Problems of compiling FDS5.0 with VS2005+Intel Fortran Compiler

[may...@gmail.com]

I tried to compile the FDS5.0 with Visual Studios 2005 and Intel
Visual Fortran Compiler 9.1. But I could not succeed in the
compilation. My situation is as following. Could anyone help me solve
this. Thanks a lot!

1. Create a VC++ project named as "isob" with "Static Library" checked
and "Precompiled Headers" unchecked and add "isob.c".
2. Add pp_noappend to the Preprocessor Definition and succeed in
compilation.
3. Create a Fortran console project named as "FDS5.0" and add all the
source code.
4. Set the project dependencies to isob.
5. I have compile with both debug and release.
For debug I could not compiled successfully with the result as
following:

1>Linking...
1>isob.lib(isob.obj) : warning LNK4075: ignoring '/EDITANDCONTINUE'
due to '/INCREMENTAL:NO' specification
1>MSVCRTD.lib(MSVCR80D.dll) : error LNK2005: _abort already defined in
LIBCMTD.lib(abort.obj)
1>MSVCRTD.lib(MSVCR80D.dll) : error LNK2005: _fprintf already defined
in LIBCMTD.lib(fprintf.obj)
1>MSVCRTD.lib(MSVCR80D.dll) : error LNK2005: ___iob_func already
defined in LIBCMTD.lib(_file.obj)
1>MSVCRTD.lib(MSVCR80D.dll) : error LNK2005: _fflush already defined
in LIBCMTD.lib(fflush.obj)
1>MSVCRTD.lib(MSVCR80D.dll) : error LNK2005: _malloc already defined
in LIBCMTD.lib(dbgheap.obj)
1>MSVCRTD.lib(MSVCR80D.dll) : error LNK2005: _free already defined in
LIBCMTD.lib(dbgheap.obj)
1>MSVCRTD.lib(MSVCR80D.dll) : error LNK2005: _realloc already defined
in LIBCMTD.lib(dbgheap.obj)
1>MSVCRTD.lib(MSVCR80D.dll) : error LNK2005: _printf already defined
in LIBCMTD.lib(printf.obj)
1>MSVCRTD.lib(MSVCR80D.dll) : error LNK2005: _fclose already defined
in LIBCMTD.lib(fclose.obj)
1>MSVCRTD.lib(MSVCR80D.dll) : error LNK2005: _fwrite already defined
in LIBCMTD.lib(fwrite.obj)
1>MSVCRTD.lib(MSVCR80D.dll) : error LNK2005: _fopen already defined in
LIBCMTD.lib(fopen.obj)
1>MSVCRTD.lib(ti_inst.obj) : error LNK2005: "private: __thiscall
type_info::type_info(class type_info const &)" (??
0type_info@@AAE@ABV0@@Z) already defined in LIBCMTD.lib(typinfo.obj)
1>MSVCRTD.lib(ti_inst.obj) : error LNK2005: "private: class type_info
& __thiscall type_info::operator=(class type_info const &)" (??
4type_info@@AAEAAV0@ABV0@@Z) already defined in
LIBCMTD.lib(typinfo.obj)
1>LINK : warning LNK4098: defaultlib 'MSVCRTD' conflicts with use of
other libs; use /NODEFAULTLIB:library
1>Debug/FDS5.0.exe : fatal error LNK1169: one or more multiply defined
symbols found


For the release, aftering waiting for several minutes for the linking,
I finally get the errors as

Linking...

Link /OUT:"Release/fds5_serial.exe" /INCREMENTAL:NO /NOLOGO /
MANIFEST /
MANIFESTFILE:"D:\Test_Fortran\FDS5_solution\fds5_serial\release
\fds5_serial.exe.intermediate.manifest" /SUBSYSTEM:CONSOLE "Release/
prec.obj" "Release/smvv.obj" "Release/pois.obj" "Release/devc.obj"
"Release/cons.obj" "Release/type.obj" "Release/mesh.obj" "Release/
func.obj" "Release/pres.obj" "Release/wall.obj" "Release/fire.obj"
"Release/ieva.obj" "Release/irad.obj" "Release/mass.obj" "Release/
part.obj" "Release/radi.obj" "Release/ctrl.obj" "Release/divg.obj"
"Release/velo.obj" "Release/evac.obj" "Release/init.obj" "Release/
read.obj" "Release/dump.obj" "Release/main.obj" "d:\Test_Fortran
\FDS5_solution\release\isob.lib"

Fatal compilation error: Out of memory asking for 2097160.
Link: error: problem during multi-file optimization compilation (code
1)
Link: error: problem during multi-file optimization compilation (code
1)

When I reset the runtime libraries to "Debug Multithreaded", the
situation did not change at all.

