Compilation errors with gcc/gfortran 4.9.1

Asked by Susanne Westhoff

Hello,

I am using the gcc compiler version 4.9.1. When generating events for the process p p > t t~ j in the fNLO mode, I get the following compilation error message:

---------------------------------------------------------------------------------------------------

WARNING: A compilation Error occurs when trying to compile /home/westhoff/programs/MG5_aMC_v2_2_3/bin/ttj-nlo/SubProcesses/P0_gg_ttxg.
The compilation fails with the following output message:
    /opt/rh/devtoolset-3/root/usr/bin/gfortran -O -fno-automatic -ffixed-line-length-132 -c -I. fks_singular.f
    fks_singular.f:6016.3:

     12 continue
       1
    fks_singular.f:6031.132:

          goto 12
                                                                                                                                        2
    Warning: Legacy Extension: Label at (1) is not in the same block as the GOTO statement at (2)
    fks_singular.f:4137.23:

          common/ccheckcnt/i_momcmp_count,xratmax
                           1
    Warning: Padding of 4 bytes required before 'xratmax' in COMMON 'ccheckcnt' at (1); reorder elements or use -fno-align-commons
    fks_singular.f:2889.20:

          common/cemsca/emsca,emsca_bare,emscasharp,scalemin,scalemax
                        1
    Warning: Padding of 4 bytes required before 'scalemin' in COMMON 'cemsca' at (1); reorder elements or use -fno-align-commons
    run.inc:75.21:
        Included at fks_singular.f:3902:

          common/to_rwgt/ do_rwgt_scale, rw_Fscale_down, rw_Fscale_up, rw_Rscale_down, rw_Rscale_up,
                         1
    Warning: Padding of 4 bytes required before 'rw_fscale_down' in COMMON 'to_rwgt' at (1); reorder elements or use -fno-align-commons
    run.inc:75.21:
        Included at fks_singular.f:2940:

          common/to_rwgt/ do_rwgt_scale, rw_Fscale_down, rw_Fscale_up, rw_Rscale_down, rw_Rscale_up,
                         1
    Warning: Padding of 4 bytes required before 'rw_fscale_down' in COMMON 'to_rwgt' at (1); reorder elements or use -fno-align-commons
    run.inc:75.21:
        Included at fks_singular.f:3801:

          common/to_rwgt/ do_rwgt_scale, rw_Fscale_down, rw_Fscale_up, rw_Rscale_down, rw_Rscale_up,
                         1
    Warning: Padding of 4 bytes required before 'rw_fscale_down' in COMMON 'to_rwgt' at (1); reorder elements or use -fno-align-commons
    run.inc:75.21:
        Included at fks_singular.f:4698:

          common/to_rwgt/ do_rwgt_scale, rw_Fscale_down, rw_Fscale_up, rw_Rscale_down, rw_Rscale_up,
                         1
    Warning: Padding of 4 bytes required before 'rw_fscale_down' in COMMON 'to_rwgt' at (1); reorder elements or use -fno-align-commons
    run.inc:75.21:
        Included at fks_singular.f:3960:

          common/to_rwgt/ do_rwgt_scale, rw_Fscale_down, rw_Fscale_up, rw_Rscale_down, rw_Rscale_up,
                         1
    Warning: Padding of 4 bytes required before 'rw_fscale_down' in COMMON 'to_rwgt' at (1); reorder elements or use -fno-align-commons
    run.inc:75.21:
        Included at fks_singular.f:747:

          common/to_rwgt/ do_rwgt_scale, rw_Fscale_down, rw_Fscale_up, rw_Rscale_down, rw_Rscale_up,
                         1
    Warning: Padding of 4 bytes required before 'rw_fscale_down' in COMMON 'to_rwgt' at (1); reorder elements or use -fno-align-commons
    run.inc:75.21:
        Included at fks_singular.f:38:

          common/to_rwgt/ do_rwgt_scale, rw_Fscale_down, rw_Fscale_up, rw_Rscale_down, rw_Rscale_up,
                         1
    Warning: Padding of 4 bytes required before 'rw_fscale_down' in COMMON 'to_rwgt' at (1); reorder elements or use -fno-align-commons
    run.inc:75.21:
        Included at fks_singular.f:298:

          common/to_rwgt/ do_rwgt_scale, rw_Fscale_down, rw_Fscale_up, rw_Rscale_down, rw_Rscale_up,
                         1
    Warning: Padding of 4 bytes required before 'rw_fscale_down' in COMMON 'to_rwgt' at (1); reorder elements or use -fno-align-commons
    run.inc:75.21:
        Included at fks_singular.f:396:

          common/to_rwgt/ do_rwgt_scale, rw_Fscale_down, rw_Fscale_up, rw_Rscale_down, rw_Rscale_up,
                         1
    Warning: Padding of 4 bytes required before 'rw_fscale_down' in COMMON 'to_rwgt' at (1); reorder elements or use -fno-align-commons
    run.inc:75.21:
        Included at fks_singular.f:526:

          common/to_rwgt/ do_rwgt_scale, rw_Fscale_down, rw_Fscale_up, rw_Rscale_down, rw_Rscale_up,
                         1
    Warning: Padding of 4 bytes required before 'rw_fscale_down' in COMMON 'to_rwgt' at (1); reorder elements or use -fno-align-commons
    run.inc:75.21:
        Included at fks_singular.f:836:

          common/to_rwgt/ do_rwgt_scale, rw_Fscale_down, rw_Fscale_up, rw_Rscale_down, rw_Rscale_up,
                         1
    Warning: Padding of 4 bytes required before 'rw_fscale_down' in COMMON 'to_rwgt' at (1); reorder elements or use -fno-align-commons
    run.inc:75.21:
        Included at fks_singular.f:887:

          common/to_rwgt/ do_rwgt_scale, rw_Fscale_down, rw_Fscale_up, rw_Rscale_down, rw_Rscale_up,
                         1
    Warning: Padding of 4 bytes required before 'rw_fscale_down' in COMMON 'to_rwgt' at (1); reorder elements or use -fno-align-commons
    run.inc:75.21:
        Included at fks_singular.f:953:

          common/to_rwgt/ do_rwgt_scale, rw_Fscale_down, rw_Fscale_up, rw_Rscale_down, rw_Rscale_up,
                         1
    Warning: Padding of 4 bytes required before 'rw_fscale_down' in COMMON 'to_rwgt' at (1); reorder elements or use -fno-align-commons
    fks_singular.f:1610.20:

          common/cemsca/emsca,emsca_bare,emscasharp,scalemin,scalemax
                        1
    Warning: Padding of 4 bytes required before 'scalemin' in COMMON 'cemsca' at (1); reorder elements or use -fno-align-commons
    run.inc:75.21:
        Included at fks_singular.f:1467:

          common/to_rwgt/ do_rwgt_scale, rw_Fscale_down, rw_Fscale_up, rw_Rscale_down, rw_Rscale_up,
                         1
    Warning: Padding of 4 bytes required before 'rw_fscale_down' in COMMON 'to_rwgt' at (1); reorder elements or use -fno-align-commons
    fks_singular.f:4048.23:

          common/ccheckcnt/i_momcmp_count,xratmax
                           1
    Warning: Padding of 4 bytes required before 'xratmax' in COMMON 'ccheckcnt' at (1); reorder elements or use -fno-align-commons
    run.inc:75.21:
        Included at fks_singular.f:5754:

          common/to_rwgt/ do_rwgt_scale, rw_Fscale_down, rw_Fscale_up, rw_Rscale_down, rw_Rscale_up,
                         1
    Warning: Padding of 4 bytes required before 'rw_fscale_down' in COMMON 'to_rwgt' at (1); reorder elements or use -fno-align-commons
    /opt/rh/devtoolset-3/root/usr/bin/gfortran -o madevent_mintFO analysis_td_template.o dbook.o open_output_files_dummy.o parton_lum_1.o parton_lum_2.o parton_lum_3.o parton_lum_4.o parton_lum_5.o parton_lum_6.o parton_lum_chooser.o matrix_1.o matrix_2.o matrix_3.o matrix_4.o matrix_5.o matrix_6.o real_me_chooser.o fks_inc_chooser.o leshouche_inc_chooser.o genps_fks.o setcuts.o setscales.o born.o sborn_sf.o b_sf_001.o b_sf_002.o b_sf_003.o b_sf_004.o b_sf_005.o b_sf_006.o b_sf_007.o b_sf_008.o b_sf_009.o b_sf_010.o b_sf_011.o b_sf_012.o fks_Sij.o trapfpe.o fastjetfortran_madfks_core.o fjcore.o fastjet_wrapper.o fks_singular.o montecarlocounter.o reweight_xsec.o boostwdir2.o configs_and_props_inc_chooser.o initcluster.o cluster.o reweight.o get_color.o FKSParamReader.o iproc_map.o MC_integer.o appl_interface_dummy.o cuts.o pythia_unlops.o recluster.o fill_MC_mshell.o born_hel.o open_output_files.o add_write_info.o BinothLHA.o madfks_plot.o driver_mintFO.o mint-integrator2.o handling_lhe_events.o -L../../lib/ libMadLoop.a -lcts -liregi -L../../lib/ -ldhelas -lgeneric -lmodel -lpdf -lcernlib -lLHAPDF -lstdc++ -lstdc++
    ../../lib//libLHAPDF.a(wrapct12.o): In function `lhct12set_':
    wrapct12.f:(.text+0x3): relocation truncated to fit: R_X86_64_PC32 against symbol `ct12co_' defined in COMMON section in ../../lib//libLHAPDF.a(wrapct12.o)
    ../../lib//libLHAPDF.a(wrapct12.o): In function `ctlhct12alphas_':
    wrapct12.f:(.text+0x79): relocation truncated to fit: R_X86_64_PC32 against symbol `ct12par2_' defined in COMMON section in ../../lib//libLHAPDF.a(wrapct12.o)
    ../../lib//libLHAPDF.a(wrapct12.o): In function `ctlhct12pdf_':
    wrapct12.f:(.text+0x21a): relocation truncated to fit: R_X86_64_PC32 against symbol `ct12par2_' defined in COMMON section in ../../lib//libLHAPDF.a(wrapct12.o)
    ../../lib//libLHAPDF.a(wrapct12.o): In function `ctlhpartonx12_':
    wrapct12.f:(.text+0x448): relocation truncated to fit: R_X86_64_PC32 against symbol `ct12par2_' defined in COMMON section in ../../lib//libLHAPDF.a(wrapct12.o)
    wrapct12.f:(.text+0x44f): relocation truncated to fit: R_X86_64_PC32 against symbol `ct12co_' defined in COMMON section in ../../lib//libLHAPDF.a(wrapct12.o)
    ../../lib//libLHAPDF.a(wrapct12.o): In function `master.0.ct12evolve_':
    wrapct12.f:(.text+0x1a0f): relocation truncated to fit: R_X86_64_PC32 against symbol `ct12par2_' defined in COMMON section in ../../lib//libLHAPDF.a(wrapct12.o)
    wrapct12.f:(.text+0x220a): relocation truncated to fit: R_X86_64_PC32 against symbol `ct12par2_' defined in COMMON section in ../../lib//libLHAPDF.a(wrapct12.o)
    ../../lib//libLHAPDF.a(EVLNNPDF.o): In function `lh_initpar2evln_':
    EVLNNPDF.f:(.text+0x63): relocation truncated to fit: R_X86_64_PC32 against symbol `nnpdf10ppar2evln_' defined in COMMON section in ../../lib//libLHAPDF.a(EVLNNPDF.o)
    ../../lib//libLHAPDF.a(EVLNNPDF.o): In function `lh_initevln2lha_':
    EVLNNPDF.f:(.text+0x1a4): relocation truncated to fit: R_X86_64_PC32 against symbol `nnpdf10eevln2lha_' defined in COMMON section in ../../lib//libLHAPDF.a(EVLNNPDF.o)
    EVLNNPDF.f:(.text+0x425): relocation truncated to fit: R_X86_64_PC32 against symbol `nnpdf10eevln2lha_' defined in COMMON section in ../../lib//libLHAPDF.a(EVLNNPDF.o)
    ../../lib//libLHAPDF.a(EVLNNPDF.o): In function `lh_fbeta_':
    EVLNNPDF.f:(.text+0xd16): additional relocation overflows omitted from the output
    collect2: error: ld returned 1 exit status
    make: *** [madevent_mintFO] Error 1

---------------------------------------------------------------------------------------------------

Is there any known problem with my compiler version? I recently upgraded it from gfortran 4.4.7., since MG requests 4.6 or later.

Thank you very much for your help!

Question information

Language:
English Edit question
Status:
Solved
For:
MadGraph5_aMC@NLO Edit question
Assignee:
marco zaro Edit question
Solved by:
Susanne Westhoff
Solved:
Last query:
Last reply:
Revision history for this message
marco zaro (marco-zaro) said :
#1

Dear Suzanne,
most likely your problem is related to lhapdf, in particular if you are using lhapdf6 on a mac, and is related to these bugs

https://bugs.launchpad.net/mg5amcnlo/+bug/1421642
https://bugs.launchpad.net/mg5amcnlo/+bug/1368721
https://bugs.launchpad.net/mg5amcnlo/+bug/1424752

Unfortunately, the only solution i have found is pretty radical, and consists in using only gnu (so gcc/g++) compilers, removing the native clang, because clang has compatibility problems with gnu compilers and lhapdf6 has many dependencies that are already compiled with clang.

