or /fpe:n.
An example follows:
IMPLICIT NONE
real*4 res_uflow, res_oflow
real*4 res_dbyz, res_inv
real*4 small, big, zero, scale Frp unlocker all in one tool (2017) for mac.
small = 1.0e-30
big = 1.0e30
zero = 0.0
scale = 1.0e-10
! IEEE underflow condition (Underflow Raised)
res_uflow = small * scale
write(6,100)'Underflow: ',small, ' *', scale, ' = ', res_uflow
! IEEE overflow condition (Overflow Raised)
res_oflow = big * big
write(6,100)'Overflow: ', big, ' *', big, ' = ', res_oflow
! IEEE divide-by-zero condition (Divide by Zero Raised)
res_dbyz = -big / zero
write(6,100)'Div-by-zero: ', -big, ' /', zero, ' = ', res_dbyz
! IEEE invalid condition (Invalid Raised)
res_inv = zero / zero
write(6,100)'Invalid: ', zero, ' /', zero, ' = ', res_inv
100 format(A14,E8.1,A2,E8.1,A2,E10.1)
end
Consider the following command line:
ifort fpe.f90 -fpe0 -fp-model strict -g (Linux and Mac OS)
ifort fpe.f90 /fpe:0 /fp:strict /traceback (Windows)
Output similar to the following should result:
Windows:
Underflow: 0.1E-29 * 0.1E-09 = 0.0E+00
forrtl: error (72): floating overflow
Image PC Routine Line Source
fpe.exe 0040115B Unknown Unknown Unknown
fpe.exe 0044DFC0 Unknown Unknown Unknown Google chrome mac os x 10.4 11.
fpe.exe 00433277 Unknown Unknown Unknown
kernel32.dll 7C816D4F Unknown Unknown Unknown
Linux and Mac OS:
./a.out
Underflow: 0.1E-29 * 0.1E-09 = 0.0E+00
forrtl: error (72): floating overflow
Image PC Routine Line Source
a.out 0804A063 Unknown Unknown Unknown
a.out 08049E78 Unknown Unknown Unknown
Unknown B746B748 Unknown Unknown Unknown
a.out 08049D31 Unknown Unknown Unknown
Aborted
The following command line uses /fpe1:
ifort fpe.f90 -fpe1 -g (Linux and Mac OS)
ifort fpe.f90 /fpe:1 /traceback (Windows)
The following output is produced:
Underflow: 0.1E-29 * 0.1E-09 = 0.0E+00
Overflow: 0.1E+31 * 0.1E+31 = Infinity
Div-by-zero: -0.1E+31 / 0.0E+00 = -Infinity
Invalid: 0.0E+00 / 0.0E+00 = NaN
The following command line uses /fpe3:
Does mac select cost money. ifort fpe.f90 -fpe3 -g (Linux and Mac OS)
ifort fpe.f90 /fpe:3 /traceback (Windows)
The following output is produced:
Underflow: 0.1E-29 * 0.1E-09 = 0.1E-39
Overflow: 0.1E+31 * 0.1E+31 = Infinity
Div-by-zero: -0.1E+31 / 0.0E+00 = -Infinity
Invalid: 0.0E+00 / 0.0E+00 = NaN
Previous | Next | Guide to compilers for GEOS-Chem
- The Intel Fortran compiler (aka ifort)
This page contains information about the Intel Fortran compiler (aka 'IFORT' compiler).
The Intel Fortran compiler is our recommended proprietary compiler for GEOS-Chem.
- 1Overview
- 2Environment settings for Intel Fortran
- 4Optimization
- 4.2Recommended compilation and optimization options for GEOS-Chem
- 4.3Precision-safe optimization
- 5Known issues
Overview
An Intel license is required
Intel Fortran requires the purchase of an expensive site license. If your institution does not have the resources to purchase the Intel Fortran Compiler, then we recommend that you use the GNU Fortran compiler—which is free and open source—instead.
Documentation
You can find more information about the Intel Fortran Compiler here:
Also, normally when you installs the Intel Fortran compilers, you also will install the C and C++ compilers.
--Bob Yantosca (talk) 19:44, 10 January 2019 (UTC)
Intel Fortran Compiler versions that have been tested with GEOS-Chem
The GEOS-Chem Support Team has tested GEOS-Chem with the compiler versions listed below. But you should be able to use other Intel Fortran Compiler versions as well.
NOTE: Certain Intel Fortran compiler versions are known to cause issues. Please see our Known issues caused by compiler bugs wiki page more information.