[Liangzhu Wang]

For DEBUG, the " MSVCR80D.dll" error can be solved by using
"Multi-threaded Debug (/MTd)" in the runtime library.

For RELEASE, the runtime library should be "Multi-threaded (/MT)".

I am using VS2005+IntelFortran 10.0.3718.2005.

Leon

[Kevin]

I added a wiki to the FDS Developer Site that explains how we compile
under MS Visual Studio 2005 and Intel Fortran 10.1

I'd be interested if you have any comments about it.

[may...@gmail.com]

Thanks Kevin,
Luckily, I have passed the compilation successfully.
But when I tried running the fds5.exe, it came to a breakpoint at
(WRITE(COMPILE_DATE,'(A)') REVISION_DATE(INDEX(REVISION_DATE,'(')
+1:INDEX(REVISION_DATE,')')-1))
Since I am new to Fortran, how to solve this one.

[Kevin]

Can you explain exactly what you did, and also post the input file.

> > > situation did not change at all.- Hide quoted text -
>
> - Show quoted text -

[Liangzhu Wang]

Kevin, the wiki looks great! Good job!

Leon

[may...@gmail.com]

This is what i have done..

C:\nist>fds5.exe room_fire.fds
forrtl: severe (66): output statement overflows record, unit -5, file
Internal F
ormatted Write
Image PC Routine Line Source
fds5.exe 00C903A2 Unknown Unknown Unknown
fds5.exe 00C8D630 Unknown Unknown Unknown
fds5.exe 00BCA0BA Unknown Unknown Unknown
fds5.exe 00BC9CD7 Unknown Unknown Unknown
fds5.exe 00BCBA52 Unknown Unknown Unknown
fds5.exe 00BCB0B0 Unknown Unknown Unknown
fds5.exe 00B9B489 _MAIN__ 55 main.f90
fds5.exe 00C94838 Unknown Unknown Unknown
fds5.exe 00C355A3 Unknown Unknown Unknown
fds5.exe 00C3536D Unknown Unknown Unknown
kernel32.dll 7C816FD7 Unknown Unknown Unknown

[Kevin]

I cannot reproduce this error.

Try adding the line below just before line 55 in main.f90

write(0,*) 'REVISION_DATE=',REVISION_DATE

and post the result. Line 55 in main.f90 is just putting together a
character string for print out.

Also, try rebuilding FDS cleanly, using "Rebuild All."

Is this an XP or Vista OS?

> > > - Show quoted text -- Hide quoted text -

[may...@gmail.com]

When I delete line 55, the program runs successfully.
BTW, my os is XP.

[Kevin]

OK, but could you please print out the information I asked for in Post
7. I would like to know if there is something wrong with the text
string 'REVISION_DATE'. It would be a bother for you to have to
comment out this line each time you recompile.

[may...@gmail.com]

Sorry, I forget to print the result for you. Here it is.

C:\nist>fds5 room_fire.fds
REVISION_DATE=
$Date: 2007-11-13 05:08:24 +0800 (鏄熸湡浜? 13 鍗佷竴鏈?2007) $

forrtl: severe (66): output statement overflows record, unit -5, file
Internal F
ormatted Write
Image PC Routine Line Source

fds5.exe 00C90422 Unknown Unknown Unknown
fds5.exe 00C8D6B0 Unknown Unknown Unknown
fds5.exe 00BCA13A Unknown Unknown Unknown
fds5.exe 00BC9D57 Unknown Unknown Unknown
fds5.exe 00BCBAD2 Unknown Unknown Unknown
fds5.exe 00BCB130 Unknown Unknown Unknown
fds5.exe 00B9B511 _MAIN__ 56 main.f90
fds5.exe 00C948B8 Unknown Unknown Unknown
fds5.exe 00C35623 Unknown Unknown Unknown
fds5.exe 00C353ED Unknown Unknown Unknown

kernel32.dll 7C816FD7 Unknown Unknown Unknown

[Zatarra]

may...@gmail.com I answer to your mail here. I couldn't get fds 5.0.2
RELEASE compiled on Studio 2005 and fortran compiler 9.1. only the
DEBUG version I got it work.