So if it is lhapdf6+macosx, I fear I cannot be of much help.
Note that NNPDF2.3 pdfs are shipped with MG5_AMC@NLO, and therefore you do not need lhapdf to use them, and you can alternatively use lhapdf 5, if the pdf you want to use is supported.

Let me know if you need further help.
Cheers,

Marco

Revision history for this message
Susanne Westhoff (suw22) said :
#2

Hi Marco,

Thank you much for your fast reply!

The problem is indeed connected to lhapdf. When I use the MG built-in
PDFs, it works fine. Is there a way to use CTEQ10 PDFs without relying
on lhapdf? Can I import them manually into MG?

Thanks again,
Susanne.

On 06/25/2015 01:01 PM, marco zaro wrote:
> Your question #268548 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/268548
>
> Status: Open => Answered
>
> marco zaro proposed the following answer:
> Dear Suzanne,
> most likely your problem is related to lhapdf, in particular if you are using lhapdf6 on a mac, and is related to these bugs
>
> https://bugs.launchpad.net/mg5amcnlo/+bug/1421642
> https://bugs.launchpad.net/mg5amcnlo/+bug/1368721
> https://bugs.launchpad.net/mg5amcnlo/+bug/1424752
>
> Unfortunately, the only solution i have found is pretty radical, and
> consists in using only gnu (so gcc/g++) compilers, removing the native
> clang, because clang has compatibility problems with gnu compilers and
> lhapdf6 has many dependencies that are already compiled with clang.
>
> So if it is lhapdf6+macosx, I fear I cannot be of much help.
> Note that NNPDF2.3 pdfs are shipped with MG5_AMC@NLO, and therefore you do not need lhapdf to use them, and you can alternatively use lhapdf 5, if the pdf you want to use is supported.
>
> Let me know if you need further help.
> Cheers,
>
> Marco
>

--
Dr. Susanne Westhoff
University of Pittsburgh
Department of Physics and Astronomy, PITT PACC
408 Allen Hall
Pittsburgh, PA 15260, USA
phone: +1 (412) 648 1024
email: <email address hidden>

Revision history for this message
Susanne Westhoff (suw22) said :
#3

Hi again,

Unfortunately the PDFs were not the only problem. The process failed
when setting up the grids. The console error is the following:

  ----------------------------------------------------------------------

WARNING: program
/home/westhoff/programs/MG5_aMC_v2_2_3/bin/ttj-nlo/SubProcesses/P0_gg_ttxg/ajob8
0 all 0 launch ends with non zero status: 5. Stop all computation
WARNING: Last 15 lines of logfile
/home/westhoff/programs/MG5_aMC_v2_2_3/bin/ttj-nlo/SubProcesses/P0_gg_ttxg/*/log.txt:
| http://www.ugr.es/~pittau/CutTools |
| |
| Compiler with 34 significant digits detetected |
  ----------------------------------------------------------------------

########################################################################
# #
# You are using OneLOop-3.4 #
# #
# for the evaluation of 1-loop scalar 1-, 2-, 3- and 4-point functions #
# #
# author: Andreas van Hameren <email address hidden> #
# date: 02-01-2014 #
# #

----------------------------------------------------------------------

This points me to the log.txt file, where I read:

------------------------------------------------------------------------------------------------------

------------------------------------------------------------------------
| You are using CutTools - Version 1.9.3 |
| Authors: G. Ossola, C. Papadopoulos, R. Pittau |
| Published in JHEP 0803:042,2008 |
| http://www.ugr.es/~pittau/CutTools |
| |
| Compiler with 34 significant digits detetected |
  ----------------------------------------------------------------------