Platform | Compiler | Status |
---|---|---|
Linux | ifort 19.0.5.281 | Supported |
Linux | ifort 18.0.5 | Supported |
Linux | ifort 17.0.4 | Supported |
Linux | ifort 15.0.0 and similar builds | Supported
|
Linux | ifort 13.0.079 and similar builds | Supported |
Linux | ifort 12 | Supported
|
Linux | ifort 11.1.069 and similar builds | Supported |
--Bob Yantosca (talk) 19:13, 22 May 2020 (UTC)
Environment settings for Intel Fortran
Here is some information about how you can customize your Unix environment to use the Intel Fortran compiler.
Using a module manager to load GNU Fortran and related libraries
On many computer systems, a module manager such as Lmod or environment-modules can be used to load the Intel Fortran compiler library (and its dependencies) into your Unix environment. For example, we use the following commands on the Harvard cluster (cannon.rc.fas.harvard.edu):
You can ask your IT staff what the corresponding commands would be on your particular cluster.
Your module manager *may* automatically define several environment variables for you:
Variable | Expected setting | Description |
---|---|---|
FC | ifort | Name of the Intel Fortran compiler |
CC | icc | Name of the Intel C compiler |
CXX | icpc | Name of the Intel C++ compiler |
NETCDF_HOME | System-dependent | Path to the root netCDF folder |
NETCDF_INCLUDE | System-dependent | Path to the netCDF include folder (e.g. $NETCDF_HOME/include) |
NETCDF_LIB | System-dependent | Path to the netCDF library folder (e.g. $NETCDF_HOME/lib or $NETCDF_HOME/lib64) |
NETCDF_FORTRAN_HOME | System-dependent | Path to the root netCDF Fortran folder |
NETCDF_FORTRAN_INCLUDE | System-dependent | Path to the netCDF Fortran include folder (e.g. $NETCDF_FORTRAN_HOME/include) |
NETCDF_FORTRAN_LIB | System-dependent | Path to the netCDF Fortran library folder (e.g. $NETCDF_FORTRAN_HOME/lib or $NETCDF_FORTRAN_HOME/lib64) |
If these variables are not automatically set by the module command on your system (or if your system does not use a module manager, then you will need to set these manually . Please see our Getting Started with GEOS-Chem: Configuring your computational environment for detailed instructions.
--Bob Yantosca (talk) 21:35, 16 January 2020 (UTC)
Requesting sufficient stack memory for GEOS-Chem
In order to run GEOS-Chem with Intel Fortran, you must request the maximum amount of stack memory in your Unix environment. If you do not request the maximum amount of stack memory, then your GEOS-Chem simulation will not have enough memory to create temporary variables (including those created within OpenMP parallel loops).
For detailed instructions, please see Getting Started with GEOS-Chem: Specifying settings for OpenMP parallelization.
--Bob Yantosca (talk) 21:36, 16 January 2020 (UTC)
Performance
Please see our Guide to GEOS-Chem performance for a summary of recent timing tests done with the Intel Fortran compiler.
--Bob Yantosca (talk) 19:29, 10 January 2019 (UTC)
Optimization
In this section we present information about the various optimization options available in the Intel Fortran Compiler.
Optimization options
Here is a quick reference table of optimization options (taken from the online Intel Fortran Compiler User and Reference Guides.
Option | Description | How invoked in GEOS-Chem? |
---|---|---|
-O0 | Turns off all optimizations. Math expressions will be evaluated in the same order in which they are written, which is necessary for debugging. If you are using a debugger (such as Totalview), compile with -g -O0. | DEBUG=yes or OPT=-O0 |
-O1 | Enables optimizations for speed and disables some optimizations that increase code size and affect speed. The -O1 option may improve performance for applications with very large code size, many branches, and execution time not dominated by code within loops. Setting -O1 automatically sets the following options:
| OPT=-O1 |
-O2 (aka -O) | Enables optimizations for speed. This is the generally recommended optimization level. This option also enables:
On Linux and Mac OS X systems, if -g is specified, -O2 is turned off and -O0 is the default unless -O2 (or -O1 or -O3) is explicitly specified in the command line together with -g. | Default setting |
-O3 | Enables -O2 optimizations plus more aggressive optimizations, such as prefetching, scalar replacement, and loop and memory access transformations. Enables optimizations for maximum speed, such as:
On Linux and Mac OS X systems, the -O3 option sets option -fomitframe-pointer. The -O3 optimizations may not cause higher performance unless loop and memory access transformations take place. The optimizations may slow down code in some cases compared to -O2 optimizations. The -O3 option is recommended for applications that have loops that heavily use floating-point calculations and process large data sets. | OPT=-O3 |
--Bob Y. 11:14, 3 October 2013 (EDT)
Recommended compilation and optimization options for GEOS-Chem
In this section, we present information about the compilation and optimization options that are invoked when you compile a GEOS-Chem simulation.