I followed the new instruction on the developer site (under wiki) but
I get a similar error to yours:


Creating library...
1>Build log was saved at "file://d:\FDS_5_Compiling\FDS5_Serial\isob
\Release\BuildLog.htm"
1>isob - 0 error(s), 18 warning(s)
2>------ Rebuild All started: Project: Serial, Configuration: Release
Win32 ------
2>Deleting intermediate files and output files for project 'Serial',
configuration 'Release|Win32'.
2>Compiling with Intel Fortran 9.1 C:\Programme\Intel\Compiler\Fortran
\9.1\IA32\...
2>..\..\fds_5_Source\prec.f90
2>..\..\fds_5_Source\smvv.f90
2>..\..\fds_5_Source\pois.f90
2>..\..\fds_5_Source\devc.f90
2>..\..\fds_5_Source\cons.f90
2>..\..\fds_5_Source\type.f90
2>..\..\fds_5_Source\mesh.f90
2>..\..\fds_5_Source\func.f90
2>..\..\fds_5_Source\pres.f90
2>..\..\fds_5_Source\wall.f90
2>..\..\fds_5_Source\fire.f90
2>..\..\fds_5_Source\ieva.f90
2>..\..\fds_5_Source\irad.f90
2>..\..\fds_5_Source\mass.f90
2>..\..\fds_5_Source\part.f90
2>..\..\fds_5_Source\radi.f90
2>..\..\fds_5_Source\ctrl.f90
2>..\..\fds_5_Source\divg.f90
2>..\..\fds_5_Source\velo.f90
2>..\..\fds_5_Source\evac.f90
2>..\..\fds_5_Source\init.f90
2>..\..\fds_5_Source\read.f90
2>..\..\fds_5_Source\dump.f90
2>..\..\fds_5_Source\main.f90
2>Linking...
2>Link: error: problem during multi-file optimization compilation
(code 1)
2>Link: error: problem during multi-file optimization compilation
(code 1)
2>
2>Build log written to "file://D:\FDS_5_Compiling\FDS5_Serial\Serial
\Release\BuildLog.htm"


I've tried with both MT/MTd and also MD/MDd but it didn't work...
What was your solution?

[may...@gmail.com]

Me too.

On Nov 22, 2:57 pm, Zatarra <martino.pore...@ticino.com> wrote:
> mayl...@gmail.com I answer to your mail here. I couldn't get fds 5.0.2

[Zatarra]

I've tried again and also if the compiler doesn't say that, it's
clearly a memory problem (the compiling fails when the memory for the
fromcom.exe exceeds 2 GB. (Win 32-bit is limited to 2 GB pro task)

How have you solved this problem by the release compiling?

Thank Martino

> > What was your solution?- Hide quoted text -

[Hostikka Simo]

I remember having similar problem at some point.

I can't reproduce it now (it may be compiler version dependent) but
could you check what are the following property settings:

Isob/Configuration properties/General/Whole Program Optimization (I
have "Use link time code generation")

Isob/Configuration properties/C/C++ / Optimization/Whole Program
optimization
(I have "Enable link-time code generation (/GL)"

Simo

> > > 2>Release

> > > Win32 ------
> > > 2>Deleting intermediate files and output files for
> project 'Serial',
> > > configuration 'Release|Win32'.
> > > 2>Compiling with Intel Fortran 9.1

> > > 2>C:\Programme\Intel\Compiler\Fortran

> > > 2>"file://D:\FDS_5_Compiling\FDS5_Serial\Serial

[Zatarra]

Thanks Simo for the quick reply.
I've checked and I've the same settings as you...

When you are compiling ( or better linking) how mach ram uses the
compiler at the peek?
Have you changed in the windows settings something? ( ex. boot.ini )

> > > - Show quoted text -- Hide quoted text -

[Zatarra]

On the Frotran / Optimisations : which options do you have?