########################################################################
# #
# You are using OneLOop-3.4 #
# #
# for the evaluation of 1-loop scalar 1-, 2-, 3- and 4-point functions #
# #
# author: Andreas van Hameren <email address hidden> #
# date: 02-01-2014 #
# #
# Please cite #
# A. van Hameren, #
# Comput.Phys.Commun. 182 (2011) 2427-2438, arXiv:1007.4716 #
# A. van Hameren, C.G. Papadopoulos and R. Pittau, #
# JHEP 0909:106,2009, arXiv:0903.4665 #
# in publications with results obtained with the help of this program. #
# #
########################################################################
  #####################################################################################
  # #
  # IREGI-alpha-1.0.0 #
  # package for one-loop tensor Integral REduction with General
propagator Indices #
  # Author: Hua-Sheng Shao
(<email address hidden><email address hidden>) #
  # a) Physics School, Peking University, Beijing,
China #
  # b) PH Department, TH Unit, CERN, Geneva,
Switzerland #
  # #
  #####################################################################################
  alpha_s value used for the virtuals is (for the first PS point):
0.10843852263722766
  ---- POLES CANCELLED ----
ABS integral = NaN +/- NaN ( NaN %)
Integral = NaN +/- NaN ( NaN %)
Virtual = NaN +/- NaN ( NaN %)
Virtual ratio = NaN +/- NaN ( NaN %)
ABS virtual = NaN +/- NaN ( NaN %)
Born*ao2pi = 0.2632E+00 +/- 0.5748E-01 ( 21.842 %)
Chi^2 per d.o.f. 0.0000E+00
accumulated results ABS integral = NaN +/- NaN ( NaN %)
accumulated results Integral = NaN +/- NaN ( NaN %)
accumulated results Virtual = NaN +/- NaN ( NaN %)
accumulated results Virtual ratio = NaN +/- NaN ( NaN %)
accumulated results ABS virtual = NaN +/- NaN ( NaN %)
accumulated results Born*ao2pi = 0.2632E+00 +/- 0.5748E-01 ( 21.842 %)
accumulated result Chi^2 per DoF = 0.0000E+00
update virtual fraction to: 1.000 NaN
   1: 0 1 2 3
4 5678 9 0

Program received signal 11 (SIGSEGV): Segmentation fault.

Backtrace for this error:
#0 0x0000003775eac584 in wait () from /lib64/libc.so.6
#1 0x00007f26db24a00d in ?? () from /usr/lib64/libgfortran.so.3
#2 0x00007f26db24b82e in ?? () from /usr/lib64/libgfortran.so.3
#3 0x00007f26db24a6ca in ?? () from /usr/lib64/libgfortran.so.3
#4 <signal handler called>
#5 0x00000000004a4e32 in ?? ()
#6 0x00000000004a649d in ?? ()
#7 0x00000000004a33d6 in ?? ()
#8 0x00000000004a4a7e in ?? ()
#9 0x0000003775e1ed5d in __libc_start_main () from /lib64/libc.so.6
#10 0x0000000000404129 in ?? ()
#11 0x00007fff818356d8 in ?? ()
#12 0x000000000000001c in ?? ()
#13 0x0000000000000001 in ?? ()
#14 0x00007fff8183738d in ?? ()
#15 0x0000000000000000 in ?? ()
  ------- iteration 2
  Update # PS points (even): 1600 --> 1600
Time in seconds: 881

----------------------------------------------------------------------------------------------------------------------

It seems the reduction of loop integrals fails for the process g g > t
tx g. Do you know why? It doesn't look like a compiler problem this time.

Thank you very much again!
Susanne.

-------------------------------------------------------------------------------------------------------------

On 06/25/2015 02:04 PM, Susanne Westhoff wrote:
> Hi Marco,
>
> Thank you much for your fast reply!
>
> The problem is indeed connected to lhapdf. When I use the MG built-in
> PDFs, it works fine. Is there a way to use CTEQ10 PDFs without relying
> on lhapdf? Can I import them manually into MG?
>
> Thanks again,
> Susanne.
>
>
>
> On 06/25/2015 01:01 PM, marco zaro wrote:
>> Your question #268548 on MadGraph5_aMC@NLO changed:
>> https://answers.launchpad.net/mg5amcnlo/+question/268548
>>
>> Status: Open => Answered
>>
>> marco zaro proposed the following answer:
>> Dear Suzanne,
>> most likely your problem is related to lhapdf, in particular if you
>> are using lhapdf6 on a mac, and is related to these bugs
>>
>> https://bugs.launchpad.net/mg5amcnlo/+bug/1421642
>> https://bugs.launchpad.net/mg5amcnlo/+bug/1368721
>> https://bugs.launchpad.net/mg5amcnlo/+bug/1424752
>>
>> Unfortunately, the only solution i have found is pretty radical, and
>> consists in using only gnu (so gcc/g++) compilers, removing the native
>> clang, because clang has compatibility problems with gnu compilers and
>> lhapdf6 has many dependencies that are already compiled with clang.
>>
>> So if it is lhapdf6+macosx, I fear I cannot be of much help.
>> Note that NNPDF2.3 pdfs are shipped with MG5_AMC@NLO, and therefore
>> you do not need lhapdf to use them, and you can alternatively use
>> lhapdf 5, if the pdf you want to use is supported.
>>
>> Let me know if you need further help.
>> Cheers,
>>
>> Marco
>>
>
>