List of commonly-used compilation options
Here are the IFORT compilation options currently used by GEOS-Chem:
Option | Description | How invoked in GEOS-Chem? |
---|---|---|
Normal compiler settings | ||
-cpp | Turns on the C-preprocessor, to evaluate #if and #define statements in the source code. | Default setting |
-w | Suppresses all compiler warnings. This is mainly a convenience to prevent excessive output to the screen or log file. NOTE: Most compiler warnings are harmless. Execution does not stop when a warning is displayed, unlike an error message, which causes program execution to halt at the point where the error occurred. | Default setting |
-O2 | Optimizes the source code for speed, without taking too many liberties with numerical precision. For more information, please see the optimization options section above. | Default setting |
-auto | This option places local variables (scalars and arrays of all types), except those declared as SAVE , on the run-time stack. It is as if the variables were declared with the AUTOMATIC attribute. It does not affect variables that have the SAVE attribute or ALLOCATABLE attribute, or variables that appear in an EQUIVALENCE statement or in a common block. | Default setting |
-noalign | Prevents the compiler from padding bytes anywhere in common blocks and structures. Padding can affect numerical precision. | Default setting |
-convert big_endian | Specifies that the format will be big endian for integer data and big endian IEEE floating-point for real and complex data. This only affects file I/O to/from binary files (such as binary punch files) but not ASCII, netCDF, or other file formats. | Default setting |
-vec-report0 | Tells the compiler to suppress printing 'LOOP HAS BEEN VECTORIZED' messages. This reduces the amount of output that is sent to the screen and/or GEOS-Chem log file. | Default setting |
-fp-model source | Rounds intermediate results to source-defined precision and enables value-safe optimizations. Basically, this tells the compiler not to take too many liberties with how numerical expressions are evaluated. For more information about this option, please see our precision-safe optimization section below. This option can be disabled by compiling GEOS-Chem with the PRECISE=no Makefile option. | Default setting |
-traceback | This option tells the compiler to generate extra information in the object file to provide source file traceback information when a severe error occurs at run time. When the severe error occurs, source file, routine name, and line number correlation information is displayed along with call stack hexadecimal addresses (program counter trace). This option increases the size of the executable program, but has no impact on run-time execution speeds. It functions independently of the debug option. |
|
Special compiler settings | ||
-r8 | This option tells the compiler to treat variables that are declared as REAL as REAL*8 (as opposed to REAL*4 .NOTE: This option is not used globally, but is only applied to certain indidvidual files (mostly from third-party codes like ISORROPIA. Current GEOS-Chem programming practice is to use either | Used as needed |
-mcmodel=medium | This option is used to tell IFORT to use more than 2GB of static memory. This avoids a specific type of memory error that can occur if you compile GEOS-Chem for use with an extremely high-resolution grid (e.g. 0.25° x 0.3125° nested grid). | Default setting |
-shared-intel (formerly -i-dynamic) | This option needs to be used in conjunction with -mcmodel=medium. It causes Intel-provided libraries to be linked in dynamically instead of statically (which is the default). | Default setting |
-ipo | This option enables interprocedural optimization between files. This is also called multifile interprocedural optimization (multifile IPO) or Whole Program Optimization (WPO). When you specify this option, the compiler performs inline function expansion for calls to functions defined in separate files. NOTE: Yuxuan Wang found that this option was useful for certain nested-grid simulations. See the this wiki post below for more information. | IPO=yes |
-static | This option prevents linking with shared libraries. It causes the executable to link all libraries statically. NOTE: Yuxuan Wang found that this option was useful for certain nested-grid simulations. See the this wiki post below for more information. | IPO=yes |
Settings only used for debugging | ||
-debug all | Tells the compiler turn on all debug error output. | DEBUG=yes |
-g | Tells the compiler to generate full debugging information in the object file. This will cause a debugger (like Totalview) to display the actual lines of source code, instead of hexadecimal addresses (which is gibberish to anyone except hardware engineers). | DEBUG=yes |
-O0 | Turns off all optmization. Source code instructions (e.g. DO loops, IF blocks) and numerical expressions are evaluated in precisely the order in which they are listed, without being internally rewritten by the optimizer. This is necessary for using a debugger (like Totalview). | DEBUG=yes |
-check bounds (aka -CB) | Check for array-out-of-bounds errors. This is invoked when you compile GEOS-Chem with the BOUNDS=yes Makefile option. NOTE: Only use -CB for debugging, as this option will cause GEOS-Chem to execute more slowly! | DEBUG=yes |
-check arg_temp_created | Checks to see if any array temporaries are created. Depending on how you write your subroutine and function calls, the compiler may need to create a temporary array to hold the values in the array before it passes them to the subroutine. For detailed information, please see our Passing array arguments efficiently in GEOS-Chem wiki page. | DEBUG=yes |
-fpe0 | This option will cause GEOS-Chem to halt if any type of floating-point error is encountered. This can happen if an equation results in a denormal value, e.g. NaN, or +/-Infinity. Common causes of floating-point errors are divisions where the denominator becomes zero. NOTE: The default compiler setting is -fpe3, which will convert many of these denormal values to zeros and then continue execution. | FPE=yes |
-ftrapuv | This option will assign a large numeric value to all local automatic variables. This makes it easier to identify numerical errors caused by improper initialization. | FPE=yes |
--Bob Y. 11:21, 3 October 2013 (EDT)
Typical settings for a GEOS-Chem simulation
The normal GEOS-Chem build uses the following IFORT compiler flags:
whereas a debugging run (meant to execute in a debugger such as TotalView) will typically use these flags:
NOTE: In order to avoid running out of memory if you compiling GEOS-Chem at extremely high resolution (e.g. the 0.25° x 0.3125° nested grids), we recommend adding the following flags:
These are automatically set when you compile with the NETCDF=yes or HDF=yes compiler options (in GEOS-Chem v9-01-03 and higher).