I've

optimisation: maximise speed
global optimisations: yes
Favor fast code
Pemtium 4
no processor extention
no parallelisation

[Hostikka Simo]

All are defaults. See the pic.

Simo

> > > > > > 2>compilation

> > > > > > (code 1)
> > > > > > 2>Link: error: problem during multi-file optimization

> > > > > > 2>compilation

[Zatarra]

I've exactly the same settings...

How much ram does the fortcom.exe use during compiling and linking?

[Hostikka Simo]

Compilation: peak to 320 MB
Linking: too fast to read, maybe 10 MB?

I think the slow linking and huge memory consumption is either a
compiler
bug or wrong setting somewhere. Have you tried "clean solution" before
compiling?

Simo

> -----Original Message-----
> From: fds...@googlegroups.com
> [mailto:fds...@googlegroups.com] On Behalf Of Zatarra
> Sent: Thursday, November 22, 2007 15:15
> To: FDS and Smokeview Discussions
> Subject: [fds-smv post:1826] Re: Problems of compiling FDS5.0
> with VS2005+Intel Fortran Compiler
>
>

[Zatarra]

Ok fortran Compiler 10.1.011 Installed and now it works! I've built my
release for I32 fds5 but, now Trying to run it I get this error from
FDS5 (testing room_fire.fds) with the offical release of fds5 it runs
perfectly..


D:\FDS_5_Compiling\test>fds5_d room_fire.fds
forrtl: severe (64): input conversion error, unit -5, file Internal
Formatted Re
ad

Image PC Routine Line Source

fds5_d.exe 0068F1FA Unknown Unknown Unknown
fds5_d.exe 0068CAA9 Unknown Unknown Unknown
fds5_d.exe 0062EF02 Unknown Unknown Unknown
fds5_d.exe 0062EA39 Unknown Unknown Unknown
fds5_d.exe 006003B9 Unknown Unknown Unknown
fds5_d.exe 005FEF19 Unknown Unknown Unknown
fds5_d.exe 005F06BA Unknown Unknown Unknown
fds5_d.exe 00693249 Unknown Unknown Unknown
fds5_d.exe 0067585F Unknown Unknown Unknown

[Zatarra]

Better with the Debug version...

D:\FDS_5_Compiling\test>fds5_d room_fire.fds
forrtl: severe (64): input conversion error, unit -5, file Internal
Formatted Re
ad
Image PC Routine Line Source

fds5_d.exe 00CAA962 Unknown Unknown Unknown
fds5_d.exe 00CA72C9 Unknown Unknown Unknown
fds5_d.exe 00BC9972 Unknown Unknown Unknown
fds5_d.exe 00BC94A9 Unknown Unknown Unknown
fds5_d.exe 00BCE4A5 Unknown Unknown Unknown
fds5_d.exe 00BCD005 Unknown Unknown Unknown
fds5_d.exe 00B9A56F _MAIN__ 51 main.f90
fds5_d.exe 00CAF329 Unknown Unknown Unknown
fds5_d.exe 00C4B0E3 Unknown Unknown Unknown
fds5_d.exe 00C4AEAD Unknown Unknown Unknown

kernel32.dll 7C816FD7 Unknown Unknown Unknown

[Zatarra]

More or less the same for the paralle version:


C:\Programme\MPICH2\bin>mpiexec -file \\PC2\FDS_5_Compiling\test
\Parallel\config
.txt
Mesh 1 of 1 is alive on PC2

forrtl: severe (64): input conversion error, unit -5, file Internal
Formatted Re
ad
Image PC Routine Line Source

parallel.exe 00CD4122 Unknown Unknown Unknown
parallel.exe 00CD0A89 Unknown Unknown Unknown
parallel.exe 00BF3132 Unknown Unknown Unknown
parallel.exe 00BF2C69 Unknown Unknown Unknown
parallel.exe 00BF7C65 Unknown Unknown Unknown
parallel.exe 00BF67C5 Unknown Unknown Unknown
parallel.exe 00B9A82E _MAIN__ 76
main_mpi.f90
parallel.exe 00CD8AE9 Unknown Unknown Unknown
parallel.exe 00C748A3 Unknown Unknown Unknown
parallel.exe 00C7466D Unknown Unknown Unknown

kernel32.dll 7C816FD7 Unknown Unknown Unknown

job aborted:
rank: node: exit code[: error message]
0: PC2: 64: process 0 exited without calling finalize


Any suggestions?