--
Dr. Susanne Westhoff
University of Pittsburgh
Department of Physics and Astronomy, PITT PACC
408 Allen Hall
Pittsburgh, PA 15260, USA
phone: +1 (412) 648 1024
email: <email address hidden>

Revision history for this message
marco zaro (marco-zaro) said :
#4

Dear Suzanne,
in principle lhapdf5 should work with ct10 (and on a mac as well).
Concerning our second problem, i have added Valentin Hirschi to this thread.
Did you install gfortran 9 after having installed amcatnlo or viceversa?
If so, can you try to recompile cutttools/iregi by going inside
vendor/cuttools/
and type
make clean; make
similarly into vendor/iregi/src/
and retrying to generate and run the process?
What version of MG5_aMC@NLO are you using?

Cheers,

Marco

Revision history for this message
Susanne Westhoff (suw22) said :
#5

Dear Marco,

Thank you much. Initially, I installed fortran 4.9.1 after having installed amcatnlo. Meanwhile, I have re-installed amcatnlo. fortran 4.9.1 is not used as the default compiler on my system (Scientific Linux 6.6 with GNOME 2.28.2). But I have set the path to fortran 4.9.1 in the file mg5_configuration.txt in ~/input. I saw that in the makefiles in sub-directories of amcatnlo (for instance, in vendors/cuttools), the variable FC is set to this path, too.

I followed your suggestion and re-compiled cuttools and iregi, then re-generated and ran the process. Unfortunately, the problem is the same as before. Do you have any other advice?

I am using MG version 2.2.3.

Best,
Susanne.

Revision history for this message
marco zaro (marco-zaro) said :
#6

Hi Suzanne,
this is with or without LHAPDF?
is lhapdf compiled with the same gfortran/gcc 4.9.1?
note that, in the case you use devtoolset on SL6 and LHAPDF6, you need to recompile LHAPDF using boost libraries compiled with the same version of the compiler
(see also this thread, if it can help: https://answers.launchpad.net/mg5amcnlo/+question/258558)
Let me know,
Cheers,Î

Marco

On 07 Jul 2015, at 17:26, Susanne Westhoff <email address hidden> wrote:

> Question #268548 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/268548
>
> Status: Needs information => Open
>
> Susanne Westhoff gave more information on the question:
> Dear Marco,
>
> Thank you much. Initially, I installed fortran 4.9.1 after having
> installed amcatnlo. Meanwhile, I have re-installed amcatnlo. fortran
> 4.9.1 is not used as the default compiler on my system (Scientific Linux
> 6.6 with GNOME 2.28.2). But I have set the path to fortran 4.9.1 in the
> file mg5_configuration.txt in ~/input. I saw that in the makefiles in
> sub-directories of amcatnlo (for instance, in vendors/cuttools), the
> variable FC is set to this path, too.
>
> I followed your suggestion and re-compiled cuttools and iregi, then re-
> generated and ran the process. Unfortunately, the problem is the same as
> before. Do you have any other advice?
>
> I am using MG version 2.2.3.
>
> Best,
> Susanne.
>
> --
> You received this question notification because you are a direct
> subscriber of the question.

Revision history for this message
Susanne Westhoff (suw22) said :
#7

Hi Marco,

Thank you for your fast reply!

I am not using LHAPDF now, but the built-in nn23nlo set, trying to separate the LHAPDF problem from this one.

Best,
Susanne.

Revision history for this message
marco zaro (marco-zaro) said :
#8

ok, can you copy-paste the error you get?
Thanks,

Marco
On 07 Jul 2015, at 17:56, Susanne Westhoff <email address hidden> wrote:

> Question #268548 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/268548
>
> Status: Answered => Open
>
> Susanne Westhoff is still having a problem:
> Hi Marco,
>
> Thank you for your fast reply!
>
> I am not using LHAPDF now, but the built-in nn23nlo set, trying to
> separate the LHAPDF problem from this one.
>
> Best,
> Susanne.
>
> --
> You received this question notification because you are a direct
> subscriber of the question.

Revision history for this message
Susanne Westhoff (suw22) said :
#9

Hello,

Together with Valentin, we solved the compiler problem. It is connected to 'write' statements of quadruple precision floating point variables. To still run MadGraph with devtoolset-3 gfortran v. 4.9.1, I modified the file

MG_aMC_v2_3_0/vendor/CutTools/src/avh/avh_olo.f90.