--Bob Y. 17:34, 29 February 2012 (EST)
Precision-safe optimization
You can use the following Intel Fortran Compiler options to select how aggressively you would like to optimize floating-point operations.
Default behavior
-fp-model fast
Example source code:
When this option is specified, the compiler applies the following semantics:
- Additions may be performed in any order
- Intermediate expressions may use single, double, or extended precision
- The constant addition may be pre-computed, assuming the default rounding mode
Using these semantics, the following shows some possible ways the compiler may interpret the original code:
or
Preferred alternative
-fp-model source (aka -fp-model precise)
Example source code:
When this option is specified, the compiler applies the following semantics:
- Additions are performed in program order, taking into account any parentheses
- Intermediate expressions use the precision specified in the source code
- The constant addition may be pre-computed, assuming the default rounding mode
Using these semantics, the following shows a possible way the compiler may interpret the original code:
Summary
If you do not select any -fp-model option, the Intel Fortran Compiler will default to -fp-model fast. As you can see from the examples above, this may not optimize the code in the same way each time. This can lead to minor numerical noise in the output, as was seen in ISORROPIA II.
To avoid this situation, we recommend compiling all source code files with -fp-model source. This will be the new default in GEOS-Chem v9-01-02.
Reference: Intel® Fortran Floating-point Operations; Document Number: 315892-003US
--Bob Y. 17:01, 25 August 2011 (EDT)
Optimization options for faster runs
Yuxuan Wang told us about the optimization options: -ipo and -static and said these options would speed up the simulations. I've tested these options on our system at Harvard. The run with the new options show very tiny differences (much less than 1% over 1 month) compared to a run with the old options only. For a full-chemistry run (43 tracers) on 4x5 resolution and 4 processors, the run time is about 10% shorter than previously.
These options are especially efficient to handle the transport. So in simulations with a faster chemistry (like tagged tracers simulations), we expect to see a higher gain in time. For example, the time for a methane run is shorten by about 30 %.
To use these options, compile GEOS-Chem with the IPO=yes Makefile option, e.g.
--Ccarouge 15:54, 8 September 2009 (EDT)
--Bob Y. 17:50, 29 February 2012 (EST)
Optimization level for debugging
If you would like to run your code in a debugger, such as Totalview, you must use the following compiler switches:
Using -O0 will ensure that the source code gets executed in the same order in which it is written (i.e. this disables all compiler optimizations). The -g switch will tell the debugger to display lines of source code instead of hexadecimal memory addresses (which are more or less gibberish unless you are a hardware engineer).
GEOS-Chem will add these switches automatically for you if you compile with the DEBUG=yes option.
--Bob Y. 15:28, 22 February 2012 (EST)
Caveat about optimizing for specific chipsets
The standard GEOS-Chem build sequence does not include any optimization flags that are specific to a certain type of CPU. If you are interested, you can certainly experiment for yourself. But be aware that this may invoke certain chip-level optimizations that could potentially change the simulation output.
Jenny Fisher wrote:
I have tested the new chips & compiler option. I found that there are small differences [in difference test output].if I use exactly the same compile commands and number of processors between our old cores and our new Broadwell cores (E5-2690 v4). The differences are very small and I think nothing to worry about.However, adding the preferred compiler flag -xCORE-AVX2 led to much bigger differences (e.g., up to 5% difference or 10 ppb in ozone…). I haven't investigated the differences in detail. I did run a one month benchmark comparison, and see that the differences can be consequential after a month (i.e. not just differences in regions where values are low.
