CutTools not compiling for NLO processes in fresh checkout of 2.6.3.2
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
MadGraph5_aMC@NLO |
Fix Released
|
Undecided
|
Valentin Hirschi |
Bug Description
Hi Everyone,
I downloaded a fresh version of MG5_aMC 2.6.3.2 in a fresh terminal (Mac Sierra 10.12.2) and installed a number of additional packages using the interactive prompt:
> install pythia8 (also prompts a lot of installing)
> install ninja
> install collier
and then try to generate a new process and output it:
> generate p p > t t~ z [QCD]
> output NLO_TTZ
at this point it tells me that it is compiling cuttools and crashes (see below).
I've seen this same bug in 2.6.2 and 2.6.1, and it occurs if I go into vendors and try to recompile it by hand. I also checked on an lxplus setup and encounter the same problems.
The version I have of gfortran on my mac is:
GNU Fortran (Homebrew GCC 8.1.0) 8.1.0
and on lxplus is:
GNU Fortran (GCC) 4.9.3
Any help you could provide with fixing these compilation errors would be greatly appreciated!
Cheers,
Jay
<compile log>
INFO: Compiling CutTools. This has to be done only once and can take a couple of minutes.
Error detected in "output NLO_TTZ"
write debug file MG5_debug
If you need help with this issue please contact us on https:/
MadGraph5Error : A compilation Error occurs when trying to compile /Users/
The compilation fails with the following output message:
cp -p ./src/cts/
cp -p ./src/cts/
cp -p ./src/cts/
cp -p ./src/cts/
mkdir -p includects
cp -p ./src/avh/* includects/
cp -p ./src/cts/* includects/
cp -p ./src/mpfun90/* includects/
cp -p ./src/qcdloop/* includects/
cp -p ./src/makefile includects/
cd includects && /Applications/
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c mpfun90.f90
ar rv libcts.a mpfun90.o
ar: creating archive libcts.a
a - mpfun90.o
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c mpmod90.f90
ar rv libcts.a mpmod90.o
a - mpmod90.o
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c mpmodm90.f90
ar rv libcts.a mpmodm90.o
a - mpmodm90.o
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c mpmodx90.f90
ar rv libcts.a mpmodx90.o
a - mpmodx90.o
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c avh_olo.f90
ar rv libcts.a avh_olo.o
a - avh_olo.o
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c cts_type.f90
ar rv libcts.a cts_type.o
a - cts_type.o
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c cts_tensors.f90
ar rv libcts.a cts_tensors.o
a - cts_tensors.o
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c cts_constants.f90
ar rv libcts.a cts_constants.o
a - cts_constants.o
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c cts_combinatori
ar rv libcts.a cts_combinatorics.o
a - cts_combinatorics.o
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c cts_dynamics.f90
ar rv libcts.a cts_dynamics.o
a - cts_dynamics.o
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c cts_kinematics.f90
ar rv libcts.a cts_kinematics.o
a - cts_kinematics.o
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c cts_loopfunctio
ar rv libcts.a cts_loopfunctions.o
a - cts_loopfunctions.o
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c cts_cuttools.f90
ar rv libcts.a cts_cuttools.o
a - cts_cuttools.o
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c cts_cutroutines.f90
ar rv libcts.a cts_cutroutines.o
a - cts_cutroutines.o
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c -o aaxbx.o aaxbx.f
ar rv libcts.a aaxbx.o
a - aaxbx.o
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c -o aaxcx.o aaxcx.f
ar rv libcts.a aaxcx.o
a - aaxcx.o
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c -o aaxdx.o aaxdx.f
ar rv libcts.a aaxdx.o
a - aaxdx.o
gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c -o aaxex.o aaxex.f
aaxex.
call ffeji(cdxi,
Error: Actual argument contains too few elements for dummy argument 'mdxi' (30/55) at (1)
aaxex.
do i=16,35
cexi(i) = ce3ijk(
Warning: Array reference at (1) out of bounds (35 > 20) in loop beginning at (2)
aaxex.
do i=16,35
cexi(i) = ce3ijk(
Warning: Array reference at (1) out of bounds (35 > 20) in loop beginning at (2)
aaxex.
do i=16,35
cexi(i) = ce3ijk(
Warning: Array reference at (1) out of bounds (35 > 20) in loop beginning at (2)
make[1]: *** [aaxex.o] Error 1
rm aaxcx.o aaxdx.o aaxbx.o
make: *** [default] Error 2
Changed in mg5amcnlo: | |
status: | New → Fix Released |
Hi,
The problem that you report is linked to GCC8.