[Zatarra]

Ok I've deleted all this stuff and now it works! ( lines form 74 to
83)



! Assign a compilation date (All Nodes)
!WRITE(REVISION_DATE,'(A)') mainmpirev(INDEX(mainmpirev,':')
+1:LEN_TRIM(mainmpirev)-2)
!READ (REVISION_DATE,'(I5)') REVISION_NUMBER
!WRITE(REVISION_DATE,'(A)') mainmpidate
!CALL GET_REVISION_NUMBER(REVISION_NUMBER,REVISION_DATE)
!SVN_REVISION_NUMBER = REVISION_NUMBER
!WRITE(COMPILE_DATE,'(A)') REVISION_DATE(INDEX(REVISION_DATE,'(')
+1:INDEX(REVISION_DATE,')')-1)
!WRITE(VERSION_STRING,'(A)') '5.0.2'
!VERSION_NUMBER = 5.0 ! Just use to indicate the major version
!PARALLEL = .TRUE.

[Kevin]

This is the problem -- when we commit source code to the SVN
repository, a character string in the file gets "sticky tagged" with
the SVN number, date, author, and other information. When you update
your local copy of the source code, our "sticky tag" gets replaced
with equivalent Chinese characters. My REVISION_DATE is

$Date: 2007-11-20 16:02:28 -0500 (Tue, 20 Nov 2007) $

FDS parses this string, and uses it to print out the version number
and compile date. We did not realize that SVN would use native
characters, but then again, we never told SVN to use English. This is
a bug that needs to be fixed.

[Kevin]

Your problem seems to be the same as the one I address in Post 25.

[dr_jfloyd]

Keyword expansion is a client side and not a server side operation.
The $Date$ keyword is expanded as the file is updated from the
repository by the machine downloading the update. It appears SVN is
using the language of the machine doing the update. This doesn't seem
to be something we can control. You can try turning off keyword
expansion before downloading an update. If you are using a third
party program (like SmartSVN) the help for that program should explain
how to do that.
You could also try setting your machine's language to English prior to
doing an update.

this appears as a known issue on the svn developers page
http://subversion.tigris.org/issues/show_bug.cgi?id=2332

For our purposes parsing the revision data automatically is useful.
If you are compiling on your own then perhaps this is not so important
to you. We could add an input to &MISC to skip over the revision
parsing and only set the version string and number.

> > > > > > > > > > situation did not change at all.- Hide quoted text -- Hide quoted text -
>
> - Show quoted text -...
>
> read more >>

[Kevin]

I have also added an ERR= branching to the internal write statement in
main.f90 that is causing the problem. Now the code should just skip
trying to parse the REVISION_DATE if there is some problem with it.
The result would be an "Undetermined" date. Hopefully the version and
SVN number will come out.

mayleaf and Zatarra -- could you both try to update the latest source
code, compile, and try to run again. You should not have to change
your language setting. Hopefully, FDS will ignore non-standard
characters in the date string.

On Nov 23, 11:35 pm, dr_jfloyd <drjfl...@gmail.com> wrote:
> Keyword expansion is a client side and not a server side operation.
> The $Date$ keyword is expanded as the file is updated from the
> repository by the machine downloading the update. It appears SVN is
> using the language of the machine doing the update. This doesn't seem
> to be something we can control. You can try turning off keyword
> expansion before downloading an update. If you are using a third
> party program (like SmartSVN) the help for that program should explain
> how to do that.
> You could also try setting your machine's language to English prior to
> doing an update.
>

> this appears as a known issue on the svn developers pagehttp://subversion.tigris.org/issues/show_bug.cgi?id=2332

> > > > > > > > > > > Link /OUT:"Release/fds5_serial.exe" /INCREMENTAL:NO /NOLOGO /- Hide quoted text -

[Zatarra]

Hi Kevin,
no the error is always here also with the newest source code.
I've only compiled a parallel release version but the error is always
an "Internal Formatting.. bla bla"

> > > > > > > > > > situation did not change at all.- Hide quoted text -- Zitierten Text ausblenden -
>
> - Zitierten Text anzeigen -...
>
> Erfahren Sie mehr >>

[Zatarra]

... on big models I'm getting now this error on my compiled version of
fds5 (parallel modus)

as usual I've to delete the line 75 to 84 on the main.f

http://groups.google.com/group/fds-smv/web/fds5_parallel_error.JPG

what could be, memory allocation problem?