I have no idea what is causing these differences. So I guess for the moment, I would recommend *not* using the specific optimisation for Broadwell/Haswell cores. However, I think it probably is ok to use the Broadwell cores without this flag. I am not sure what impact this choice will have on performance.--Bob Yantosca (talk) 14:50, 28 March 2017 (UTC)
Known issues
The following issues occur with certain versions of the Intel Fortran Compiler. Several of these issues have now been resolved in the most recent GEOS-Chem versions. We recommend upgrading to the most recent public release if possible.
Compilation issues with Intel Fortran 18
This update (Git ID: 9833fff4) was included in GEOS-Chem 12.0.1, which was released on 24 Aug 2018.
The GEOS-Chem Support Team recently tested GEOS-Chem 12.0.0 with ifort 18, which is a recent release of the Intel Fortran Compiler. An 'out-of-the-box' compilation with ifort 18 resulted in these errors:
Location | Problem | Solution |
---|---|---|
Makefile_header.mk | The -openmp switch has been retired in Intel Fortran 18. The new option to turn on OpenMP parallelization is now called -qopenmp. | Now issue a command to get the compiler version. This is saved into the COMPILER_VERSION variable in Makefile_header.mk. If COMPILER_VERSION is 18 or higher, use -qopenmp to activate OpenMP. Otherwise use -openmp. |
--Bob Yantosca (talk) 15:51, 10 September 2018 (UTC)
Cannot compile GEOS-Chem v10-01 with Intel Fortran Compiler v17
This issue is now resolved in GEOS-Chem v11-01.
Myroslav Hordiichuk wrote:
When compiling with make -j2 MET=geosfp GRID=4x5 UCX=y CHEM=benchmark there is a problem, I can't fix. The screenshots are attached.
Bob Yantosca wrote:
Thanks for writing. [Because.] you [are] using a very new version of the Intel compiler (v2017). then that may be not able to parse the module interfaces in NcdfUtil/ncdf_mod.F90 and in HEMCO/Core/hco_diagn_mod.F90. We recently encountered this issue when trying to port GEOS-Chem v11-01 to the GNU Fortran compiler. I ended up rewriting code in these modules to avoid this issue.What I think is happening is that, like GNU Fortran, Intel Fortran 2017 is by default using the newer Fortran 2003 or Fortran 2008 standard. This standard is more strict than the Fortran-90 language standard when it comes to the module interfaces. Long story short: I had to rewrite the DIAGN_UPDATE
interface in HEMCO/Core/hco_diagn_mod.F90 to remove the OPTIONAL
array arguments. This got the code to compile with GNU Fortran. Basically, instead of having only 2 subroutines contained in the DIAGN_UPDATE
module interface, I had to have 6. The arguments Scalar
, Array2D
, Array3D
cannot be OPTIONAL
arguments in this case. This syntax used to be OK in the original Fortran-90 standard but evidently not in the newer F2003 or F2008 standards. We also had to modify a similar module interface in NcdfUtil/ncdf_mod.F90 accordingly.
--Bob Yantosca (talk) 20:47, 20 January 2017 (UTC)
Resetting stacksize for Linux
Overall machine memory limits are set with the Unix limit command. If you use csh or tcsh, you can set the following commands in your ~/.cshrc file:
Or if you use bash, you can add these to commands your ~/.bashrc file:
NOTE: depending on your particular OS build (Linux, CentOS, Fedora, Ubuntu), not all of these limits will be used.
It is important to set the stacksize memory to the maximum value, because this will determine the amount of memory available for temporary variables, which are:
- variables that are not declared with the SAVE attribute
- variables that are not located at the top of a module
- variables that are not located in a common block.
However, one quirk is that the stacksize memory for child processes (i.e. processes spawned by CPUs within !$OMP PARALLEL DO
loops) are not set by the stacksize limit, but instead by the OMP_STACKSIZE environment variable. If OMP_STACKSIZE is not set with a high enough value, then your GEOS-Chem simulation may think it doesn't have enough memory to proceed, and may die with a segmentation fault error.
The fix for this situation is to make sure that you set OMP_STACKSIZE to a high value. It's OK if the value you give to OMP_STACKSIZE is larger than the largest amount of memory on your system. As long as it's set to a high positive number it will work.
If you use csh or tcsh, you can add this command to your ~/.cshrc file.
Or if you use bash, add this command to your ~/.bashrc file:
Resetting the OMP_STACKSIZE environment variable in this manner usually will correct the following errors:
NOTE: We now recommend that you use the OMP_STACKSIZE variable instead of the KMP_STACKSIZE variable. This will make it easier to switch between compilers on your system. OMP_STACKSIZE works with all compilers, but KMP_STACKSIZE only works with Intel Fortran.
