MadGraph5_aMC@NLO

Can't Compile vendor/CutTools for MG5_aMC_v2_5_2

Asked by Chris

Working with MG5_aMC_v2_5_2 with gfortran version 4.7.2, I can not compile CutTools.

When I call "make" I get the following error: (Have confirmed mpmodule.mod is located in includects)

cp -p ./src/cts/cts_qpr.in ./src/cts/cts_mpr.h
mkdir -p includects
cp -p ./src/avh/* includects/
cp -p ./src/cts/cts_qpc.in ./src/cts/cts_mpc.h
cp -p ./src/cts/cts_qprec.in ./src/cts/cts_mprec.h
cp -p ./src/cts/cts_qpinit.in ./src/cts/cts_mpinit.h
cp -p ./src/cts/* includects/
cp -p ./src/mpfun90/* includects/
cp -p ./src/qcdloop/* includects/
cp -p ./src/makefile includects/
cd includects && make EXE="" FC="gfortran" FFLAGS="-fPIC -fno-automatic -O2 -funroll-all-loops" default
make[1]: Entering directory `/home1s/cranelli/MonoZ/MadGraph_StandAlone/MG5_aMC_v2_5_2/vendor/CutTools/includects'
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c mpfun90.f90
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c mpmod90.f90
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c mpmodm90.f90
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c mpmodx90.f90
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c avh_olo.f90
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c cts_type.f90
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c cts_tensors.f90
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c cts_constants.f90
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c cts_combinatorics.f90
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c cts_dynamics.f90
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c cts_kinematics.f90
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c cts_loopfunctions.f90
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c cts_cuttools.f90
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c cts_cutroutines.f90
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c -o aaxbx.o aaxbx.f
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c -o aaxcx.o aaxcx.f
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c -o aaxdx.o aaxdx.f
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c -o aaxex.o aaxex.f
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c -o aaxinv.o aaxinv.f
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c -o aacbc.o aacbc.f
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c -o aaccc.o aaccc.f
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c -o aacinv.o aacinv.f
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c -o spence.o spence.f
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c -o npoin.o npoin.f
mpmod90.f90:43.4:

use mpfunmod
    1
Fatal Error: Can't open module file 'mpfunmod.mod' for reading at (1): No such file or directory
make[1]: *** [libcts.a(mpmod90.o)] Error 1
make[1]: *** Waiting for unfinished jobs....
mpmodm90.f90:34.4:

use mpmodule
    1
mpmodx90.f90:32.4:
Fatal Error: Can't open module file 'mpmodule.mod' for reading at (1): No such file or directory

use mpmodule
    1
Fatal Error: Can't open module file 'mpmodule.mod' for reading at (1): No such file or directory
make[1]: *** [libcts.a(mpmodx90.o)] Error 1
make[1]: *** [libcts.a(mpmodm90.o)] Error 1
cts_mprec.h:1.6:
    Included at cts_dynamics.f90:3:

  use mpmodule
      1
Fatal Error: Can't open module file 'mpmodule.mod' for reading at (1): No such file or directory
make[1]: *** [libcts.a(cts_dynamics.o)] Error 1
cts_mprec.h:1.6:
    Included at cts_cuttools.f90:2:

  use mpmodule
      1
Fatal Error: Can't open module file 'mpmodule.mod' for reading at (1): No such file or directory
make[1]: *** [libcts.a(cts_cuttools.o)] Error 1
aaxex.f:218.22:

      call ffeji(cdxi,mcxi,ccxi,mcxi,cbxi,mbxi,caxi,maxi,
                      1
Warning: Actual argument contains too few elements for dummy argument 'mdxi' (30/55) at (1)
cts_mprec.h:1.6:
    Included at cts_cutroutines.f90:51:

  use mpmodule
      1
Fatal Error: Can't open module file 'mpmodule.mod' for reading at (1): No such file or directory
make[1]: *** [libcts.a(cts_cutroutines.o)] Error 1
cts_mprec.h:1.6:
    Included at cts_tensors.f90:2:

  use mpmodule
      1
Fatal Error: Can't open module file 'mpmodule.mod' for reading at (1): No such file or directory
make[1]: *** [libcts.a(cts_tensors.o)] Error 1
cts_mprec.h:1.6:
    Included at cts_type.f90:57:

  use mpmodule
      1
Fatal Error: Can't open module file 'mpmodule.mod' for reading at (1): No such file or directory
make[1]: *** [libcts.a(cts_type.o)] Error 1
cts_mprec.h:1.6:
    Included at cts_constants.f90:5:

  use mpmodule
      1
Fatal Error: Can't open module file 'mpmodule.mod' for reading at (1): No such file or directory
make[1]: *** [libcts.a(cts_constants.o)] Error 1
cts_loopfunctions.f90:7.6:

  use denominators
      1
Fatal Error: Can't open module file 'denominators.mod' for reading at (1): No such file or directory
make[1]: *** [libcts.a(cts_loopfunctions.o)] Error 1
cts_mprec.h:1.6:
    Included at cts_kinematics.f90:19:

  use mpmodule
      1
Fatal Error: Can't open module file 'mpmodule.mod' for reading at (1): No such file or directory
make[1]: *** [libcts.a(cts_kinematics.o)] Error 1
ar rv libcts.a cts_combinatorics.o
ar: creating libcts.a
a - cts_combinatorics.o
avh_olo.f90:23477.6:

  use mpmodule
      1
Fatal Error: Can't open module file 'mpmodule.mod' for reading at (1): No such file or directory
make[1]: *** [libcts.a(avh_olo.o)] Error 1
ar rv libcts.a mpfun90.o
a - mpfun90.o
rm aacbc.o aaxcx.o npoin.o aaxdx.o aacinv.o aaxbx.o aaxex.o aaccc.o aaxinv.o spence.o
make[1]: Leaving directory `/home1s/cranelli/MonoZ/MadGraph_StandAlone/MG5_aMC_v2_5_2/vendor/CutTools/includects'
make: *** [default] Error 2

Question information

Language:
English Edit question
Status:
Answered
For:
MadGraph5_aMC@NLO Edit question
Assignee:
Valentin Hirschi Edit question
Last query:
Last reply:
Revision history for this message
Valentin Hirschi (valentin-hirschi) said : 		#1

Given that this code compiles for most users and platform, I can only think of one thing that can go wrong (I assumed you already made sure to clean up the build directory first with 'make clean').

CutTools doesn't not support parallel builds because it didn't correctly order its dependencies, so my guess is that your default make configuration is such that it will attempt to use as many cores as available when called.
Could you then try to compile it with 'make -j1' so as to forbid parallel building and check if it solves your problem?

If so then we might think of automatically adding '-j1' to our automated build of CutTools, so as to solve your problem for people with similar setups.

Revision history for this message
Chris (christopher-anelli) said : 		#2

Hi Valentin,

Thanks for your help.

Yes, I tried make clean.

Running make -j1 fixes the issue, CutTools compiles.

Thanks!

Chris

On Tue, Mar 7, 2017 at 12:32 PM, Valentin Hirschi <
<email address hidden>> wrote:

> Your question #521122 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/521122
>
> Status: Open => Needs information
>
> Valentin Hirschi requested more information:
> Given that this code compiles for most users and platform, I can only
> think of one thing that can go wrong (I assumed you already made sure to
> clean up the build directory first with 'make clean').
>
> CutTools doesn't not support parallel builds because it didn't correctly
> order its dependencies, so my guess is that your default make configuration
> is such that it will attempt to use as many cores as available when called.
> Could you then try to compile it with 'make -j1' so as to forbid parallel
> building and check if it solves your problem?
>
> If so then we might think of automatically adding '-j1' to our automated
> build of CutTools, so as to solve your problem for people with similar
> setups.
>
> --
> To answer this request for more information, you can either reply to
> this email or enter your reply at the following page:
> https://answers.launchpad.net/mg5amcnlo/+question/521122
>
> You received this question notification because you asked the question.
>

Revision history for this message
Valentin Hirschi (valentin-hirschi) said : 		#3

Ok, I have modified the dev. version of MG5_aMC so that it now enforces sequential build of CutTools.
This should be fine since nowadays almost all versions of make support the '-j' option.

Can you help with this problem?

Provide an answer of your own, or ask Chris for more information if necessary.

To post a message you must log in.