In particular, the statement

write(*,*) x

needs to be replaced by

double precision xp
xp = x
write (*,*) xp

in the following functions:

- function printi
- function printr
- function printi (occurs a second time below function printr)

Many thanks to Marco and Valentin for their help and for their fast and good support!

Revision history for this message
Andre (andre-david-8) said :
#10

Hi Susanne,

Can you post a diff of the changes you made? Because on a fresh copy of MG_aMC_v2_3_0, I do not see two copies of printi, nor do I see write(*,*) x in MG_aMC_v2_3_0/vendor/CutTools/src/avh/avh_olo.f90.

They rather look like:

  function printr( xx_in ,ndec_in ) result(rslt)
  real(kindr2) &
                  ,intent(in) :: xx_in
  integer,optional,intent(in) :: ndec_in
  character(ndecim(prcpar)+nxtr+novh ) :: rslt
  character(ndecim(prcpar)+nxtr+novh+1) :: cc
  character(10) :: aa,bb
  integer :: ndec
  real(kindr2) :: xx
  xx = xx_in
  if (present(ndec_in)) then ;ndec=ndec_in
                        else ;ndec=ndecim(prcpar)+nxtr
  endif
  write(aa,'(i10)') min(len(cc),ndec+novh+1) ;aa=adjustl(aa)
  write(bb,'(i10)') min(len(cc),ndec ) ;bb=adjustl(bb)
  aa = '(e'//trim(aa)//'.'//trim(bb)//')'
  write(cc,aa) xx ;cc=adjustl(cc)
  if (cc(1:2).eq.'-0') then ;rslt = '-'//cc(3:len(cc))
  else ;rslt = ' '//cc(2:len(cc))
  endif
  end function

  function printi( ii ) result(rslt)
  integer ,intent(in) :: ii
  character(ndecim(prcpar)) :: rslt
  character(ndecim(prcpar)) :: cc
  character(10) :: aa
  write(aa,'(i10)') ndecim(prcpar) ;aa=adjustl(aa)
  aa = '(i'//trim(aa)//')'
  write(cc,aa) ii ;cc=adjustl(cc)
  if (cc(1:1).ne.'-') then ;rslt=' '//cc
  else ;rslt=cc
  endif
  end function

Thanks,

Andre

Revision history for this message
marco zaro (marco-zaro) said :
#11

Hi Andre,
it should be line 548
  real(kindr2) :: xx
  xx = xx_in

to be changed in

  double precision :: xx
  xx = xx_in

Let us know if it works.
Cheers,

Marco

Revision history for this message
Andre (andre-david-8) said :
#12

Hi,

It helped, but moved the problem elsewhere:

$ make check_poles
[...]
make[1]: Leaving directory `/afs/cern.ch/work/a/adavidzh/private/workspace/mg5_230_devtoolset3/ppzhQCD/SubProcesses/P0_uux_zh/V0_uux_zh'
gfortran -o check_poles parton_lum_1.o parton_lum_2.o parton_lum_3.o parton_lum_chooser.o matrix_1.o matrix_2.o matrix_3.o real_me_chooser.o fks_inc_chooser.o leshouche_inc_chooser.o genps_fks.o setcuts.o setscales.o born.o sborn_sf.o b_sf_001.o fks_Sij.o trapfpe.o fastjetfortran_madfks_core.o fjcore.o fastjet_wrapper.o fks_singular.o veto_xsec.o montecarlocounter.o reweight_xsec.o boostwdir2.o configs_and_props_inc_chooser.o initcluster.o cluster.o reweight.o get_color.o FKSParamReader.o iproc_map.o MC_integer.o reweight_xsec_events_pdf_dummy.o appl_interface_dummy.o cuts.o pythia_unlops.o recluster.o check_poles.o BinothLHA.o born_hel.o -L../../lib/ libMadLoop.a -lcts -liregi -L../../lib/ -ldhelas -lgeneric -lmodel -lpdf -lcernlib -lstdc++
../../lib//libcts.a(avh_olo.o): In function `__avh_olo_qp_print_MOD_printr':
avh_olo.f90:(.text+0x95b33): undefined reference to `_gfortran_transfer_real128_write'
collect2: error: ld returned 1 exit status
make: *** [check_poles] Error 1

I suspect (but cannot confirm) that this is due to the

xx = xx_in

Ideas?