--Bob Y. 14:05, 5 November 2014 (EST)
Out of memory asking for NNNNN
Debra Weisenstein wrote:
- I'm trying to compile GEOS-CHEM (the stratospheric beta version that Seb Eastham at MIT is working on) on hpc at Harvard, and though it compiled for me early before, I've now been getting an 'out of memory' error every time it tries to compile GeosCore/diag49_mod.F. I changed the compile option for mcmodel to large. Here is the compile line and error.
Fort Macon Marina
- I've gone back and tried compiling older versions that I've compiled and run before and get a similar error:
- Any idea what is going on? The 'top' command shows 301780k of free memory (out of 8GB) and 906164k of free swap.
Bob Yantosca wrote:
- I googled the error message ('Out of memory asking for ….') and I found this post online. The person who replied to this post gave 3 suggestions:
- Use -O2.
- Split up the source into smaller files if possible (may be difficult if it's a module.)
- Raise your datasize limit
- So it may be that your datasize limit is not maxed out in your computational environment. You can check the setting of the system limits by typing:
- at the Unix prompt. You'll get a screen like this:
- Depending on your platform (i.e. combination of hardware + operating system) you may not have all of these limits available. The two most important limits for GEOS-Chem are datasize and stacksize. These should be set to unlimited. For good measure, you can also set the other available limits for your platform to unlimited as well. Follow the instructions in this wiki post. One caveat: if you try to set a limit that does not exist on your platform to unlimited, you will get a warning message from the Unix shell. You can then delete the offending limit setting from your .cshrc or .bashrc file.
--Bob Y. 14:58, 25 July 2013 (EDT)
Bug fix for GEOS-Chem compiled with Intel Fortran Compiler 12
Prasad Kasibhatla wrote:
- I am experiencing the IFORT 12 compile failure of GEOS-Chem v9-01-03 with nested grid options turned on in define.h. The compile seems to be failing in strat_chem_mod.F90 (see end of this msg). And to add - the compile succeeds for v9-01-03l with IFORT 12. I noticed that strat_chem_mod.F90 appeared post-v9-01-03l.
Bob Yantosca wrote:
- I have this version of IFORT installed locally:
- And I got the exact same error as you did. I looked on the internet and it seemed that the issue is related to how IFORT 12 deals w/ OpenMP DO loops (click HERE and HERE).
- A little brute-force debugging revealed the offending parallel DO loop in strat_chem_mod.F90, in routine CALC_STE:
- I also noticed that this DO loop is only executed for global simulations. Immediately preceding it there is an #if statement that exits out of the routine if you are doing an nested grid run:
- So given that this code doesn't get used for nested runs, I added an #else block so that the code isn't compiled in when you compile for nested grids:
- I've compiled this with the NA nested grid, the CH nested grid, and for 2 x 2.5 and 4 x 5 global GEOS-5 simulations, and the code does not get that compilation error. I've pushed this to the GEOS-Chem repository as a last-minute bug fix. I'm not sure why the error happens in the first place w/ the IFORT 12 compiler, but in any case, this fixes it.
--Bob Y. 13:52, 5 October 2012 (EDT)
Error compiling with IFORT 12 and Mac OS X
David Lary wrote:
- I am using Intel(R) 64, Version 12.1.3.289 Build 20120130 on Mac OS X 10.7.4. When I typed make, I get the following error:
Bob Yantosca replied:
- I think this is an issue particular to IFORT running on Mac OS X. It is probably not finding the proper library. I found a couple of posts (post #1 and post #2) on the internet that describe this error.
- Post #2 recommends adding the following path to your LD_LIBRARY_PATH environment variable:
--Bob Y. 10:43, 26 June 2012 (EDT)
Compilation error with IFORT 12
Geert Vinken wrote:
- I was trying to compile the GEOS-Chem v9-01-03k version, but I was getting a strange error when trying to do this with ifort 12.1.2:
- Compiling with ifort 11.1 shows no problems at this module. Any of you ever heard of this problem or know a way around it? If I uncomment this line the model compiles, so it's just this line (and apparently the combination of 8_1x1 =) that's causing the problem.
- Also, renaming the GENERIC8_1x1 array to GENERIC8 seems to solve the problem.
Bob Yantosca replied:
- I think the default behavior of IFORT 12.1 has changed to adopt one of the more modern Fortran standards (F2003 or F2008). I found this entry in the Intel Fortran Compiler 12.1 manual which describes how one specifies floating-point constants.
- As shown in the entry above, the IFORT 12 compiler now interprets an underscore immediately preceded and followed by numbers as an alternate way to specify numbers in scientific notation. For example, the traditional way of representing a REAL*4 and REAL*8 constants:
- can now also be represented as:
- Therefore, the string:
- is probably now parsed by the compiler as an exponential. This more than likely separates the string into substrings GENERIC8_1x1 and x1, which is what generates the error.