> > > > > > > > > > > read.obj" "Release/dump.obj" "Release/main.obj"- Zitierten Text ausblenden -

[Zatarra]

ok problem solved was the Stack Reserve and Commit that had to be set
to 32768000

> > > > > > > > > > > > \fds5_serial.exe.intermediate.manifest"- Zitierten Text ausblenden -

[Kevin]

Could you post exactly what you see on the screen, especially the line
number and the exact error statement. Plus, double check the compile
date and SVN number. We have not officially released the executables
for this version, which remains at 5.0.2, but the SVN number is
different.

> > > > > > > > > > > read.obj" "Release/dump.obj" "Release/main.obj"- Hide quoted text -
>
> - Show quoted text -...
>
> read more >>

[Zatarra]

Hi Kevin,

here ist the error reporting from the debug-compiled fds5 (Revision
1036: /trunk/FDS/trunk/FDS_5)


E:\Program Files\MPICH2\bin>mpiexec -file \\PC11\FDS_64\config.txt
Mesh 3 of 4 is alive on pc11
Mesh 4 of 4 is alive on pc11
Mesh 1 of 4 is alive on pc11
Mesh 2 of 4 is alive on pc11

forrtl: severe (64): input conversion error, unit -5, file Internal
Formatted Re
ad

Image PC Routine Line
Source

fds5_d.exe 0000000140A5F434 Unknown Unknown
Unknown
fds5_d.exe 0000000140A5B009 Unknown Unknown
Unknown
fds5_d.exe 0000000140990812 Unknown Unknown
Unknown
fds5_d.exe 0000000140946819 Unknown Unknown
Unknown
fds5_d.exe 0000000140945F72 Unknown Unknown
Unknown
fds5_d.exe 000000014094D269 Unknown Unknown
Unknown
fds5_d.exe 000000014094B81D Unknown Unknown
Unknown
fds5_d.exe 00000001408DCE21 MAIN__ 76
main_mpi.f90

fds5_d.exe 0000000140A64F7C Unknown Unknown
Unknown
fds5_d.exe 00000001409E9A6C Unknown Unknown
Unknown
fds5_d.exe 00000001409E983E Unknown Unknown
Unknown
kernel32.dll 0000000077D5964C Unknown Unknown
Unknown

forrtl: severe (64): input conversion error, unit -5, file Internal
Formatted Re
ad

Image PC Routine Line
Source

fds5_d.exe 0000000140A5F434 Unknown Unknown
Unknown
fds5_d.exe 0000000140A5B009 Unknown Unknown
Unknown
fds5_d.exe 0000000140990812 Unknown Unknown
Unknown
fds5_d.exe 0000000140946819 Unknown Unknown
Unknown
fds5_d.exe 0000000140945F72 Unknown Unknown
Unknown
fds5_d.exe 000000014094D269 Unknown Unknown
Unknown
fds5_d.exe 000000014094B81D Unknown Unknown
Unknown
fds5_d.exe 00000001408DCE21 MAIN__ 76
main_mpi.f90

fds5_d.exe 0000000140A64F7C Unknown Unknown
Unknown
fds5_d.exe 00000001409E9A6C Unknown Unknown
Unknown
fds5_d.exe 00000001409E983E Unknown Unknown
Unknown
kernel32.dll 0000000077D5964C Unknown Unknown
Unknown

forrtl: severe (64): input conversion error, unit -5, file Internal
Formatted Re
ad

Image PC Routine Line
Source

fds5_d.exe 0000000140A5F434 Unknown Unknown
Unknown
fds5_d.exe 0000000140A5B009 Unknown Unknown
Unknown
fds5_d.exe 0000000140990812 Unknown Unknown
Unknown
fds5_d.exe 0000000140946819 Unknown Unknown
Unknown
fds5_d.exe 0000000140945F72 Unknown Unknown
Unknown
fds5_d.exe 000000014094D269 Unknown Unknown
Unknown
fds5_d.exe 000000014094B81D Unknown Unknown
Unknown
fds5_d.exe 00000001408DCE21 MAIN__ 76
main_mpi.f90

fds5_d.exe 0000000140A64F7C Unknown Unknown
Unknown
fds5_d.exe 00000001409E9A6C Unknown Unknown
Unknown
fds5_d.exe 00000001409E983E Unknown Unknown
Unknown
kernel32.dll 0000000077D5964C Unknown Unknown
Unknown

forrtl: severe (64): input conversion error, unit -5, file Internal
Formatted Re
ad