Revision history for this message
marco zaro (marco-zaro) said :
#13

Hi Andre,
did you recompile cuttools?
It loooked to me that the problem is the same as before
Otherwise try xx = dble(xx_in)

Cheers,

Marco

Revision history for this message
Andre (andre-david-8) said :
#14

Hi Marco,

Yes, I recompiled.
Your suggestion compiles, but same problem:

make[1]: Leaving directory `/afs/cern.ch/work/a/adavidzh/private/workspace/mg5_230_devtoolset3/ppzhQCD/SubProcesses/P0_uux_zh/V0_uux_zh'
gfortran -o check_poles parton_lum_1.o parton_lum_2.o parton_lum_3.o parton_lum_chooser.o matrix_1.o matrix_2.o matrix_3.o real_me_chooser.o fks_inc_chooser.o leshouche_inc_chooser.o genps_fks.o setcuts.o setscales.o born.o sborn_sf.o b_sf_001.o fks_Sij.o trapfpe.o fastjetfortran_madfks_core.o fjcore.o fastjet_wrapper.o fks_singular.o veto_xsec.o montecarlocounter.o reweight_xsec.o boostwdir2.o configs_and_props_inc_chooser.o initcluster.o cluster.o reweight.o get_color.o FKSParamReader.o iproc_map.o MC_integer.o reweight_xsec_events_pdf_dummy.o appl_interface_dummy.o cuts.o pythia_unlops.o recluster.o check_poles.o BinothLHA.o born_hel.o -L../../lib/ libMadLoop.a -lcts -liregi -L../../lib/ -ldhelas -lgeneric -lmodel -lpdf -lcernlib -lstdc++
../../lib//libcts.a(avh_olo.o): In function `__avh_olo_qp_print_MOD_printr':
avh_olo.f90:(.text+0x95b33): undefined reference to `_gfortran_transfer_real128_write'
collect2: error: ld returned 1 exit status
make: *** [check_poles] Error 1

Revision history for this message
marco zaro (marco-zaro) said :
#15

Hi Andre`,
I have notified Valentin Hirschi, who knows for sure how to solve your problem (he did so many times...)

Cheers,

Marco

Revision history for this message
Andre (andre-david-8) said :
#16

My bad: there are three printr functions (in a dp, mp, and qp module) and only the qp one matters.

Things works now. For reference the following diff works for:
- lxplus SLC6 6.6
- scl enable devtoolset-3 bash [gcc version 4.9.1 20140922 (Red Hat 4.9.1-10) (GCC)]
- MG5_aMC_v2.3.0.tar.gz
- p p > z h [QCD]

--- MG5_aMC_v2_3_0/vendor/CutTools/src/avh/avh_olo.f90 2015-07-02 04:18:10.000000001 +0200
+++ mg5_230_devtoolset3/vendor/CutTools/src/avh/avh_olo.f90 2015-07-31 18:49:58.000000001 +0200
@@ -12242,7 +12242,7 @@
   character(ndecim(prcpar)+nxtr+novh+1) :: cc
   character(10) :: aa,bb
   integer :: ndec
- real(kindr2) :: xx
+ double precision :: xx
   xx = xx_in
   if (present(ndec_in)) then ;ndec=ndec_in
                         else ;ndec=ndecim(prcpar)+nxtr

Revision history for this message
Andre (andre-david-8) said :
#17
Revision history for this message
marco zaro (marco-zaro) said :
#18

Ok great!

Cheers,

Marco
On 31 Jul 2015, at 18:56, Andre <email address hidden> wrote:

> Question #268548 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/268548
>
> Andre posted a new comment:
> My bad: there are three printr functions (in a dp, mp, and qp module)
> and only the qp one matters.
>
> Things works now. For reference the following diff works for:
> - lxplus SLC6 6.6
> - scl enable devtoolset-3 bash [gcc version 4.9.1 20140922 (Red Hat 4.9.1-10) (GCC)]
> - MG5_aMC_v2.3.0.tar.gz
> - p p > z h [QCD]
>
> --- MG5_aMC_v2_3_0/vendor/CutTools/src/avh/avh_olo.f90 2015-07-02 04:18:10.000000001 +0200
> +++ mg5_230_devtoolset3/vendor/CutTools/src/avh/avh_olo.f90 2015-07-31 18:49:58.000000001 +0200
> @@ -12242,7 +12242,7 @@
> character(ndecim(prcpar)+nxtr+novh+1) :: cc
> character(10) :: aa,bb
> integer :: ndec
> - real(kindr2) :: xx
> + double precision :: xx
> xx = xx_in
> if (present(ndec_in)) then ;ndec=ndec_in
> else ;ndec=ndecim(prcpar)+nxtr
>
> --
> You received this question notification because you are a direct
> subscriber of the question.