- I have looked for a compiler switch in the IFORT 12.1 manual that would disable this feature, but have not been successful. The short-term solution is to rename the variables so that it does not have an underscore surrounded by two numbers (e.g. from GENERIC8_1x1 to GENERIC_8_1x1).
--Bob Y. 18:13, 23 May 2012 (EDT)
Relocation truncated to fit error
If your code uses many large arrays, or if you are compiling an ultra-fine resolution version of GEOS-Chem (e.g. a 0.25° x 0.3125° GEOS-FP nested grid), then you may see this type of error:
The wording you get may differ slightly than the example shown above.
Long story short: IFORT is telling you that your program is trying to use more than 2GB of statically-allocated data (i.e. data space that is not declared with an ALLOCATABLE statement) at compile time. The default setting in IFORT is to expect to use less than 2GB of memory, so you are hitting the upper limit.
The solution is simple: recompile your code with the following compiler flags:
The -mcmodel=medium flag will tell IFORT that you expect to use more than 2GB of statically-allocated memory in your program. However, this also requires that you use link using dynamic libraries instead of the normal shared libraries. Using the -shared-intel flag will turn on the dynamic library linking. (Starting with GEOS-Chem v9-01-03, these compiler flags will be applied to the build sequence automatically.)
IMPORTANT NOTE! If your code links to any libraries such as HDF or netCDF, then you MUST rebuild each library, making sure that the Fortran and C compilers use the -mcmodel=medium option. Please see our Installing libraries for GEOS-Chem page for examples.
GEOS-Chem v9-01-03 and higher will automatically set these flags for you.
For more information, please see the following links:
--Bob Y. (talk) 18:24, 25 September 2015 (UTC)
Problems with IFORT 11.0.xxx
You should use GEOS-Chem with IFORT 11.1.058 or higher versions. Please see the discussion below about problems in the earlier versions of IFORT 11.0.xxx:
Tzung-May Fu wrote:
- I tested the Intel Fortran v11.0.074 compiler, but found that it is incompatible with the GC code. This is related to the partition.f bug that I reported earlier. (Actually, I'm not sure there is a bug in partition.f any more, unless you have also run into it with IFORT v10).
- I ran a 1-day simulation, using Bob's v8-01-03 standard run release, with no change at all. Using Intel Fortran v10.1.015, I was able to replicate Bob's standard run. However, when I switched to Intel Fortran v11.0.074, I ran into the error in partition.f, due to the CONCNOX-SUM1 < 0d0 check. Here's the error message in log:
- I then tried Bob's fix to partition.f. This time the run finishes, warning the user about the CONCNOX-SUM1 < 0d0 issue. But the output result is completely wacky!!! Below you can compare the surface Ox concentrations, using
- (A) IFORT v10
- (B) IFORT v11 and the partition.f fix
- The (B) spatial pattern is completely off. NOx is also affected and shows the similar weird pattern.
- I'm pretty sure the problem is in the chemistry part. I've tried turning off the optimization but the problem persists. Perhaps there is some problem with the way IFORTv11 treats floating points? Also, I am not sure if IFORTv11 caused the weird model result, or if IFORTv11 caused some issues in chemistry, and the partition.f 'fix' subsequently lead to the weird result.
- Long story short, it seems like IFORTv11 is not a good choice for now, and that the 'fix' to partition.f should not be implemented.
Philippe Le Sager wrote:
- Thanks for testing Ifort11. We did run into the partition bug with Ifort10 after fixing tpcore. So I doubt that the weird result is related to that partition fix, and it is probably just a problem with IFORT 11.
Bob Yantosca wrote:
- You might have to go thru the IFORT 11 manuals to see if any default behavior has changed (i.e. optimization, compiler options, etc). It may not just be the concnox thing but something else in the numerics that is particular to IFORT 11.
- There is usually a 'What's new' document w/ every Intel compiler release. Maybe that has some more information, you could look at it.
Bob Yantosca wrote:
- I've also heard from some folks @ NASA that IFORT 11.0 was problematic. They claim that IFORT 11.1 is much better. You may want to look into this in the meantime.
--Bob Y. 16:50, 7 October 2009 (EDT)
Eric Sofen wrote:
- Both Becky Alexander and I have run into problems with IFORT 11.1. When either of us run offline aerosol simulations compiled on IFORT 11.1, the simulation compiles and runs without errors, but the sulfur budgets are way off. The problems seem to be occurring in the deposition code, as Becky's simulations end up with very little deposition, but at the same time, the S burdens are too low. In my case, the deposition ends up being an order of magnitude too high. Changing back to IFORT 10 fixed both of these problems.