Image PC Routine Line
Source

fds5_d.exe 0000000140A5F434 Unknown Unknown
Unknown
fds5_d.exe 0000000140A5B009 Unknown Unknown
Unknown
fds5_d.exe 0000000140990812 Unknown Unknown
Unknown
fds5_d.exe 0000000140946819 Unknown Unknown
Unknown
fds5_d.exe 0000000140945F72 Unknown Unknown
Unknown
fds5_d.exe 000000014094D269 Unknown Unknown
Unknown
fds5_d.exe 000000014094B81D Unknown Unknown
Unknown
fds5_d.exe 00000001408DCE21 MAIN__ 76
main_mpi.f90

fds5_d.exe 0000000140A64F7C Unknown Unknown
Unknown
fds5_d.exe 00000001409E9A6C Unknown Unknown
Unknown
fds5_d.exe 00000001409E983E Unknown Unknown
Unknown
kernel32.dll 0000000077D5964C Unknown Unknown

Unknown

job aborted:
rank: node: exit code[: error message]

0: PC11: 64: process 0 exited without calling finalize
1: PC11: 64: process 1 exited without calling finalize
2: PC11: 64: process 2 exited without calling finalize
3: PC11: 64: process 3 exited without calling finalize

> > > > > > > > > > > > MANIFESTFILE:"D:\Test_Fortran\FDS5_solution\fds5_serial\release- Zitierten Text ausblenden -

[Kevin]

OK, there are still problems with the non-English character strings.
Could you post the lines from the top of main_mpi.f90 that look like
these:

! Miscellaneous declarations

CHARACTER(255), PARAMETER :: mainmpiid='$Id: main_mpi.f90 1023
2007-11-23 21:52:38Z mcgratta $'
CHARACTER(255), PARAMETER :: mainmpirev='$Revision: 1023 $'
CHARACTER(255), PARAMETER :: mainmpidate='$Date: 2007-11-23 16:52:38
-0500 (Fri, 23 Nov 2007) $'

When I update my source files from the repository, SVN puts
information between the $ signs. This tells me who made the last
change to the file, and when. My text is written in English, because
that is the native language of my computer. I'd like to see what your
text looks like. We might have to check the character strings first
before we try to parse them. The calculation fails when FDS tries to
get the revision number and date from the above text strings.

> > > > > > > > > fds5.exe 00C3536D Unknown Unknown Unknown- Hide quoted text -

[Zatarra]

Here my main_mpi.f90

PROGRAM FDS

! Fire Dynamics Simulator, Main Program, Multiple CPU version.

USE PRECISION_PARAMETERS
USE MESH_VARIABLES
USE GLOBAL_CONSTANTS
USE TRAN
USE DUMP
USE READ_INPUT
USE INIT
USE DIVG
USE PRES
USE MASS
USE PART
USE VELO
USE RAD
USE MEMORY_FUNCTIONS
USE COMP_FUNCTIONS, ONLY : SECOND, WALL_CLOCK_TIME, SHUTDOWN
USE MATH_FUNCTIONS, ONLY : GAUSSJ
USE DEVICE_VARIABLES
USE WALL_ROUTINES
USE FIRE
USE RADCONS
USE CONTROL_FUNCTIONS
!EVAC:USE EVAC

