CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   OpenFOAM Installation (http://www.cfd-online.com/Forums/openfoam-installation/)
-   -   Compilation on SGI Altix using intel compiler (http://www.cfd-online.com/Forums/openfoam-installation/57632-compilation-sgi-altix-using-intel-compiler.html)

markusuhlmann July 13, 2006 05:14

I run into the following error
 
I run into the following error when compiling on an SGI Altix system (IA-64 processors, linux kernel 2.4.21), using the intel 9.0 compiler:

------------------------------------------------
...
SOURCE_DIR=OSspecific/Unix/signals
SOURCE=OSspecific/Unix/signals/sigFpe.C ; icpc -gcc-version=330 -DlinuxIA64 -DDP -wd654,819,1125,1476,1505,1572 -O3 -ftz -DNoRepository -DWM_PROJECT_VERSION='"'1.3'"' -I/disco07/c_fosiles/uhlmann/OpenFOAM/OpenFOAM-1.3/src/zlib-1.2.1 -I/disco07/c_fosiles/uhlmann/OpenFOAM/OpenFOAM-1.3/src/OpenFOAM/lnInclude -IlnInclude -I. -KPIC -c $SOURCE -o Make/linuxIA64I64DPOpt/sigFpe.o
OSspecific/Unix/signals/sigFpe.C(88): error: identifier "uint64_t" is undefined
const uint64_t sNAN = 0x7ff7ffffffffffffllu;
^

OSspecific/Unix/signals/sigFpe.C(92): error: identifier "uint64_t" is undefined
uint64_t* dPtr = reinterpret_cast<uint64_t*>(result);
^

OSspecific/Unix/signals/sigFpe.C(92): error: identifier "dPtr" is undefined
uint64_t* dPtr = reinterpret_cast<uint64_t*>(result);
^

OSspecific/Unix/signals/sigFpe.C(92): error: "uint64_t" is not a type name
uint64_t* dPtr = reinterpret_cast<uint64_t*>(result);
^

compilation aborted for OSspecific/Unix/signals/sigFpe.C (code 2)
--------------------------------------------

any idea?

(one suggestion might be "use the gnu compiler", but we are still having trouble installing a more recent version)

thanks in advance.

mattijs July 13, 2006 11:56

Try replacing uint64_t with 'u
 
Try replacing uint64_t with 'unsigned long long' (or anything else which is 64 bit unsigned integer)

markusuhlmann July 17, 2006 11:14

Yes, replacing uint64_t with '
 
Yes, replacing uint64_t with 'unsigned long long' indeed worked.

Thanks!

markusuhlmann July 18, 2006 05:47

After the above modification,
 
After the above modification, I finished the compilation of openfoam.

Then I started with the example "icoFoam" from the tutorial, as suggested
in the "README". "blockMesh . cavity" generates a singular "points" file,
as follows:

-------------------------------------------------------------------------------
/*---------------------------------------------------------------------------*\
| ========= | |
| \ / F ield | OpenFOAM: The Open Source CFD Toolbox |
| \ / O peration | Version: 1.3 |
| \ / A nd | Web: http://www.openfoam.org |
| \/ M anipulation | |
\*---------------------------------------------------------------------------*/

FoamFile
{
version 2.0;
format ascii;

root "/disco07/c_fosiles/uhlmann/OpenFOAM/uhlmann-1.3/run/tutorials/icoFoam"
;
case "cavity";
instance ""constant"";
local "polyMesh";

class vectorField;
object points;
}

// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * //

882{(0 0 0)}

// ************************************************** *********************** //
-----------------------------------------------------------------------------
Any ideas of what is going on?

Thanks.

mattijs July 19, 2006 03:23

Possibly your run-time selecti
 
Possibly your run-time selection tables are not constructed correctly? (compiler issue)

Try running a mesh generated on another machine and see if you get any messages about non-existing boundary conditions.

markusuhlmann July 19, 2006 03:49

>Try running a mesh generated
 
>Try running a mesh generated on another machine

We did just that (icoFoam-cavity,laplacianFoam; also in parallel using SGI MPT-1.9.1) and it works fine on grids/decompositions generated on a different machine.


All times are GMT -4. The time now is 04:07.