How to install printer driver on mac. --Eric Sofen 13:32, 22 October 2009
Yuxuan Wang wrote:
- From our interaction with the Intel people, ifort 11.1.056 should work for GEOS-Chem. The GC version we tested at Tsinghua is v8-02-01 (nested-grid China with GEOS-5 meteorology). The platform we tested is Nehalem from Intel, with the following compilation options:
- Not sure whether these options will work for Mac OSX. From the testing, we found that codes compiled with ifort 11.1.056 ran at 2% faster than ifort 10.1.008.
--Bob Y. 14:59, 4 November 2009 (EST)
Problem with IFORT 11 and GEOS-Chem adjoint
Nicolas Bousserez wrote:
- We have been struggling for some time with the following problem when running GC adjoint (v8-02-01):
- but found libguide.so already initialized'.
- After some investigations it seems like it is a linker error generated when different parts of the program try to link both static and dynamic verions of the OpenMP runtime. There is an option in ifort 11 to have openmp linked statically, which theoretically should fix this problem.
- But using ifort 11 for GC seems to cause other problems and this compilation option doesn't exist with ifort 10. The fact is that Daven Henze, who is using ifort 10 and a linux platform similar to ours never got the above problem. Has anyone got this error before? My platform configuration is the following:
- If anyone running GC adjoint has a similar configuration please let me know what is your Makefile (using netcdf libraries) configuration and which version of ifort you're using so that I can do some testing.
--Bob Y. 09:43, 8 April 2011 (EDT)
Incompatibility between IFORT 11 and OS version
If you are using a the Intel Fortran Compiler version 11, you may encounter some incompatibilities with your operating system, which might require an OS upgrade.
Nicolas Bousserez wrote:
- We have been struggling for some time with the following problem when running GC adjoint (v8-02-01). We get this error:
- After some investigations it seems like it is a linker error generated when different parts of the program try to link both static and dynamic verions of the OpenMP runtime. There is an option in ifort 11 to have openmp linked statically, which theoretically should fix this problem. But using ifort 11 for GC seems to cause other problems and this compilation option doesn't exist with ifort 10. The fact is that Daven Henze, who is using ifort 10 and a linux platform similar to ours never got the above problem. Has anyone got this error before? My platform configuration is the following:
Nicolas Bousserez wrote:
- For what it's worth, this is the oldest OS we're using:
- and this is the newest:
- What is relevant is the 2.6.9 and the 2.6.18, not the compilation dates. It means you're running something equivalent to a RHEL4/CentOS-4 kernel instead of RHEL5/CentOS-5, which has implications for your libraries, compatibility, bugs, security, etc. I would guess that you've been updating RHEL4 (first released 2005) or equivalent for several years, and ifort11 was released during the era of RHEL5 (first released 2007), so it wouldn't be too surprising if there were a library incompatibility. I don't know whether that is the cause of your symptom, but it might be. (RHEL6 was released late last year and v12 of the Intel compilers have also been released. CentOS-6 will be out soon.)
- You can continue to use the older compiler with the older OS, but I'd recommend upgrading the OS, which is worth doing anyway.
--Bob Y. 10:25, 13 April 2011 (EDT)
Speedup With Hyperthreading on Nehalem chips
Hyperthreading is when a job uses more threads than there are actual CPU cores. I've noticed that using 16 threads ($OMP_NUM_THREADS = 16) on an 8-core system (2 x quad core Intel Nehalem X5570's) leads to a 15% speedup over using 8 threads. These tests were with GEOS-Chem v8-02-03, full chemistry, 2x2.5, ifort 10.1.021, and
Fort Machine
This does not have a positive impact when using earlier generations of Intel chips (Harpertown or Clovertown).
--Daven Henze 1:42, 16 December 2009 (MDT)
Performance bottleneck caused by inefficient subroutine calls
Special care has to be taken when passing pointer arrays or sub-fields of dervied type objects to subroutines. If this is done incorrectly, it can cause a huge performance slowdown. Please see the discussion on our Passing array arguments efficiently in GEOS-Chem wiki page for full details.
--Bob Y. 10:49, 10 June 2013 (EDT)
Bugs in the IFORT compiler cause HEMCO to segfault
The GEOS-Chem Support Team has recently determined that bugs in the Intel Fortran Compiler versions 14 and 15 have caused the Harvard-NASA Emissions Component (aka HEMCO) to halt with segmentation faults. For more information, please see these wiki posts:
--Bob Y. (talk) 20:38, 25 August 2015 (UTC)
Fort Macomb
Previous | Next | Guide to compilers for GEOS-Chem