Sounds weird that you face the exact same error on gcc4.9.3 (but if you setup gcc via a environment variable to turn warning into crash).
Not sure what we can do concerning Cuttools and the compatibility with gcc8.
I will assign Valentin to comment on this.
Cheers,
Olivier
> On 25 Jun 2018, at 11:59, James Howarth <email address hidden> wrote: /answers. launchpad. net/mg5amcnlo jhowarth/ Documents/ Work/MadGraph/ MG5_aMC_ v2_6_3_ 2/vendor/ CutTools. cts_qpr. in ./src/cts/cts_mpr.h cts_qpc. in ./src/cts/cts_mpc.h cts_qprec. in ./src/cts/ cts_mprec. h cts_qpinit. in ./src/cts/ cts_mpinit. h Xcode.app/ Contents/ Developer/ usr/bin/ make EXE="" FC="gfortran" FFLAGS="-fPIC -fno-automatic -O2 -funroll-all-loops" default
>
> Public bug reported:
>
> Hi Everyone,
>
> I downloaded a fresh version of MG5_aMC 2.6.3.2 in a fresh terminal (Mac
> Sierra 10.12.2) and installed a number of additional packages using the
> interactive prompt:
>
>> install pythia8 (also prompts a lot of installing)
>> install ninja
>> install collier
>
> and then try to generate a new process and output it:
>
>> generate p p > t t~ z [QCD]
>> output NLO_TTZ
>
> at this point it tells me that it is compiling cuttools and crashes (see
> below).
>
> I've seen this same bug in 2.6.2 and 2.6.1, and it occurs if I go into
> vendors and try to recompile it by hand. I also checked on an lxplus
> setup and encounter the same problems.
>
> The version I have of gfortran on my mac is:
>
> GNU Fortran (Homebrew GCC 8.1.0) 8.1.0
>
> and on lxplus is:
>
> GNU Fortran (GCC) 4.9.3
>
> Any help you could provide with fixing these compilation errors would be
> greatly appreciated!
>
> Cheers,
>
> Jay
>
> <compile log>
>
> INFO: Compiling CutTools. This has to be done only once and can take a couple of minutes.
> Error detected in "output NLO_TTZ"
> write debug file MG5_debug
> If you need help with this issue please contact us on https:/
> MadGraph5Error : A compilation Error occurs when trying to compile /Users/
> The compilation fails with the following output message:
> cp -p ./src/cts/
> cp -p ./src/cts/
> cp -p ./src/cts/
> cp -p ./src/cts/
> mkdir -p includects
> cp -p ./src/avh/* includects/
> cp -p ./src/cts/* includects/
> cp -p ./src/mpfun90/* includects/
> cp -p ./src/qcdloop/* includects/
> cp -p ./src/makefile includects/
> cd includects && /Applications/
> gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c mpfun90.f90
> ar rv libcts.a mpfun90.o
> ar: creating archive libcts.a
> a - mpfun90.o
> gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c mpmod90.f90
> ar rv libcts.a mpmod90.o
> a - mpmod90.o
> gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c mpmodm90.f90
> ar rv libcts.a mpmodm90.o
> a - mpmodm90.o
> gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c mpmodx90.f90
> ar rv libcts.a mpmodx90.o
> a - mpmodx90.o
> gfortran -fPIC -fno-automatic -O2 -funroll-all-loops -c avh_olo.f90
> ar rv libcts.a avh_olo.o
> a - avh_olo.o
> gfortran -fPIC -fno-automatic -O2 -funroll-a...