IMPLICIT NONE

! Miscellaneous declarations

CHARACTER(255), PARAMETER :: mainmpiid='$Id$'
CHARACTER(255), PARAMETER :: mainmpirev='$Revision$'
CHARACTER(255), PARAMETER :: mainmpidate='$Date$'
LOGICAL :: EX,DIAGNOSTICS,EXCHANGE_RADIATION=.TRUE.
INTEGER :: LO10,NM,IZERO,DATE_TIME(8),NN,REVISION_NUMBER,IOS
CHARACTER(10) :: BIG_BEN(3)
CHARACTER(255) :: REVISION_DATE
REAL(EB) :: T_MAX,T_MIN
REAL(EB), ALLOCATABLE, DIMENSION(:) :: T,TC_GLB,TC_LOC,DT_SYNC,
DTNEXT_SYNC,DSUM_ALL,PSUM_ALL,USUM_ALL
INTEGER, ALLOCATABLE, DIMENSION(:) ::
MESH_STOP_STATUS,COUNT_GLB,COUNT_LOC
LOGICAL, ALLOCATABLE, DIMENSION(:) :: ACTIVE_MESH,STATE_GLB,STATE_LOC
INTEGER NOM,IWW,IW
INTEGER, PARAMETER :: N_DROP_ADOPT_MAX=10000
TYPE (MESH_TYPE), POINTER :: M,M4
TYPE (OMESH_TYPE), POINTER :: M2,M3

! MPI stuff

! INCLUDE '/usr/local/include/mpif.h' ! Uncomment this line only if
the compiler needs help finding mpif.h
INCLUDE 'mpif.h'
INTEGER :: N,MYID=0,NUMPROCS=1,I,IERR,STATUS(MPI_STATUS_SIZE)
INTEGER :: RNODE,BUFFER_SIZE,TAG,PNAMELEN
INTEGER, ALLOCATABLE, DIMENSION(:,:,:) :: TAGS
INTEGER, ALLOCATABLE, DIMENSION(:) :: REQ,PREQ
INTEGER, ALLOCATABLE, DIMENSION(:,:) :: ARRAY_OF_STATUSES
INTEGER, ALLOCATABLE, DIMENSION(:,:) :: ARRAY_OF_STATUSES2
INTEGER :: N_REQ,N_PREQ
CHARACTER(MPI_MAX_PROCESSOR_NAME) PNAME

! Initialize MPI (First executable lines of code)

CALL MPI_INIT(IERR)
CALL MPI_COMM_RANK(MPI_COMM_WORLD, MYID, IERR)
CALL MPI_COMM_SIZE(MPI_COMM_WORLD, NUMPROCS, IERR)
CALL MPI_GET_PROCESSOR_NAME(PNAME, PNAMELEN, IERR)

WRITE(LU_ERR,'(A,I2,A,I2,A,A)') 'Mesh ',MYID+1,' of ', NUMPROCS,' is
alive on ',PNAME(1:PNAMELEN)

! Start wall clock timing

WALL_CLOCK_START = WALL_CLOCK_TIME()

! Assign a compilation date (All Nodes)
!WRITE(REVISION_DATE,'(A)') mainmpirev(INDEX(mainmpirev,':')
+1:LEN_TRIM(mainmpirev)-2)
!READ (REVISION_DATE,'(I5)') REVISION_NUMBER
!WRITE(REVISION_DATE,'(A)') mainmpidate
!CALL GET_REVISION_NUMBER(REVISION_NUMBER,REVISION_DATE)
!SVN_REVISION_NUMBER = REVISION_NUMBER

!WRITE(COMPILE_DATE,'(A)',IOSTAT=IOS,ERR=10)
REVISION_DATE(INDEX(REVISION_DATE,'(')+1:INDEX(REVISION_DATE,')')-1)
!10 IF (IOS>0) COMPILE_DATE = 'Undetermined'

!WRITE(VERSION_STRING,'(A)') '5.0.2'
!VERSION_NUMBER = 5.0 ! Just use to indicate the major version
!PARALLEL = .TRUE.

! Read input from CHID.data file (All Nodes)

> > > > > > > > > Also, try- Zitierten Text ausblenden -

[Kevin]

Aha! Now I see. FDS is trying to get information from

> CHARACTER(255), PARAMETER :: mainmpiid='$Id$'
> CHARACTER(255), PARAMETER :: mainmpirev='$Revision$'
> CHARACTER(255), PARAMETER :: mainmpidate='$Date$'

but because you have not downloaded via an SVN update, there is no
information in the $ $ strings. This is a bug. We need to make sure
that FDS ignores these strings if the information is not present. For
the moment, simply comment out the problem lines.

> > > > > no the error is always here also with the newest source code.- Hide quoted text -